Announcement

Collapse
No announcement yet.

Samsung 2693HM - Doesn't turn on (no standby)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #61
    Re: Samsung 2693HM - Doesn't turn on (no standby)

    Originally posted by twist3r View Post
    HOW they get charged? And with what voltage? How you can determ how much voltage is going to be inside one charged capacitor (without reading the markings on the capacitor itself).
    Simple.

    If you connect a 1.5V battery across a capacitor and then disconnect the battery, the capacitor will have.... guess what? 1.5V across it. If you do the same with a 9V battery, then the capacitor will have 9V across it.

    Now imagine you put a resistor between the battery and the capacitor. If you "very quickly" connect and disconnect your capacitor to the battery with the resistor between them, the capacitor will not be at the same voltage as the battery - it will be lower, because the resistor limits the rate of current going into the capacitor, and thus allows it to charge up slower. However, if you leave the capacitor connected to the battery (again with the resistor in between them) "long enough", eventually the capacitor will reach the same voltage as the battery.

    In the case of the circuits around IC602 and IC924, think of the SW pin as the battery and capacitors C621/C3045 as the capacitors in the above experiment. The main difference here is that instead of using a resistor like in the above experiment, the circuits use an inductor (L600/L902.) Why an inductor and not a resistor? Because with a resistor, there will be a lot of losses as current passes through it and generates heat. But with an inductor, the losses will be much smaller, because it can "store" some of the current that passes through it (at the peaks of the square wave pulses), and then give it back to the circuit (to capacitors C621/C3045) when there is nothing output between the peaks of the square waves.

    Originally posted by twist3r View Post
    I keep getting stuck completely when observing a cirucits and see that one end of capacitor is connected to a 'lines' which goes to ground, but...same lines are connected to other elements which should 'go' to close the circle. So I always keep asking myself 'how that dmn voltage is not lost, since its connected to the ground'.
    I know what you mean and I was getting stuck in a similar way when I was a beginner in electronics and didn't understand how circuits work.

    The reasons circuits are called "circuits" is because there is always a return path (or multiple return paths eventually all recombining back into a single return path.) Again, going back to batteries, since they are easier to understand... you CANNOT have a current go out of one terminal of the battery and not go back to the other. The battery is a electro-chemical reaction waiting to happen. By drawing current from one terminal of the battery and returning it to the other, you are essentially allowing the chemicals inside the battery to react and become more neutral. While doing so, however, you can use the current to do useful work for you. In the case of connecting a simple light bulb across a battery... as the current passes through the filament of the bulb, it makes it heat up so much that the filament starts glowing, which is what produces light. But the current does NOT get lost in the filament of the bulb. If let's say 0.5 Amps of current goes out of the (+) terminal of the battery and makes it to the bulb, then the same 0.5 Amps will continue to flow through the bulb and out of the bulb, back to the (-) terminal of the battery. So current CANNOT get lost. However, the voltage can... or rather, it gets reduced as it passes through the bulb. This is called a voltage drop.

    So in the case of complex circuits, like the ones surrounding IC602 and IC924, you have different components that manipulate the current to produce the desired functions you want the circuit to do. Some components will store current and release it later (like L600 and L902). Others, like the resistive divider formed by R604 and R606 simply "waste" power and turn it into heat... but it's needed so that the ICs can "read/see" the output voltages properly on their FB pins. And moreover, these resistors waste very little power, because they have relatively high resistances given the voltages in the circuit. And capacitors C621 and C3045 may seem like they waste power when they allow spikes of current to pass to ground when they charge up... but remember: as these capacitors charge up, they then can act like little sources/batteries themselves, thus returning the current that they seemed to waste initially.

    That said, here are some videos on the matter that you may find useful:
    https://www.youtube.com/watch?v=X4EUwTwZ110
    https://www.youtube.com/watch?v=rbCXKhhzBN0

    Originally posted by twist3r View Post
    Ok....but now comes the stuped question: why the negative leg of capacitor is connected 'further' to other cirucit?

    It's not. It's connected to ground... which is basically just a common point that other parts of the circuit use as a "reference" for the voltage levels they see/get.

    In other words, if we are both standing next to each other on the 1st floor of a building and look up to a window on the 3rd floor, we will both agree that this window is xx meters above us. But if I'm standing on the 1st floor and you're standing on the 2nd floor and we look at that same window on the 3rd floor again, we will not agree that the distance between us and that window is the same xx meters. You may say it's 3 meters and I may say it's 6, and we may both be right.

    So ground is just a common reference point for voltages. In the case of capacitors C621 and C3045, when they charge to 3.3V and 1.2V respectively, other parts in the circuit will see those voltages ONLY if both of these capacitors and the other parts of the circuit are using one same point as a reference. In this case, it's ground (though it doesn't have to be, and this will vary with different circuits.)

    Originally posted by twist3r View Post
    And 'when' that cirucuit is 'activated'. And...why it goes to other elements (like that schottky diode) but at the same time it 'goes' to the ground.
    The Schottky diodes D600 and D942 play a very important role when there is a current passing through inductors L600 and L902 but when the voltage on the SW pin is 0V between the peaks of the square wave pulses: it provides a return path for the current "stored" in inductors L600/L902 to continue to flow out of them and into capacitors C621/C3045. Without these Schottky diodes, inductors L600/902 will try to do "anything" to keep that current going through them, because... remember, inductors don't like instant changes in current. And that "anything" could mean the inductor could generate 1000's of Volts just so that current can jump through and continue to flow. If you're wondering what the heck I mean, think about a time when you have disconnected an inductive high-current device when it was turned On and drawing current - for example, an electric radiator/heater. If you have, notice sometimes you see blue sparks/lights at the plug when you did that? - That's the inductance of the heating element generating high voltage and jumping the air gap between the disconnected plug prongs and wall outlet to try to maintain the current going through it. Another example would be the sparks you see at the brushes on a motor (e.g. a drill.) The sparks are created from the inductive property of the coils in the rotor of the motor.

    So basically, this phenomenon is called inductive kickback.
    Diodes D600 and D942 simply provide a return path so that the inductive kickback from L600 and L902 doesn't generate extremely high voltages that can damage IC602 and IC924. Moreover, these diodes allow the inductive kickback from L600 and L902 to supply current into C621 and C3045 between the voltage pulses from the SW pins. So L600/L902 and C621/C3045 are used in tandem as energy storage devices to transform the square wave pulses into a smooth DC voltage and provide power to the rest of the components on the logic board.

    Again, look up how a buck regulator circuit works, and it may get more clear.

    Originally posted by twist3r View Post
    I have real issue while reading schematics, especially when I observe the lines connected to the ground. Everytime I see a ground sign, I say 'if the 'voltage' touch the ground, everything is wasted and nothing exist. But seeing other elmeents connected to the ground and at the same time being part of the full circuit makes me 'hang' and my brain 'freeze'.
    Well, the voltage never really directly "touches" ground. If it did, that would be the same as placing a wire across the (+) and (-) of a battery or in the plug of you wall, and you know what that does, right.

    But when you have resistors, capacitors, inductors, and other elements going to ground, that doesn't mean they are uncontrollably passing current to ground. Rather, it's exactly the opposite of that - they pass current in a controlled manner, again with the idea to manipulate it to do various function in the circuit. In the case of caps, they'll store charge. In the case of inductors, they'll maintain current. In the case of resistors... they may waste power, but divide voltages or limit currents going to other components.... and so on.

    Originally posted by twist3r View Post
    I'm trying to figureout how you got to some conclusions.
    No shame in that. It's what this forum is for, after all: sharing knowledge. I've received a ton of help and knowledge from here, so now it's my turn to return the favor if I can.
    Last edited by momaka; 12-23-2020, 12:02 AM.

    Comment


      #62
      Re: Samsung 2693HM - Doesn't turn on (no standby)

      Thanks momaka for the ^ explanation! I would have to re-read some parts of your post and try to understand them. There are certain things which are 'basics' but my mind just refuse to 'click' and understand them. The more I'm trying, the more I'm getting angry for not understanding simple things.

      Let me try to answer on some of your request from your first post ^^.

      Originally posted by momaka View Post
      So for C306 and C307, you can try removing them to see if the resistance to ground of the AVDD_DVI line increases. On that note, same applies to C370, since it is in parallel with C373, which you measured as 1 Ohms, and which is connected to pin 69 (AUVREF).
      Ok, so...

      I measured resistance to ground of 'left' pad (of L307) and it was 245.6 Ohm. Then I measured resistance of the 'right' pad (of L307) and it was 1.1 Ohm.

      I then, removed C306 and measured resistance of both pads (of L307). No change.

      Then removed C307 and again measured resistance of both pads (of L307). And again, no change.

      I put back both C306 & C307. I must say it was not an easy task. I hope I solder them back properly cause they are a bit 'forward slash' position (or backward slash), compared to their previous postion on their pads. I measured resistance of both C306 & C307 while they were off the board and resistance was on MOhms range.
      Once I put them back resistance to ground of one end of capacitor was 0.2 Ohm and the other is around 0.9 Ohm. I guess its a bit lower compared to previous state because of my bad (forward slash) soldering.

      Ok, ofter that I removed C370, but it was all the same when it comes to measuring resistance at both pads of L307. So I soldered C370 back to its place (again with forward-slash position). It was so hard to do it with my tools. Now desoldering such small components is so easy, compared to soldering them back with my soldering iron.


      Originally posted by momaka View Post
      Check/remove the above ceramic capacitors mentioned. If they don't measure short-circuit, then they are probably OK. However, that could imply IC300 is bad. Before we conclude that, though, if you can / have the nerves, try measuring across all ceramic capacitors to see if any more measure 2 Ohms or less. Once you find any that do, just note them down for now.
      Here are the ones that have very small resistance. I think that I managed to check all the caps.

      C 304 = 3.8
      C 305 = 2.6
      C 306 = 1.0 -> desoldered/soldered back
      C 307 = 1.0 -> desoldered/soldered back
      C 309 = 1.0
      C 322 = 1.2
      C 324 = 1.2
      C 325 = 1.2
      C 326 = 1.1
      C 327 = 1.3
      C 328 = 1.1
      C 329 = 1.2
      C 330 = 1.3
      C 331 = 1.2
      C 332 = 3.7
      C 333 = 1.2
      C 334 = 3.7
      C 335 = 3.6
      C 336 = 3.7
      C 337 = 3.6
      C 338 = 3.7
      C 340 = 0.9
      C 341 = 0.9
      C 342 = 0.9
      C 343 = 2.0
      C 358 = 3.7
      C 370 = 1.0 -> desoldered/soldered back
      C 510 = 0.3
      C 518 = 0.3


      Electrolytic caps:
      C 367 = 3.7 Ohm
      C 373 = 0.9 Ohm
      C 3045 = 3.6 Ohm

      ^ from what I see at page 57 of service manual, C510 & C510 are part of 'PC_SOUND_INPUT circle' and I don't remember that I ever used that port for anything. Don't know why their resistance is so low.


      Originally posted by momaka View Post
      On that note, if you do perform the power-up experiment again with the main/logic board connected to the PSU board, does anything on the main/logic board get very hot, like IC300 or any other chip?
      I checked and only IC300 gets very gentle worm. I barely noticed. Power board & main board were connected for probably 5-10 min. And nothing changed. If I didn't expect IC300 to get worm I almost woudn't even notice. Other ICs were cold.


      Originally posted by momaka View Post
      Electrolytic cap C373 is rated for 35V, so it is unlikely to get damaged before IC300. Same goes for C367, which is rated for 16V - a lot higher than what IC300 could withstand, so it's unlikely that either of these electrolytic caps got damaged. But the ceramic caps across them, C370 and C358, respectively, could have taken a hit.
      Well, I started re-reading your reply but went to desolder C306, C307 and C370. I was thinking to desolder C358 as well, but since you didn't asked me to do that in the first part of your post, I didn't. If you want I can do that. ....As soon as I gather my patience and feeling that my hands are calm again, because it wasn't easy to put the capacitors back since they kept being sticky to my caveman version of tweezers. OMG those caps are so small!


      Originally posted by momaka View Post
      With that said, I think the next thing for you to try is to solder back L302, L303, and L304, since you measured fairly high resistance on their "right" pads on the board.
      OK, I put them back.


      Originally posted by momaka View Post
      Also, what resistance (to ground) did the other of these inductors measure on their "right" pads - i.e. L300, L301, and L305.
      L300 = 246.1 Ohm (left pad) | 0.8 Ohm (right pad)
      L301 = 246.1 Ohm (left pad) | 2.0 Ohm (right pad)
      L305 = 246.0 Ohm (left pad) | 1.2 Ohm (right pad)


      Originally posted by momaka View Post
      We already know AVDD_DVI line (L307) is no good, so DO NOT solder that one back in. But if the others show high enough resistance (over 50 Ohms minimum, I'd say), then you can solder back their corresponding inductors too. That way, maybe we can get IC300 to work partially without L307 soldered back... which with a bit of hope, maybe you just loose the DVI/HDMI input but could still use the VGA input? IDK, we will see.
      Well, since resistance at 'right pad' of L300, L301 and L305 is very low (far below 50 Ohms) I didn't solder them back.


      Originally posted by momaka View Post
      Once you solder these back, power on the monitor again. You may also want to connect the pannel / front button board to the logic board to see if your power LED works and if you can turn the monitor On and Off. If you can't, check again if the +3.3V_LIVE and +1.2V_LIFE are still OK and report back what you saw/measured.
      Since I didn't soldered back L300, L301 and L305 because of their lower resistance, I didn't proceed further with powering up power and main/logic board (along with pannel/front button). I'm awaiting further instructions, if there are any left.
      Last edited by twist3r; 12-24-2020, 11:34 PM.

      Comment


        #63
        Re: Samsung 2693HM - Doesn't turn on (no standby)

        Well, few days ago I was obsereving a main/logic board from a 2493HM monitor, since the model is the same as for my (2693HM) monitor.

        Model of the main/logic board is: BN41-00962A

        At first I thought that the guy (the seller) was selling some faulty board, because I noticed that some elements are missing. But then, I remember that in the service manual for my monitor (which is the same for 2493HM monitor as well), had some parts where it sais "2493HM DELETE".
        So I was like: Ahaaa, ok!

        There are 3 'areas' in the service manual with '2493HM DELETE' note.
        Two of them are on page 60. The note sais that 2493 model doesn't have a D602 shottky diode.
        Also on the same page there is a note for 2 capacitors:
        C609 (ceramic cap) &
        C610 (electrolytic cap)

        They are also missing on 2493HM main/logic board (and also in the manual).

        One thing that I don't understand is why 2493HM mainboard doesn't have D602 diode? That diode is first on '+18V_IN' line. I mean, if there is not a diode there, the line is broken and the line can't provide 18v. How is that possible?


        And... the 3rd note on the service manual is on page 59, and it shows that there should not be present two ceramic capacitors. Ok, but I tried to find them on my board. These caps are:

        C387 &
        C388

        These 2 caps are the last 2 elements on the way to 'PWM_DIM' connection.

        Main idea was (if it comes to that) to buy a 2493HM main/logic board and 'transfer' missing elements from my main board to a main/logic board of 2493HM monitor (since the model of the board is the same, BN41-00962A). But thing is that even after almost one hour of looking through a magnifying glass, I coudn't find them on my main board. They are just not there, like they don't exist. They should be located somewhere in top, top-right area of the board. But they are not there. In the service manula, prior these 2 caps there is a R348 resistor, which I found (and marked it with yellow arrrow on the main board image I'm gonna attach). So now I think that using a main board from 2493HM monitor for my 2693HM monitor wont work, since I can't find and transfer these 2 ceramic caps from 2693HM to 2493HM main board.

        What do you think?
        Attached Files
        Last edited by twist3r; 12-29-2020, 08:12 PM.

        Comment


          #64
          Re: Samsung 2693HM - Doesn't turn on (no standby)

          Originally posted by twist3r View Post
          I measured resistance to ground of 'left' pad (of L307) and it was 245.6 Ohm. Then I measured resistance of the 'right' pad (of L307) and it was 1.1 Ohm.

          I then, removed C306 and measured resistance of both pads (of L307). No change.

          Then removed C307 and again measured resistance of both pads (of L307). And again, no change.

          I put back both C306 & C307. I must say it was not an easy task. I hope I solder them back properly cause they are a bit 'forward slash' position (or backward slash), compared to their previous postion on their pads. I measured resistance of both C306 & C307 while they were off the board and resistance was on MOhms range.
          Once I put them back resistance to ground of one end of capacitor was 0.2 Ohm and the other is around 0.9 Ohm. I guess its a bit lower compared to previous state because of my bad (forward slash) soldering.
          No, you soldering sounds to be OK. 0.2 Ohms to ground on the ground side and 0.9 Ohms on the AVDD_AU side with or without C306 & C307 means AVDD_AU is just likely shorted inside IC300... which is not good... but it is what it is.

          Originally posted by twist3r View Post
          Ok, ofter that I removed C370, but it was all the same when it comes to measuring resistance at both pads of L307. So I soldered C370 back to its place (again with forward-slash position). It was so hard to do it with my tools. Now desoldering such small components is so easy, compared to soldering them back with my soldering iron.
          Yes, the soldering back on, especially of the very small SMDs, is the hardest part. I usually use my nails to hold down those SMDs while soldering them. It sounds weird, but works very easily. You could probably do it with a sharp metal edge/object too. Just make sure the pads of the SMD are somewhat clean from solder, but not completely clean. Then place flux on pads, after that position the SMD on top properly. Finally, gently press down on the SMD with the sharp object (a cut out from an aluminum can... or in my case, I just use the nail on my thumb) and touch ones side of the SMD and pad with the soldering iron. Useful tip #1: I rest the hand I use for holding down the SMD right on the PCB and on a steady soft surface. That way, the PCB can't move or slide anywhere. Useful tip #2: make sure to use a sharp soldering tip. BC2, BC3, C4, DL32, and K should do the trick. It doesn't have to be a small tip, just have a sharp edge somewhere on it. Once one side of the SMD is soldered, you can turn the board and solder the other side of the SMD and pad.

          Originally posted by twist3r View Post
          Here are the ones that have very small resistance. I think that I managed to check all the caps.

          C 304 = 3.8
          C 305 = 2.6
          C 306 = 1.0 -> desoldered/soldered back
          C 307 = 1.0 -> desoldered/soldered back
          C 309 = 1.0
          C 322 = 1.2
          C 324 = 1.2
          C 325 = 1.2
          ...
          ...
          Hmmm.. C322, C324-C331, & C333 are all tied to VDDP, which I can't find what generates that rail anywhere in the SM. It appears to come out of pin 97 on IC300. Would be interesting to see what voltage one of these caps measure when there is power given to the monitor. I doubt any of these ceramic caps would be shorted, though.

          As for C332, C334-C338, and pretty much all the caps the read around 3.6-3.8 Ohms... those are all connected to the +1.2V_LIVE, so no surprise about their resistance.

          We'll keep those resistances in mind for now, but not much else to worry about here.

          Originally posted by twist3r View Post
          ^ from what I see at page 57 of service manual, C510 & C510 are part of 'PC_SOUND_INPUT circle' and I don't remember that I ever used that port for anything. Don't know why their resistance is so low.
          Don't worry about those.
          I think the headphone jack CN500 just grounds them out when they are not used, so probably that's why.

          Originally posted by twist3r View Post
          I checked and only IC300 gets very gentle worm. I barely noticed. Power board & main board were connected for probably 5-10 min. And nothing changed. If I didn't expect IC300 to get worm I almost woudn't even notice. Other ICs were cold.
          OK, that's good so far.

          Originally posted by twist3r View Post
          Well, I started re-reading your reply but went to desolder C306, C307 and C370. I was thinking to desolder C358 as well, but since you didn't asked me to do that in the first part of your post, I didn't.
          Yeah, that's fine. You can do that later, if needed.

          Originally posted by twist3r View Post
          OK, I put them back.
          Just to make it clear for anyone else reading, this is for L302, L303, and L304, since they read higher (normal?) resistance.

          With that done, try giving power to the monitor (PSU board + main/logic board connected) again, and see if +3.3V_LIVE and +1.2V_LIVE are still there. And also check if IC300 is very hot too after a minute or two.

          If 3.3V and 1.2V rails are OK and IC300 is not too hot (maybe warm, but not "burning-my-finger" hot), then continue with instructions below.

          Originally posted by twist3r View Post
          L300 = 246.1 Ohm (left pad) | 0.8 Ohm (right pad)
          L301 = 246.1 Ohm (left pad) | 2.0 Ohm (right pad)
          L305 = 246.0 Ohm (left pad) | 1.2 Ohm (right pad)
          OK, these don't look very good, so let's try these things:

          1st: solder back L301. Then connect main/logic board to PSU board and power up. See if 3.3V and 1.2V are still there... and if yes, then check if IC300 is getting hot or not again.

          If 3.3V and 1.2V are present and IC300 not too hot, solder back L305. Then do the same experiment as with L301 above.

          If that passes too, repeat once more with L300.

          Basically, we want to see which one of these will bring down the +3.3V_LIVE. Whichever one of these does it (could be all of these... i.e. L300, L301, and/or L305), desolder it's corresponding inductor (L30_) again. We already know L307 (AVDD_DVI) is pretty much shorted to ground, so don't bother soldering that back in.

          Once you find which ones don't short IC300 to ground, feel free to do an experiment and connect the front panel button board, T-con board, and inverter again. It probably would be a miracle if the monitor powers up and displays a "No Signal" message or Samsung logo with no device connected... but worth a try.

          If not, and monitor is still unresponsive to front panel button board commands, then I suppose that's the end of the line for IC300.

          You could try (if you want) to force 3.3V into the right pads of L307 and any other L30x right pads that short the +3.3V_LIVE (i.e. not allow +3.3V_LIVE to come up when main board is powered by PSU) with an external PC PSU... and hope that whatever is shorted inside IC300 can burn out and not cause a problem anymore. Though in all honesty, this will probably achieve nothing... or maybe blow even more magic smoke out of IC300. Probably not worth wasting time with, though.
          Last edited by momaka; 12-30-2020, 07:27 PM.

          Comment


            #65
            Re: Samsung 2693HM - Doesn't turn on (no standby)

            Originally posted by twist3r View Post
            One thing that I don't understand is why 2493HM mainboard doesn't have D602 diode? That diode is first on '+18V_IN' line. I mean, if there is not a diode there, the line is broken and the line can't provide 18v. How is that possible?
            It means the +3.3V_LIVE and +1.2V_LIVE on the 2493HM are only generated from the 5V_MICOM line.

            One possibility is that the 2493HM board uses less power and doesn't need the +18V_IN when the monitor is turned On.

            But you can always solder D602 and the corresponding capacitors back on the 2493HM board from your 2693HM board... and that should make IC602 & IC924 just work with the 18V line, like it did on your 2693HM board.

            In any case, I wouldn't worry too much about this discrepancy.

            Originally posted by twist3r View Post
            And... the 3rd note on the service manual is on page 59, and it shows that there should not be present two ceramic capacitors. Ok, but I tried to find them on my board. These caps are:

            C387 &
            C388

            These 2 caps are the last 2 elements on the way to 'PWM_DIM' connection.
            Yeah, I don't see them at all on your main board nor any spots for them either. It looks like the trace from resistor R348 goes directly to L601 and then out to CN601. So maybe they did a revision on the PCB but not to the SM... which doesn't really surprise me too much. So probably don't worry about this either.

            Originally posted by twist3r View Post
            Main idea was (if it comes to that) to buy a 2493HM main/logic board and 'transfer' missing elements from my main board to a main/logic board of 2493HM monitor (since the model of the board is the same, BN41-00962A).
            ...
            So now I think that using a main board from 2493HM monitor for my 2693HM monitor wont work, since I can't find and transfer these 2 ceramic caps from 2693HM to 2493HM main board.
            I don't think those two caps are a deal breaker. There's a good chance they might be missing from the 2493HM board too. And those are just for backlight / dimming control. So worst case, if you connect the 2493HM board to your monitor, the backlights might not work properly or dim properly... but we can probably get them to work again. Most likely, it would require that the PWM_DIM, A_DIM, BRT_CTL, and LED_GREEN circuits on the 2493HM board are matche to the ones on your 2693HM board... and *maybe* swap the firmware chips, if needed (though you actually might not need to.)

            With that said, if you can find a 2493HM board that is relatively inexpensive to you, then you can give it a try. But if it isn't cheap and/or you don't feel it's worth the gamble, then keep your 2693HM monitor for parts, in case you do run into a spare board some day. Alternatively, you can also look for a replacement of IC300... though that will require hot rework and may not be too easy if you've never soldered chips of that size before.

            Comment


              #66
              Re: Samsung 2693HM - Doesn't turn on (no standby)

              momaka, I wish you a Hapy New Year! (and same to everyone else, ofcorse)

              Also, thanks for again replying to my topic.

              Originally posted by momaka View Post
              Yes, the soldering back on, especially of the very small SMDs, is the hardest part. I usually use my nails to hold down those SMDs while soldering them. It sounds weird, but works very easily. You could probably do it with a sharp metal edge/object too. Just make sure the pads of the SMD are somewhat clean from solder, but not completely clean. Then place flux on pads, after that position the SMD on top properly. Finally, gently press down on the SMD with the sharp object (a cut out from an aluminum can... or in my case, I just use the nail on my thumb) and touch ones side of the SMD and pad with the soldering iron.
              Useful tip #1: I rest the hand I use for holding down the SMD right on the PCB and on a steady soft surface. That way, the PCB can't move or slide anywhere.
              Useful tip #2: make sure to use a sharp soldering tip. BC2, BC3, C4, DL32, and K should do the trick. It doesn't have to be a small tip, just have a sharp edge somewhere on it. Once one side of the SMD is soldered, you can turn the board and solder the other side of the SMD and pad.
              Yea, it was very hard. I used my tweezers, which are not the kind of tweezers meant for soldering. And while I was using the tweezers along with Edsyn FL-22 flux-gel, capacitor gets stickied to a tweezer. So a lot of time i was like 'where the f* that capacitor dissapeared. It was there just a second ago'

              As for your 2nd tip, ...yea, I didn't thought about it. I used the same chisel shaped soldering tip, that I used for desoldering. Long time ago I was told that 'thorn' shaped tip can't help you a lot cose they can't transfer a lot of heat at the peak of the tip. I guess there are times where they are usefull, like with such a small capacitors (or resistors). I have a 'C 'shapred tip somewhere (I just currently don't remember its full markings).

              Thing that also made me wonder if I did solder it back proprely is the fact that such a small capacitor also cover its pads, so when I try to solder it, I can't put 'solder under it' but 'outside' at the edge. Its because I'm not using a hot air station, but a soldering iron. If I had soldering station I guess this would be much easier, but... I had to use what I have at hand.


              Originally posted by momaka View Post
              Originally posted by twist3r View Post
              Here are the ones that have very small resistance. I think that I managed to check all the caps.

              C 322 = 1.2
              C 324 = 1.2
              C 325 = 1.2
              C 326 = 1.1
              C 327 = 1.3
              C 328 = 1.1
              C 329 = 1.2
              C 330 = 1.3
              C 331 = 1.2
              C 333 = 1.2
              Hmmm.. C322, C324-C331, & C333 are all tied to VDDP, which I can't find what generates that rail anywhere in the SM. It appears to come out of pin 97 on IC300. Would be interesting to see what voltage one of these caps measure when there is power given to the monitor. I doubt any of these ceramic caps would be shorted, though.
              Well, I think they are connected to 3.3v rail through L305 (VDDP, SM page 59, top-right corner). Same way C306 & C307 are connected to a 3.3v rail through L307 (AVDD_DVI). I thought about it few days ago while I was desoldering C306 & C307, thinking 'ok...if we are checking this, maybe for the same reason I should desolder all these small caps (C322, C234-C331, C333) which are also connected to 3.3v rail. But after I barelly managed to solder back those 3 caps (C306, C307 & C370), I was so much tired and decided its 'best' to forget about that thought.

              In theory, could it be that some of these caps caused all the trouble? Although, you did said thats probably not the cause of all this mess.

              Originally posted by momaka View Post
              As for C332, C334-C338, and pretty much all the caps the read around 3.6-3.8 Ohms... those are all connected to the +1.2V_LIVE, so no surprise about their resistance.
              Ok. One thing that I want to ask here... It seams to me that there are certain values that are always 'expected'. For instance...

              For coil, expected resistance is 0.2-0.4 Ohm (right?)
              For 1.2v rails I guess expected resistance is 3.6-3.8 Ohms (not just for this, but any other board)

              So... what would be expected resistacnes for caps in a 3.3v rails? (or 5v rails)? If there is a 'patern' here, I would like to memorize it.

              Originally posted by momaka View Post
              Originally posted by twist3r View Post
              Well, I started re-reading your reply but went to desolder C306, C307 and C370. I was thinking to desolder C358 as well, but since you didn't asked me to do that in the first part of your post, I didn't.
              Yeah, that's fine. You can do that later, if needed.
              Hmm, ok. I'm already paranoid, about all the elements. I keep thinking 'what if this small piece of s* is the one that made my logic board become faulty?'


              Originally posted by momaka View Post
              Try giving power to the monitor (PSU board + main/logic board connected) again, and see if +3.3V_LIVE and +1.2V_LIVE are still there. And also check if IC300 is very hot too after a minute or two.
              I was creating my reply by trying to respond/comment on some of your replies. When I came to this part (to check +3.3V_LIVE and +1.2V_LIVE if they still exist, as well to check if the IC300 is getting hot), everything fell apart. I'll explain why, bear with me

              Ok, so I went to check the voltages.

              C3045 = 1.223v
              C619 = 3.294

              And IC300 was barelly warm.

              But then, I was like...'lets check few more things'. And very soon 'few more' became 'everything that I thought is important'. Some things surprised me and made me unable to figureout why I'm getting certain values which I thought are wrong (and maybe they are), so I'll write down what I did (while trying to keep correct order of the measurements that I did).

              Ok, so...

              I checked other 2 electrolytic capacitors:

              C610 = 1.545v
              C619 = 4.880v


              Then I checked shottky diodes:

              Note:
              =====
              first value is the voltage on the diode while one probe is touching one side of the diode, and the other probe is touching other side of the diode. Then, next value is th 'voltage to ground' of one side of the diode where black is on the ground point and red touching one side of the diode. And third value, is the same like 2nd value, just red probe touching other side of the diode, while black kept touching 'ground point'.

              D602 = 3.139v / 4.888v / 1.769v
              D601 = 0.326v / 4.887v / 5.214v
              D942 = 1.241v / 1.240v / 0v
              D600 = 3.298v / 3.297v / 0v



              Next thing, I checked voltages on some ICs:

              IC304: tab had 0v, one of the pins 5.213v and the other pin 5.212v (so I guess this one was good)

              IC113: tab had 0v, one pin had 3.290v and the other had 2.486v (which is again what it should be)


              IC602: (below are the pins and their voltages)
              ======
              1 = 6.880v (BS)
              2 = 4.880v (IN)
              3 = 3.297v (SW)
              4 = 0v (GND)
              5 = 1.218v (FB)
              6 = 1.097v (COMP)
              7 = 4.047v (EN)
              8 = 1.411v (SS)


              IC924: (below are the pins and their voltages)
              ======
              1 = 5.324v (BS)
              2 = 4.880v (IN)
              3 = 1.240v (SW)
              4 = 0v (GND)
              5 = 1.223v (FB)
              6 = 1.041v (COMP)
              7 = 3.300v (EN)
              8 = 1.382v (SS)

              ^ here everything looks ok. Only for IC602 pin 1 (BS) I guess that higher voltage is created because of the capacitor between BS and SW (C615). Datasheet for that IC (MP1583DN), for BS pin sais:

              High-Side Gate Drive Bootstrap Input. BS supplies the drive for the high-side N-Channel MOSFET
              switch. Connect a 4.7nF or greater capacitor from SW to BS to power the high-side switch
              So I guess all 4 checked ICs are ok.




              Then I went back to (again) measure voltages across the electrolytic capacitors and I noticed something.
              Now C610 didn't have 1.545v anymore, but 2.653v. Because its connected to a D602 shottky diode, I check that diode again. Here are the values (again, first number is when probes are on both sides of the diode, then 2nd number when black on the ground and red probe touching one side of the diode, and third number is when red touching the other side of diode, while black still on the ground).

              So...

              C610 = 2.653v
              D602 = 1.030v / 4.888v / 1.769v (soon it went to 1.901v)

              ^ thats what I marked on papper and continue measureing other things (will be back again to this subject for the 3rd time, since I'm writing correct order of what I was checking).

              Comment


                #67
                Re: Samsung 2693HM - Doesn't turn on (no standby)

                Next thing I did was to measure a voltage on L300-L305 and L307 ferrite beads. One thing that I learned here (by doing it) is that I can't measure a voltage of a ferrite bead which is soldered on the board by putting probes on both side of the ferrite bead, because measuremant is then 0.000v (or 0.001v). But I had to touch the ground point with a black probe and use red to touch sides of the ferrite bead.
                For those ferrited beads which were desoldered I coudl touch their paids with both probes touching one of the pads and I could get the voltages (I also could touch one ground with black and use red on the pad, just like in the case of the ferrite beads which is already on the board). So I did that too just to be 'sure' (and reason is below)...


                L300 = 3.289v / left pad to GND = 3.289v, right pad to GND = 0.001v
                L301 = 3.290v / left pad to GND = 3.290v, right pad to GND = 0.001v
                L302 = 3.290v (this is voltage to ground, since it was soldered on the board)
                L303 = 3.291v (this is voltage to ground, since it was soldered on the board)
                L304 = 3.293v (this is voltage to ground, since it was soldered on the board)
                L305 = 2.980v / left pad to GND = 3.293v, right pad to GND = 0.304v ...WTF?
                L307 = 3.293v / left pad to GND = 3.293v, right pad to GND = 0.001v
                BD1005 = 0.174v / left pad to GND = 0.174v, right pad to GND = 0.001v

                For all ferrite beads ^ above I would say...voltages are expected given the current state of the board (with ferrite beads desoldered), BUT...what makes me freakout is L305. It has lower voltage between its pads (2.980v). And while left pad to ground had 'corrrect' voltage of 3.293v, what makes me wonder is voltage of right pad to ground, which is 0.304v. Where the F* that voltage came from?

                So I checked Service Manual (page 59) and noticed that L305 was connected to VDDP line, but also bunch of ceramic capacitors were connected too. I then measured voltages of most of them I think (since it was hard to find them so easily).

                BTW. from time to time I was checking big IC300 and it was just gently warm.

                So:

                C322 = 0.305v
                C333 = 0.303v
                C326 = 0.305v
                C327 = 0.305v
                C328 = 0.305v
                C330 = 0.305v
                C331 = 0.304v

                ^ at the momment I coudn't find C324, C325 & C329, but I guess they woudl have the same voltage (of ~ 0.305v). Now, could it be that some of these caps caused the trouble, since their voltage is the same as the voltage of the right pad of L305. And since capacitors were connected (on the other end) to ground, ...I know its stuped, but... I guess they are getting that voltage from 'somewhere' (from some rail) which is 'touching' the ground somewhere where it shoudn't)?

                Btw, by accident, I also measured few more caps:

                C334 = 1.217v
                C337 = 1.213v
                C347 = 3.293v

                C334 & C337 are part of 1.8V_LIVE (or 1.2V_LIVE ?) rail, just right from the VDDP rail (page 59 of service manual). So.... I would guess their voltage is...ok?

                C347 is connected to L303 as part of AVDD_MPLL, so I guess its voltage of 3.293v is good.

                In this momment, I knew I wont be able to fullfill your request of testing the main/logic board the way you instructed me, at least not before I get the comment on all this from you.



                Ok, next thing I (again) went to check voltage on that C610 capacitor and the D602 shottky diode, so here are the results:


                C610 = 3.795v
                D602 = *forgot* / 4.888v / 3.799v (soon it went to 1.901v)

                note: *forgot* means that I forgot to measure voltages on the diode while touching every end of the diode with a probe. But I measured the voltage of left pad to ground and right pad to ground.

                Now, its clear to me that voltage of D602 'right' pad is the same as the voltage on C610 capacitor.
                What I don't understand is why it kept increasing over time? Also, why voltage across the diode dropped initially from 3.139v to 1.030v when I did 2nd measurement?

                D602 is part of 18V_IN rail. There is no (currently, because of faulty main/logic board) 18v rail. So, everything about that D602 diode is confusing me. Can you try to help me understand what you think is happening?




                Ok, so... next thing I did was to measure voltages of every pin of those diodes D105-D112 U D121-D128. Remember the D108 which was damaged and we didn't noticed it?

                Here is how I measured it: When you observe the diode with two pins on top, I marked the right one as 'pin 1', and left one as pin 2. The bottom one was 'pin 3'. I kept black probe on the ground, and only moved red probe from pin to pin.

                Here are the values:

                D### = pin 1 | pin 2 | pin 3
                =============================
                D105 = 0v | 3.293v | 3.293v
                D106 = 0v | 3.293v | 1.212v
                D107 = 0v | 3.293v | 0.215v
                D108 = 0v | 3.293v | 0v <===== damaged diode
                D109 = 0v | 3.293v | 3.293v
                D110 = 0v | 3.293v | 3.294v
                D111 = 0v | 3.293v | 3.293v
                D112 = 0v | 3.293v | 3.293v

                ^ Here I would 'guess' that most diodes show 'correct' values except for pin 3 of D106, D107 and ofcorse D108. Maybe the reason for low voltages on pin 3 of all of these three diodes is the D108 diode. I don't know.

                D### | pin 1 | pin 2 | pin 3
                =============================
                D121 | 0v | 3.293v | 0.041v
                D122 | 0v | 3.293v | 0.040v
                D123 | 0v | 3.293v | 0.039v
                D124 | 0v | 3.293v | 0.036v
                D125 | 0v | 3.293v | 0.034v
                D126 | 0v | 3.293v | 0.011v
                D127 | 0v | 3.293v | 0.171v
                D128 | 0v | 3.293v | 0.188v


                ^ Here.... I have no idea why voltages are so low on pin 3 for all the diodes (D121-D128).



                Ok, after this, I again check if the IC300 is hot. It was not hot, just maybe a bit more warm after 2 and a half hours of being under voltage. Thats how long it took me, because before I connected power and main/logic board to a socket wall, I checked resistance of some ceramic caps that were close to BD1005 ferrite bead, but they were all ok. Again, after 2.5 hours, IC300 was just warm. Other elements on main/logic board were colder than IC300, but mostly all of them were cold (it was hard to 'sense' that they are even warm).


                The last thing I did, was to measure a voltage of a BIG FAT cap on the power board.

                BIG FAT CAP = 303v

                Thats not the voltage I measured before (I think it was around 387v or somtehing like that, like 3 weeks ago). I didn't checked the capacitor as soon as I power te main/logic board, so I don't know what voltage was there, but I guess it was ~387. But why voltage dropped after 2.5 hours, thats what I don't get.

                After all this, I was 100% exhausted. A lot of questions and no strength to even write my respond.



                So, here it is. A 'respond'. Sorry about its length.


                And because of all this, I didn't tried to solder the remaining ferrite beads (L301, L305 and L300) yet, and power up the monitor. Btw. you probably forgot to tell me, but I guess you meant that I should also solder back BD1005 too
                Last edited by twist3r; 01-03-2021, 11:20 PM.

                Comment


                  #68
                  Re: Samsung 2693HM - Doesn't turn on (no standby)

                  Ok, some news here...

                  First, I was trying to figureout why the main board for 2493HM monitor has some elements (I counted 5) missing compared to the main board of my 2693HM monitor, even though the board model (and revision) is the same: BN41-00962A. Then, I ended up on shopjimmy website where he had 2 boards, but for different monitors (one for 2493HM and one for 2693HM). And finally I noticed the difference. At the back of the main/logic board there is a sticker with some code. Part of the code is what makes the difference (its a part number). Even when I compared Service manuals for both 2493HM and 2693HM it was 'obvious' but... I missed it.

                  Anyway, main/logic board part numbers are:

                  BN41-00962A (PART NO: BN91-01886A) - 2493HM
                  BN41-00962A (PART NO: BN91-01886B) - 2693HM

                  ^ such 'a small' difference


                  Originally posted by momaka View Post
                  Originally posted by twist3r View Post
                  Main idea was (if it comes to that) to buy a 2493HM main/logic board and 'transfer' missing elements from my main board to a main/logic board of 2493HM monitor (since the model of the board is the same, BN41-00962A).
                  ...
                  So now I think that using a main board from 2493HM monitor for my 2693HM monitor wont work, since I can't find and transfer these 2 ceramic caps from 2693HM to 2493HM main board.
                  I don't think those two caps are a deal breaker. There's a good chance they might be missing from the 2493HM board too. And those are just for backlight / dimming control. So worst case, if you connect the 2493HM board to your monitor, the backlights might not work properly or dim properly... but we can probably get them to work again. Most likely, it would require that the PWM_DIM, A_DIM, BRT_CTL, and LED_GREEN circuits on the 2493HM board are matche to the ones on your 2693HM board... and *maybe* swap the firmware chips, if needed (though you actually might not need to.)

                  With that said, if you can find a 2493HM board that is relatively inexpensive to you, then you can give it a try. But if it isn't cheap and/or you don't feel it's worth the gamble, then keep your 2693HM monitor for parts, in case you do run into a spare board some day. Alternatively, you can also look for a replacement of IC300... though that will require hot rework and may not be too easy if you've never soldered chips of that size before.
                  OK, soooo.... I found a main/logic board for (what I think cheap price). Its a board pulled out from 2493HM monitor. (price was ~ $8.5).

                  I then desoldered 3 elements which I managed to find on my monitor board (the other 2 remaining ceramic caps I coudn't find)

                  I desoldered:

                  C609 - ceramic capacitor
                  C610 - electrolytic capacitor
                  D602 - shottky diode

                  While desoldering small tiny C609 ceramic capacitor...I lost that dmn thing somewhere. I still don't understand how it just dissapeared. I'm sure it was in the solder 'bubble' but once I dropped the bubble on the table, I couldn't find it. Even if I melted the bubble few more times, and even squeezing the hot solder bubble with my fingers, trying to feel a 'sand' object between the tips of my fingers. But nothing. I then checked service manual to see what was the capacitor capacity. It was 100 nF. Then I desoldered some small caps from some spare boards. Used my ESR meter and check the capacity while it was set to 120 Hz. I found one with 102 nF and use that instead of the one I lost.

                  So I put that ceramic cap on 2493HM board. (I wont mantion that I was so pizzed cause it was so hard to solder it with soldering iron while the dmn capacitor was so small. Didn't helped me momaka's advice to hold it with my fingernails... cause I don't have any (guess I have to grow some).

                  Then I soldered back D602 shottky diode and finally C610 capacitor. I have to say that desoldering surface mount elements is not easy as well, with soldering iron. Simply because their pads are not just 'outside' in the open, but also under them. But ok, I did that. It was much easier than soldering back that small C609 ceramic baztard.


                  Then, I connected power board to main/logic boarrd, and also connected my (shallow) monitor frame (to be able to press 'power on' touch-sense button). I also soldered two wires to Power_EN pin on the power board, as well as another wire to GND pin of power board. And then connected both wires to red and black probes of my multimeter. Finally I connected power board to a wall socket.

                  I checked the multimeter but it was already showing 3.214 volts. For a sec, I thought 'how'? I didn't touched my monitor frame power button. Then I noticed led diode on my monitor frame is already lit. I guess...it was let in that state before ...or I don't know, maybe it reacted while I was attaching monitor frame (tiny) cable to a main/logic board. Anyway, I pressed the monitor power button and blue led turned off, and voltages on POWER_EN pin of power board dropped to zero. Again, I pressed the power on button on monitor's frame and led light show up shinning blue, and voltages are back to 3.214v on power board POWER_EN pin.

                  Ok, then I checked the voltages on power board, on connector pins where it was expected to be 18v and 5v, and there they were. 18.* and 5.* volts. So, I concluded that it seams like it 'works'. The only thing that confused me was that the POWER_EN pin on the power board had 3.214v (which I think is something like 3.3v).

                  Anyway .... ok then, I checked one arhive that I got (I think here, somewhere) for 2493HM monitor, which included schematic for power board as well (since most datasheets only has PDF file for main/logic board). I was thinking ....'2493HM main/logic board,.... lets see if they mantion what voltage is expected on POWER_EN pin (datasheet was for power board BN44-00192B & BN44-00195A) And then...I think I found it:

                  It said (yellow box):
                  PFC Voltage Booster circuits.
                  When QB851 turns on by the PS_ON ( = about +3V) signal form the main board, Transistor in OPTO PC80s is then turn on, which will turn on QB802, which will supply VCC for PFC IC Controller ICP801S, that is when the DC gets boosted up by the MOSFET QP801S, Diode D802, Inductor LP801.
                  So... if my logic doesn't tricks me... I would say that expected voltage for POWER_EN pin is ~3.3v and not 5v as I was expecting. Correct?




                  I was expecting to get 5v there, at least thats how I undertood it. I'm reffering to post #26, here is the quote:


                  Originally posted by momaka View Post
                  Originally posted by twist3r View Post
                  I was told that standby voltage should be present at powerboard connector pin, and only then, it will be able (once its shorted to the PS_ON) to provide voltages to other boards which are connected to a power board. Now I'm thinking.... my monitor is disassembled. If I don't touch (press) monitor power button, how can then the voltage be sent to other boards of the monitor? I guess when I press power button on a monitor, it short 'standby' pin with PS_ON pin and then voltage is 'sent' to main(logic) board? Did you assumed that all other monitor parts are all connected together so that I can check voltage on POWER_EN pin? They are not connected...
                  So when you try to measure POWER_EN / PS_ON, you should have, at a minimum, the PSU board connected to the logic board.

                  The way it works is that PSU board generates 5V standby on pin ST_BY (on connector CNM803). This 5V standby voltage powers the logic board in standby / soft-off mode (and on the logic board, ST_BY is connected to 5V_MICOM on connector CN600.) When you press the power button on the monitor, the logic board interprets that and sends a signal on pin POWER_EN of CN600 (which goes to PS_ON pin on the power supply board on connector CNM803.) The power board "reads" the POWER_EN / PS_ON signal and turns On the rest of the voltages - i.e. 18V, 5V, and whatever other voltages there are.

                  Depending on implementation / design, typically POWER_EN / PS_ON either has to be pulled high (to 5V) or low (to 0V or grounded) to make the power supply board turn On. From what I have seen, most monitors use an active high signal to turn On the PSU (meaning you have to feed 5V into PS_ON pin to turn on the monitor.)

                  But in the above case with the voltage measurement at PS_ON / POWER_EN, I'm trying to have you do that so we can determine if the monitor is trying to turn On or not and also to confirm if PS_ON is active high or not... though I would guess that it is.)
                  ^ could you help me understand? Could 3.3v actually be the correct voltage for POWER_EN pin? (since it also managed to 'activate' the power board and make it produce 18v and 5v on its connector pins). Or, I have to get 5v on POWER_EN pin and something is not fully right here?



                  I also checked the service manual for my 2693HM monitor, by searching 'POWER_EN' ...and on page 17 its mantioned to check if there is a signal on R391 (they never mantioned what voltage value should be expected). Then, I saw 2 things. Even though its not connected to, there are resistors there right next to R391 which are connected to 3.3v line. Maybe that sais something but there is no connection there to R391. Then I checked to which pin that R391 is connected to a big IC300 chip. It connects to GPIOE[O] (pin 228). But there is no mantion what voltage comes out of that pin. Then I tried to google and try if I can find datasheet for that big IC300 chip: MST97889CLD-LF. But ofc, there is no datasheet available for download anywhere. Would be great if someone has such datasheet, but I doubt that.
                  Attached Files

                  Comment


                    #69
                    Re: Samsung 2693HM - Doesn't turn on (no standby)

                    Freaking 10K limit...


                    Anyway, I'm trying to allow myself to think that the new main/logic board should work 'ok'. But I got some questions before I proceed:

                    1. The fact that I used spare 102nF instead of 100nF capacitor (which I lost) is no big deal, correct?

                    2. I want to dissolder firmware chip from my 2693HM main board and solder it to 2493 main board. Two questions here:

                    a) whats the easiest way to desolder it (and not damage all 8 legs) with an soldering iron? (I got no hot-air). I was desoldering these ICs before but I didn't cared much since I just played 'desoldering' game. But now I must not brake things so that puts preasure on me.

                    b) I thought that there is a one firmware chip but now I'm confused because there are 2 and I don't know which chip is the one that I should desolder.

                    IC303 - MX25L4005 (this is the one I thought to desolder, and service manual list it as: IC-FLASH MEMORY; MX25L4005, 4Mbit, 512Kx8Bi - SM list it as IC115)

                    OR

                    IC302 - S-24CS08AFJ-TB-1GE (this one got me thinking because service manual list is as: IC-EEPROM; S-24CS08AFJ-TB-1GE, 8Kbit, 1Kx8 - SM list is as IC112)


                    ... OR.... I should desolder and switch BOTH chips?


                    My guess was MX25L4005 (even my MiniPro programmer found it in its base) but then this IC302 (in service manual is close to a 'scaler' written in red) maybe that is for making a monitor 'stretch the picture' of whatever its function is (I guess I'm making no sense). Btw. where EDID information is written?

                    What do you think? I'm not russhing things (if you didn't noticed that by now ), so I'll not proceed until someone more experienced and smarter than me answer.

                    Comment


                      #70
                      Re: Samsung 2693HM - Doesn't turn on (no standby)

                      One extra information that I've found in another topic (post #26):

                      Samsung SyncMaster 2693hm - Dead backlight

                      While ensign.fodder's monitor had different issue, he measured some voltages there and, as it can be seen, he also found that the POWER_EN (or PSUs Power ON/OFF) pin has 3.2v, which is exactly what I got (3.214v).
                      So that might also confirm that ~3.2v (~3.3v) is expected voltage on POWER_EN (Power ON/OFF) pin. What do you think?
                      Attached Files

                      Comment


                        #71
                        Re: Samsung 2693HM - Doesn't turn on (no standby)

                        Originally posted by twist3r View Post
                        momaka, I wish you a Hapy New Year! (and same to everyone else, ofcorse)
                        Thanks!
                        Happy new year to you as well!

                        Originally posted by twist3r View Post
                        As for your 2nd tip, ...yea, I didn't thought about it. I used the same chisel shaped soldering tip, that I used for desoldering. Long time ago I was told that 'thorn' shaped tip can't help you a lot cose they can't transfer a lot of heat at the peak of the tip.
                        Yes, sharp conical "thorn" tips are NOT good indeed. I despise them not just for SMD, but anything in general. Only very occasionally when I have to solder something in a very tight spot (between other parts that I can't remove, for example) is the only time I might try to use it. But in general, I usually find a different way.

                        Originally posted by twist3r View Post
                        I have a 'C 'shapred tip somewhere (I just currently don't remember its full markings).
                        Yes, those C-shaped tips are good for SMD. But the chisel you have is also good.

                        Originally posted by twist3r View Post
                        Thing that also made me wonder if I did solder it back proprely is the fact that such a small capacitor also cover its pads, so when I try to solder it, I can't put 'solder under it' but 'outside' at the edge. Its because I'm not using a hot air station, but a soldering iron.
                        That's fine. As the solder on the side of the cap melts, it will also melt the solder on the bottom of the cap. So it should be OK. The key is to touch both the cap lead AND the SMD pad to which the cap is supposed to be soldered at the same time (one one side, that is.) Once you do that, your SMD component should only be soldered on one side. To make sure that it is soldered properly, try moving it. If it can't be moved and seems soldered to the board, then it's good. After this, you can then solder the cap on the other side.

                        Originally posted by twist3r View Post
                        If I had soldering station I guess this would be much easier, but...
                        If you have a good hot air station... yes.
                        But with the cheaper ones... no.
                        I have a cheaper one and I *avoid* using it as much as possible for SMD components, unless I really have to. I find it easier to do SMDs with a soldering iron - especially stuff with only 2 leads.

                        Originally posted by twist3r View Post
                        In theory, could it be that some of these caps caused all the trouble?
                        With small ceramic caps in parallel, it only takes ONE to short out and make the rest look like they are all shorted. But you appear to have multiple pins of IC300 show low resistance to ground @ L300-L307. So I think more likely IC300 is shot.

                        Originally posted by twist3r View Post
                        Ok. One thing that I want to ask here... It seams to me that there are certain values that are always 'expected'. For instance...

                        For coil, expected resistance is 0.2-0.4 Ohm (right?)
                        For 1.2v rails I guess expected resistance is 3.6-3.8 Ohms (not just for this, but any other board)

                        So... what would be expected resistacnes for caps in a 3.3v rails? (or 5v rails)? If there is a 'patern' here, I would like to memorize it.
                        No, there is no pattern or expected resistance value at all.

                        The resistances are just a collection of values I've gathered "in the field" when doing monitor repairs. Like for example, if you try to measure the resistance of the GPU on a PC graphics card, you will notice the resistance is typically very low - often 1.5 to 4 Ohms (though could be higher too.) CPUs and RAM, on the other hand, tend to have slightly higher resistances... though not always. And in the case of low-power ICs like IC300 in your monitor, I -expect- the resistance to be somewhere in the 100's of Ohms or more. That doesn't mean that it will be. But generally, it is.

                        And remember, GPUs, CPUs, RAM, and all other ICs are active devices - meaning the resistance you measure with a multimeter is only static when the monitor is off. When power is applied to the monitor, the resistance can change dramatically. Of course, you can't measure it then, because you CANNOT use a multimeter to measure resistance when there is power in the circuit.

                        So when I saw the low 1.x Ohms resistance on 3.3V_LIVE, I knew that had to be a red flag, because a relatively "small" IC like IC300 generally should NOT exhibit that low of a resistance. Thus, it seemed more like something was shorted. And by removing L300-L307 and the 3.3V_LIVE returning back to normal working order, you confirmed that something is likely bad inside of IC300.

                        Originally posted by twist3r View Post
                        But then, I was like...'lets check few more things'. And very soon 'few more' became 'everything that I thought is important'. Some things surprised me and made me unable to figureout why I'm getting certain values which I thought are wrong (and maybe they are), so I'll write down what I did (while trying to keep correct order of the measurements that I did).
                        ...
                        Note:
                        =====
                        first value is the voltage on the diode while one probe is touching one side of the diode, and the other probe is touching other side of the diode.
                        That usually doesn't give any useful information when checking voltages, so in the future, don't do it unless you're asked to.

                        The only thing this is useful for is to determine forward voltage drop of a diode in a DC circuit. Thus, you can sometimes use it to see if a diode has failed open-circuit. But that is extremely rare. Diodes -usually- fail short-circuit.

                        Originally posted by twist3r View Post
                        Then, next value is th 'voltage to ground' of one side of the diode where black is on the ground point and red touching one side of the diode.
                        Good.
                        That's the correct way to measure voltage in a circuit.
                        The reason you use ground is because you want to have a common reference point for all voltages.

                        Originally posted by twist3r View Post
                        Next thing, I checked voltages on some ICs:
                        ...
                        ^ here everything looks ok. Only for IC602 pin 1 (BS) I guess that higher voltage is created because of the capacitor between BS and SW (C615).
                        Yes, that's normal.
                        The bootstrap pin on PWM driver ICs is always at some voltage higher than the input. It's needed in order to turn On the Gate of the upper MOSFET in a switching/buck regulator circuit.

                        Originally posted by twist3r View Post
                        Then I went back to (again) measure voltages across the electrolytic capacitors and I noticed something.
                        Now C610 didn't have 1.545v anymore, but 2.653v. Because its connected to a D602 shottky diode, I check that diode again.
                        That's normal.
                        Diodes do have a small reverse leakage current, which basically means they don't 100% block current in the reverse direction. And that's why the voltage on C610 increased a little over time.

                        ... more replies coming below as I type them.

                        Comment


                          #72
                          Re: Samsung 2693HM - Doesn't turn on (no standby)

                          Originally posted by twist3r View Post
                          Next thing I did was to measure a voltage on L300-L305 and L307 ferrite beads.
                          ...
                          L305 = 2.980v / left pad to GND = 3.293v, right pad to GND = 0.304v ...WTF?
                          Likely this is from leakage current.
                          Remember, ICs are complex devices with many transistors inside. Because you still have some power pins connected to the 3.3V and 1.2V rails, it is possible for a small "leakage" voltage to appear on any disconnected pins, including disconnected power pins. So this isn't something to worry about necessarily.

                          Originally posted by twist3r View Post
                          BUT...what makes me freakout is L305. It has lower voltage between its pads (2.980v). And while left pad to ground had 'corrrect' voltage of 3.293v, what makes me wonder is voltage of right pad to ground, which is 0.304v. Where the F* that voltage came from?
                          Left pad = +3.3V_LIVE, so of course you will see 3.3V to ground.
                          Right pad = power pin internally connected to IC, hence possibility of leakage from other pins... and in yours case, = 0.304V
                          Measuring across the two pads is essentially = left pad - right pad = 3.3 - 0.304 = 2.996V... which is pretty close to the 2.980V you got.
                          So there, mystery solved!

                          Originally posted by twist3r View Post
                          Now, its clear to me that voltage of D602 'right' pad is the same as the voltage on C610 capacitor.
                          What I don't understand is why it kept increasing over time? Also, why voltage across the diode dropped initially from 3.139v to 1.030v when I did 2nd measurement?

                          D602 is part of 18V_IN rail. There is no (currently, because of faulty main/logic board) 18v rail. So, everything about that D602 diode is confusing me. Can you try to help me understand what you think is happening?
                          Again, this is made possible by the small reverse leakage current of D602. Every diode has a small reverse leakage current. Check datasheet. Because this current is so tiny, it doesn't take much for the voltage to change. If you had a bit more sweat on your fingers and happened to touch across C610, that alone could decrease the voltage easily by a few Volts.

                          In fact, if you try putting a very tiny load (like 10-100 KOhm resistor) across C610, that alone would probably keep the voltage down to 0.xV most of the time.

                          Then again, the higher voltage could be due to a malfunction of IC300, triggering the PSU to pulse on very shortly every once in a while.

                          When you have a malfunctioning IC that is responsible for turning on and off the PSU, you also have to factor in that this could be what is upsetting some voltages.

                          Originally posted by twist3r View Post
                          Ok, so... next thing I did was to measure voltages of every pin of those diodes D105-D112 U D121-D128.
                          No point in doing that. (Other than if you're just trying to learn/see how the circuit works.)

                          We only care if these diodes are good or not (i.e. showing shorted or in the rare case, open like that physically damaged diode.) Sp the resistance tests were enough here.

                          Therefore, I'll skip any questions about these, unless you really would like me to answer something that you want to learn about.

                          Originally posted by twist3r View Post
                          Ok, after this, I again check if the IC300 is hot. It was not hot, just maybe a bit more warm after 2 and a half hours of being under voltage.
                          I suppose that confirms that at least the 1.2V_LIVE and 3.3V rails connected to L302, L303, and L304 are not shorted or affected by the damage.

                          Originally posted by twist3r View Post
                          The last thing I did, was to measure a voltage of a BIG FAT cap on the power board.
                          No need.

                          We've already determined the PSU is OK.

                          IC300 is likely the culprit here. So any of the other measurements on other components and boards are just not needed at this point.

                          Comment


                            #73
                            Re: Samsung 2693HM - Doesn't turn on (no standby)

                            Originally posted by twist3r View Post
                            Anyway, main/logic board part numbers are:

                            BN41-00962A (PART NO: BN91-01886A) - 2493HM
                            BN41-00962A (PART NO: BN91-01886B) - 2693HM

                            ^ such 'a small' difference
                            Well, it's the same board, technically, +/- a few components. So kind of makes sense why it has the same part number... except one being type "A" board and the other being type "B".

                            Originally posted by twist3r View Post
                            OK, soooo.... I found a main/logic board for (what I think cheap price). Its a board pulled out from 2493HM monitor. (price was ~ $8.5).
                            Not bad at all!

                            Originally posted by twist3r View Post
                            While desoldering small tiny C609 ceramic capacitor...

                            I lost that dmn thing somewhere.
                            ...
                            Then I desoldered some small caps from some spare boards. Used my ESR meter and check the capacity while it was set to 120 Hz. I found one with 102 nF and use that instead of the one I lost.
                            ...
                            1. The fact that I used spare 102nF instead of 100nF capacitor (which I lost) is no big deal, correct?
                            Yes, that is fine.
                            You can expect at least 5% tolerance on most power components in circuit. For ceramic caps, even 10-20% from original value is not too far off.
                            And that cap really is just for filtering any possible high frequency noise.
                            Board will probably run OK without it in most circumstances. Or if you used up to 1 uF ceramic, that would be OK too. So don't sweat it - the value of these caps is rarely that important.

                            Originally posted by twist3r View Post
                            Anyway, I pressed the monitor power button and blue led turned off, and voltages on POWER_EN pin of power board dropped to zero. Again, I pressed the power on button on monitor's frame and led light show up shinning blue, and voltages are back to 3.214v on power board POWER_EN pin.
                            Normal operation there.

                            POWER_EN = high ---> monitor is OFF/standby
                            POWER_EN = low ---> monitor is ON

                            Originally posted by twist3r View Post
                            The only thing that confused me was that the POWER_EN pin on the power board had 3.214v (which I think is something like 3.3v).
                            The precise voltage of POWER_EN is not that important.

                            Generally, for HIGH/LOW signal lines like that, there is a threshold that determines if a signal is considered "LOW" or "HIGH". And there is also a "deadzone" / undetermined /transition area that is ignored.

                            So for example of a 3.3V signal line (like your POWER_EN signal seems to be), probably something like 0 to 0.8V is considered low, and 2.7 to 3.3V is considered high. The "deadzone" area between 0.8V and 2.7V may be ignored and the monitor may not change states until it goes either below the maximum LOW voltage threshold or above the minimum HIGH voltage threshold. That said, I'm using the values of 0.8V and 2.7V just as an example here. They may be something different. It should be stated in the datasheet of the supervisor IC used on the PSU board.

                            But again, it looks like your monitor's PSU is turning On and Off properly, so no need to bother investigating this any further.

                            Originally posted by twist3r View Post
                            So... if my logic doesn't tricks me... I would say that expected voltage for POWER_EN pin is ~3.3v and not 5v as I was expecting. Correct?
                            Yes, correct.

                            I incorrectly assumed POWER_EN is 5V, simply because +5V_MICOM going to the logic board is 5V. It doesn't really matter too much, though. If you "dump" 5V into POWER_EN with a high-resistance resistor from +5V_MICOM, the PSU board will still turn on, because even that will satisfy the HIGH/LOW logic requirements for the signaling.

                            Originally posted by twist3r View Post
                            Anyway, I'm trying to allow myself to think that the new main/logic board should work 'ok'. But I got some questions before I proceed:
                            ...
                            2. I want to dissolder firmware chip from my 2693HM main board and solder it to 2493 main board.
                            Don't.
                            Not yet, anyways.

                            First try the 2493HM board "as-is" in your monitor.
                            In regards to the components you already moved (C610, D602, and etc.) - that's fine, you can leave those.

                            But since your monitor seems to respond normally to On-Off commands from the 2493HM board, just put that in there, assemble the monitor enough that you can test it, and see what happens.

                            Again, the WORST that could happen is that the backlight controls/brightness *might* be slightly off or not working due to some small differences in the backlight control circuits. But I'm pretty sure Samsung designed this so that nothing would fry or break by using the 2493HM board directly as is on your monitor.

                            So try it and see what happens. If it works, I say DON'T mess with desoldering firmware chips. In fact, the firmware chip from your 2693HM logic board may try to "ask" the 2493HM board to do something that it isn't capable of, and that IMO is more likely to introduce "bugs" and incompatibility than straight-up running the 2493HM board with the 2493HM firmware.

                            In other words, do the firmware swap only if the backlight controls don't appear to work normally (or you notice some other bug.)

                            Originally posted by twist3r View Post
                            a) whats the easiest way to desolder it (and not damage all 8 legs) with an soldering iron?
                            Got a second soldering iron (25-40W)? If yes, you can use one iron to melt a line of solder on one side of the chip and the other iron to melt a line on the other side. Then you touch both side simultaneously and move/slide the firmware chip away from its position to hopefully an area that is clear of SMDs and other components.

                            If that's not possible or you have only 1 soldering iron, then melt bigger pools of solder on both sides of the firmware chip and use a big tip to melt solder on both sides simultaneously... then grab the chip with thin tweezers and lift it away when the solder is melted. <--- I usually use this method.

                            Originally posted by twist3r View Post
                            b) I thought that there is a one firmware chip but now I'm confused because there are 2 and I don't know which chip is the one that I should desolder.

                            IC303 - MX25L4005 (this is the one I thought to desolder, and service manual list it as: IC-FLASH MEMORY; MX25L4005, 4Mbit, 512Kx8Bi - SM list it as IC115)

                            OR

                            IC302 - S-24CS08AFJ-TB-1GE (this one got me thinking because service manual list is as: IC-EEPROM; S-24CS08AFJ-TB-1GE, 8Kbit, 1Kx8 - SM list is as IC112)
                            I'm thinking it's the MX25L4005, simply because that's a 4-Mbit chip, which sounds more inline with what would be the "main" firmware/BIOS chip. The other one, I'm not sure what it could be.

                            When it comes to BIOS/flashing/firmware, that's where you're reaching the limits of my knowledge. I prefer to work with analog / power circuits.

                            Comment


                              #74
                              Re: Samsung 2693HM - Doesn't turn on (no standby)

                              Hey momaka. Thanks again for another great reply!

                              I'm sorry that I made your PM box full. I Apologize for doing that.


                              Originally posted by momaka View Post
                              Originally posted by twist3r View Post
                              Thing that also made me wonder if I did solder it back proprely is the fact that such a small capacitor also cover its pads, so when I try to solder it, I can't put 'solder under it' but 'outside' at the edge. Its because I'm not using a hot air station, but a soldering iron.
                              That's fine. As the solder on the side of the cap melts, it will also melt the solder on the bottom of the cap. So it should be OK. The key is to touch both the cap lead AND the SMD pad to which the cap is supposed to be soldered at the same time (one one side, that is.) Once you do that, your SMD component should only be soldered on one side. To make sure that it is soldered properly, try moving it. If it can't be moved and seems soldered to the board, then it's good. After this, you can then solder the cap on the other side.
                              Thanks for the tips. I tried first to solder the cap while first putting solder on both pads. But... once I soldr one side, the other side of the capacitor would be 'up in the sky'. And when I melt the solder on that side, it wont help because the first page that I solder the capacitor on, was firm in holding that side and not allowing opposite side to 'lay' on the pad.
                              So then, I useld solder wick to completly remove the solder from one side, and first solder one side of the capacitor. That was 'better' approach, but because of the size of the capacitor, it wasn't easy at all. Capacitor kept being stickied to my tweezers so everytime I tried to hold it in place it ended up being on the tweezers.

                              Can you check the image of the main/logic board that I attached, and view the area in top-left corner of the board. You'll see a pads where C609 ceramic capacitor should be placed (its not my board, its from the shopppjimmy). Can you try to guess whats the package size of that capacitor that I soldered? I just want to know what is my soldering 'achievement'

                              Before soldering that capacitor I desoldered couple other of simmilar size and put it in the small box. I showed the box (with ceramic capacitors in) to my father and told him 'check the size of the capacitor that I finally soldered on the board'. He checked the inside of the box and said something like 'what capacitor, where?'. I was like 'exactly' After that he asked 'this? those are capacitors? They look like dirt or dust'

                              It wasn't easy for me to solder that small capacitor back at all. Was looking for too long through magnifying glass and my eyes started to complain long after I finished it. I guess I would have to try to make a DIY microscope from a webcam or something like that.



                              Originally posted by momaka View Post
                              Originally posted by twist3r View Post
                              If I had soldering station I guess this would be much easier, but...
                              If you have a good hot air station... yes.
                              But with the cheaper ones... no.
                              I have a cheaper one and I *avoid* using it as much as possible for SMD components, unless I really have to. I find it easier to do SMDs with a soldering iron - especially stuff with only 2 leads.
                              Well, 'good soldering hot air station' is not gonna happen here anytime soon Maybe a ~ $60-$80 cheaper one but it seams thats not gonna be good SMD.

                              I guess you are used to, after years of practice, to preffer soldering iron over hot air station. Maybe I'll think the same way, once I have hot air station, so I can do comparison between hot air and soldering iron.




                              Originally posted by momaka View Post
                              With small ceramic caps in parallel, it only takes ONE to short out and make the rest look like they are all shorted. But you appear to have multiple pins of IC300 show low resistance to ground @ L300-L307. So I think more likely IC300 is shot.
                              I hoped that the faulty element could be anything else but BIG IC300 chip. But I guess, in this case, as you said, its probably IC300.

                              I found that chip on aliexpress. Its not 'too expensive'...like $4-$6.... BUT it comes with ~$75-$80 shipping cost to my country, so thats that. End of the line for possible future 'fix' of my old main/logic board.

                              btw. If we imagine for a second that I find the same IC300 chip, and I want to solder it on my board instead of the faulty one.... I'll use hot air station but... (two questions here):

                              1. Do I need to pre-heat it (under the board)?
                              2. Do I need a really good hot air station (the one that reach 480-500 C degrees), or I can use cheaper ones that reach ~300-350 C degrees? I'm just asking so I can have some info of what kind of hot air station is required?


                              One offtopic qustion here (two exactly), sorry about that:

                              Imagine you have a faulty PC GPU card with faulty memory chip on it. Or you have 2 faulty RAM sticks (modules) for computer motherboard....

                              1. Do I need expensive hot arii station to change memmory chips or cheaper one can do it (without, ofcorse damaging the chips)?

                              2. If one chip on the RAM stick (module) is faulty, will the RAM module be 'fixed' when I solder the good chip from another ram module (of the same tipe), or... these memmory modules has some 'control' IC that wont allow 'alien' chips to be used on a RAM module? I don't know if I explaind this good. I'm asking if the RAM module needs to be 'programmed' (in a way) to accept new 'brother' (new memmory chip) among other memmory chips which were already on the RAM module?



                              Originally posted by momaka View Post
                              The resistances are just a collection of values I've gathered "in the field" when doing monitor repairs. Like for example, if you try to measure the resistance of the GPU on a PC graphics card, you will notice the resistance is typically very low - often 1.5 to 4 Ohms (though could be higher too.) CPUs and RAM, on the other hand, tend to have slightly higher resistances... though not always. And in the case of low-power ICs like IC300 in your monitor, I -expect- the resistance to be somewhere in the 100's of Ohms or more. That doesn't mean that it will be. But generally, it is.

                              So when I saw the low 1.x Ohms resistance on 3.3V_LIVE, I knew that had to be a red flag, because a relatively "small" IC like IC300 generally should NOT exhibit that low of a resistance. Thus, it seemed more like something was shorted. And by removing L300-L307 and the 3.3V_LIVE returning back to normal working order, you confirmed that something is likely bad inside of IC300.
                              Ok, I (think) I got it. Thanks! May I conclude from this that everything that measures under 1.0 Ohm is 'short'?



                              Originally posted by momaka View Post
                              And remember, GPUs, CPUs, RAM, and all other ICs are active devices - meaning the resistance you measure with a multimeter is only static when the monitor is off. When power is applied to the monitor, the resistance can change dramatically. Of course, you can't measure it then, because you CANNOT use a multimeter to measure resistance when there is power in the circuit.
                              Ok, so the resistance is always checked while the board is turned OFF. But I want to ask you about the ^ bold part above. When you say 'CANNOT' does that mean 'I wont be getting correct resistance values if I try to check resistance while the board is turned ON?. ...or I 'CANNOT' use multimeter ot measure resistance because its DANGEROUS?



                              Originally posted by momaka View Post
                              Originally posted by twist3r View Post
                              first value is the voltage on the diode while one probe is touching one side of the diode, and the other probe is touching other side of the diode.
                              That usually doesn't give any useful information when checking voltages, so in the future, don't do it unless you're asked to.

                              The only thing this is useful for is to determine forward voltage drop of a diode in a DC circuit. Thus, you can sometimes use it to see if a diode has failed open-circuit. But that is extremely rare. Diodes -usually- fail short-circuit.
                              Ok, I'll try to remmember that. Just one question here: When you say that diodes (usually) fail short-circuit, does that mean that faulty diodes will most of the time BRAKE (cut off) the circuit? And that faulty diode wont allow anymore a current to pass through in both directions if its a faulty one? Is that the definition of a faulty diode (most of the time)? Did I understood this correctly?
                              Attached Files

                              Comment


                                #75
                                Re: Samsung 2693HM - Doesn't turn on (no standby)

                                Originally posted by momaka View Post
                                Originally posted by twist3r View Post
                                Next thing I did was to measure a voltage on L300-L305 and L307 ferrite beads.
                                ...
                                L305 = 2.980v / left pad to GND = 3.293v, right pad to GND = 0.304v ...WTF?
                                Likely this is from leakage current.
                                Remember, ICs are complex devices with many transistors inside. Because you still have some power pins connected to the 3.3V and 1.2V rails, it is possible for a small "leakage" voltage to appear on any disconnected pins, including disconnected power pins. So this isn't something to worry about necessarily.
                                Originally posted by momaka View Post
                                Originally posted by twist3r View Post
                                BUT...what makes me freakout is L305. It has lower voltage between its pads (2.980v). And while left pad to ground had 'corrrect' voltage of 3.293v, what makes me wonder is voltage of right pad to ground, which is 0.304v. Where the F* that voltage came from?
                                Left pad = +3.3V_LIVE, so of course you will see 3.3V to ground.
                                Right pad = power pin internally connected to IC, hence possibility of leakage from other pins... and in yours case, = 0.304V
                                Measuring across the two pads is essentially = left pad - right pad = 3.3 - 0.304 = 2.996V... which is pretty close to the 2.980V you got.
                                So there, mystery solved!

                                ^ sorry for too many quoted-quotes. I think its 'good' if anyone else decide to peak inside this topic and read this epic conversation, he wont get lost in here, hopefully.

                                Ok, I have to say that, since I got no knowledge, somtimes (most of the time) I got to ask stuped quetions. Sometimes I don't realize how stuped they may be, and sometimes, I ask them anyway, even when I know that the question sound stuped. Simply because I'm trying to 'learn' things and gets to their 'core'. Something that I would 'use it' as a constant. Something that, no matter what, wont change some fact that I learned before. Thats the reason why I keep asking simmilar questions. Because I want to 'feel' how solid the facts are around specific subjects, so I can say 'ok, I learned this and lets move on'.

                                So...
                                we measure resistance to GROUND when the board is turned off?
                                and
                                we measure voltage to GROUND when the board is turned ON?

                                The only time we measure voltage on some element by putting multimeter probes on both sides of that element is when we want to see if there is a voltage drop? Correct?



                                Originally posted by momaka View Post
                                Every diode has a small reverse leakage current. Check datasheet. Because this current is so tiny, it doesn't take much for the voltage to change. If you had a bit more sweat on your fingers and happened to touch across C610, that alone could decrease the voltage easily by a few Volts.

                                In fact, if you try putting a very tiny load (like 10-100 KOhm resistor) across C610, that alone would probably keep the voltage down to 0.xV most of the time.
                                This thing, the relationship between voltage, current and resistance confusing me for my entire life.
                                And the reason for that is that freaking formula:

                                V = I * R

                                So... everytime I think about it, its like "ok, if we increase the resistance, then the resulting voltage can't be lower, but higher"

                                But...now when you say that that if I had sweat fingers and touched the capacitor, that would decrease the voltage. What sweat (or water) does to a resistance? Lower it or increase it? My mind can't decide. Once I think about it, I would say water (sweat) decrese the voltage. But next time I think about it...again I get confused and not sure if it actually increase the voltage.

                                When you said that if we put even tiny load (10-100 KOhm reistor) that would probably keep voltage to zero on the capacitor. First I thought 'why would we put (even) the small load, small resistor'. That would increase the voltage. But then I thought... we need some 'load' (some resistance) if we want 'voltage' to 'happen' across some element. Maybe thats the reason you mantioned 'small load'. Then again... if we put higher restor (bigger load), that must increase voltage? Right? (because of V=I*R)

                                Also... (maybe there is no connection here but) I have to ask: what about electrolytic capacitor's ESR? If its internal resistance (which increase over time) is high, will the voltage on that capacitor be lower or higher?




                                Originally posted by momaka View Post
                                We only care if these diodes are good or not (i.e. showing shorted or in the rare case, open like that physically damaged diode.) Sp the resistance tests were enough here.
                                Again diodes ...
                                I was thinking why I'm getting stuck with some things. Probably big reason might be the english language by itself, because my brain translate english to my language, and... that not always goes 100% correct.

                                I already mantioned (commented and ask if I'm correct) about diodes at the begining of my reply. Now I have to ask again. (I'm really sorry about me being an idiot here, who can't understand this)

                                What I thoguht I learned, is that diodes allow current to flow in on direction and not in the opposite direction. (Yes you said its always a leakage current, which is...(I guess) small, so not important).
                                So when the diode fails, most of the time what that failure does to that 'line'? Does it brake the line so the current cant' flow in any direction, or it short the line, so current flow (like I guess happen when 230v went through my monitor's HDMI port)? Then again, there were 'protection' diodes, so... I really don't know if they failed by lettiner higher current flow, which means 'line' was not broken, ...OR they let the current flow and then failed by braking the line. I'm on the edge of misery here.



                                Originally posted by momaka View Post
                                Originally posted by twist3r View Post
                                The last thing I did, was to measure a voltage of a BIG FAT cap on the power board.
                                No need.

                                We've already determined the PSU is OK.
                                Ok, one 'quick' question here: How you guys always calculate what voltage will be measured across that BIG FAT filter capacitor? What is the calculation formula?

                                If 230v AC comes at primary tranformer side, what voltage comes out as DC at its secondary side after bridge rectifier? I was trying to guess that 230v AC was multiplied by 1.41 (square root (of 2), I don't know how it is properly said in english). That BIG FAT capacitor is AFTER bridge rectifier, right (and NOT between secondary tranformer side and bridge rectifier)? How you always correctly guess ~ the voltage on that capacitor? I would like to know.



                                Originally posted by momaka View Post
                                Originally posted by twist3r View Post
                                OK, soooo.... I found a main/logic board for (what I think cheap price). Its a board pulled out from 2493HM monitor. (price was ~ $8.5).
                                Not bad at all!
                                Well... I would guess thats a cheap price from your point of view, and even from my own. But checking some sites where sells spare parts (spare PSU or LOGIC boards), like the one I mantinoed few times here (where I got the images of both 2493HM and 2693HM main boards), price range is so high that it kills the whole idea of 'fixing' any monitor. Because... if the used (but working) monitor cost... lets say 50 bucks, whats the point of buying a board that cost $45 to fix a faulty monitor? And sometimes the cost of the board itself is higher than the price of used but working monitor.
                                So yes, I got it for cheap, but I think its only because I was fixing my 26" monitor, which has a higher value compared to 24" (or 22") monitors. But if I was for fixing 22" monitor that price would be on the 'line' where I would have to decide if the price would kill the whole purpose of fixing it because used 22" here can be bought for 25-30 bucks)

                                Also, another reason for me trying to fix my monitor (and it the future other ones) is to learn things while doing it, even if the end price might be against common logic.




                                Originally posted by momaka View Post
                                You can expect at least 5% tolerance on most power components in circuit. For ceramic caps, even 10-20% from original value is not too far off.
                                And that cap really is just for filtering any possible high frequency noise.
                                Board will probably run OK without it in most circumstances. Or if you used up to 1 uF ceramic, that would be OK too. So don't sweat it - the value of these caps is rarely that important.
                                Ok, got it. One offtopic question though:
                                Back in 2017 when I had to change capacitors on my monitor's power board, I read on couple different forums (probably also on badcaps) a suggestion that when you solder new electrolytic capacitor, its a good idea to solder on its legs (on the back of the board) a small ceramic capacitor to extend the life of Electrolytic capacitor. I guess they didn't meant an SMD ceramic cap, but the one with legs (for though-hole). So... do you agree with that suggestion? (of soldering ceramic cap paralell, on the legs of electrolytic one for life extension)






                                Originally posted by momaka View Post
                                I incorrectly assumed POWER_EN is 5V, simply because +5V_MICOM going to the logic board is 5V. It doesn't really matter too much, though. If you "dump" 5V into POWER_EN with a high-resistance resistor from +5V_MICOM, the PSU board will still turn on, because even that will satisfy the HIGH/LOW logic requirements for the signaling.
                                Ok the PSU board will turn on, and...what would happen next? I guess nothing, because.... there want be a picture since the main board (brain IC300 chip) is dead. Correct?

                                Comment


                                  #76
                                  Re: Samsung 2693HM - Doesn't turn on (no standby)

                                  Ok, and finally the interesting part (I know I bored you complitly with my questions ^ above. Again, sorry about that, I really had to ask them)


                                  Originally posted by momaka View Post
                                  First try the 2493HM board "as-is" in your monitor.
                                  In regards to the components you already moved (C610, D602, and etc.) - that's fine, you can leave those.

                                  But since your monitor seems to respond normally to On-Off commands from the 2493HM board, just put that in there, assemble the monitor enough that you can test it, and see what happens.

                                  Again, the WORST that could happen is that the backlight controls/brightness *might* be slightly off or not working due to some small differences in the backlight control circuits. But I'm pretty sure Samsung designed this so that nothing would fry or break by using the 2493HM board directly as is on your monitor.

                                  So try it and see what happens. If it works, I say DON'T mess with desoldering firmware chips. In fact, the firmware chip from your 2693HM logic board may try to "ask" the 2493HM board to do something that it isn't capable of, and that IMO is more likely to introduce "bugs" and incompatibility than straight-up running the 2493HM board with the 2493HM firmware.

                                  In other words, do the firmware swap only if the backlight controls don't appear to work normally (or you notice some other bug.)
                                  You guessed most of it (if not even everything) correct.

                                  Ok, I didn't changed the firmware chips (one of the reason being for not being 100% sure which one is the correct one, or should I switch not one but two chips from one board to another.

                                  DVI and HDMI ports work (didn't tested VGA port, but there should be no reason not to work).

                                  BUT...
                                  Once I was in my desktop, brightness apears to 'breath'. Like... I don't know if I'll explain this well...
                                  It was like trying to ...'adjust' itself. Like... switching to brighter, then go darker, then brighter, then more brither, then back to darker, then again brigth, brighter, darker.... etc.

                                  Now I'm not sure.... but after some time that calmed down. Still, because at the time I was aware of what was happening, I took a closer look by observing it and...I kinda (think) to see tiny fluctuation in brightness.

                                  Why I'm not sure? Its because I tried to see WTF is wrong by using OSD menu. And yes, you were right about brightness. It was set to 100%, but changing it all the way to 0% did nothing. It kept being very bright. Thats 'good', BUT...back in 2017 when I switched capacitor on power board, I decided to turn down the brightness (and contrast) to, I think, ~ 35%-40% because that kept my monitor from not being very hot at the back (plastic cover). But now, while I was observing my 26" monitor for probably 30 minutes, back cover of the monitor was pretty hot (somewhere in the area where power board is). And... I could smell...the 'electronic smell'. I don't know how else to express myself about that smell.
                                  Contrast did changed while I was changing it over OSD menu, but...(maybe I'm wrong) I think the change was not as it was in the past (while old main board was still working).

                                  One thing for sure is: changing brightness by using OSD menu, does nothing. And I need to change it (to lower it) so that I can extend the life of my CCFL lamps. It was noticable if you pay closer look at the right vertical edge ofo monitor screen that there is a 'darker' fat (of a thumb size) or better call it 'strip' across that edge. Ofcorse that can be seen only because I know the flaws of my monitor. And I know where to look. If you observe the monitor as a whole, that right edge of the screen is not that noticable.

                                  So...I turned off my monitor and let it wait for some better time until I buy hot air station. Then I could try to switch firmware chips and try to see if brightness could be lowered from the OSD menu. I just have to determ how many (and which) chips are firmware chips.

                                  One extra thing that I notice while being in OSD... I don't remember if I saw that option on my old main board, but...there is a setting for something called 'HDMI dark level' or something like that. That was there, but 'not available' for changing. I don't know what that is for.







                                  Originally posted by momaka View Post
                                  Originally posted by twist3r View Post
                                  whats the easiest way to desolder it (and not damage all 8 legs) with an soldering iron?
                                  Got a second soldering iron (25-40W)? If yes, you can use one iron to melt a line of solder on one side of the chip and the other iron to melt a line on the other side. Then you touch both side simultaneously and move/slide the firmware chip away from its position to hopefully an area that is clear of SMDs and other components.

                                  If that's not possible or you have only 1 soldering iron, then melt bigger pools of solder on both sides of the firmware chip and use a big tip to melt solder on both sides simultaneously... then grab the chip with thin tweezers and lift it away when the solder is melted. <--- I usually use this method.
                                  Well, I don't have 2nd soldering iron so that wont work. As for the 2nd method, that might work if I have a 'big tip' which (again) I don't.
                                  I was desoldering (maybe not of that size) chips with 8 legs before, but...you know, its different when you desolder out of 'fun' and don't care if you brake things, and you still managed not ot brake it most of the time, and the other thing is when you are in fear not to fck up things, like I could.





                                  Originally posted by momaka View Post
                                  Originally posted by twist3r View Post
                                  I thought that there is a one firmware chip but now I'm confused because there are 2 and I don't know which chip is the one that I should desolder.

                                  IC303 - MX25L4005 (this is the one I thought to desolder, and service manual list it as: IC-FLASH MEMORY; MX25L4005, 4Mbit, 512Kx8Bi - SM list it as IC115)
                                  OR
                                  IC302 - S-24CS08AFJ-TB-1GE (this one got me thinking because service manual list is as: IC-EEPROM; S-24CS08AFJ-TB-1GE, 8Kbit, 1Kx8 - SM list is as IC112)
                                  I'm thinking it's the MX25L4005, simply because that's a 4-Mbit chip, which sounds more inline with what would be the "main" firmware/BIOS chip. The other one, I'm not sure what it could be
                                  Thats what I thought at fist as well. That MX25L4005 is the 'one'. BUT I think I saw in some russian site among firmware dumps for my monitor that there were two files in the package, so that made me thinking that there are 2 chips which needs to be programmed in order everything to work correct.





                                  Originally posted by momaka View Post
                                  When it comes to BIOS/flashing/firmware, that's where you're reaching the limits of my knowledge. I prefer to work with analog / power circuits.

                                  At least you know in which area you have a knowledge. It's much better than knowing that you are not good at all for anything about electronics (like in my case) and I still keep trying to figureout things while bothering a lot of people with my questions.


                                  Here is one more:

                                  You got 2 faulty monitors. One uses CCFL and the other uses LED. You don't know what is wrong with them. But by knowing about all sorts of things that can be wrong with them, and picking the hardest one, which monitor would have greater chance for a fix?
                                  I'm asking this becaseuse... you see, I changed caps on my monitor power board back in 2017. Now I changed main/logic board (and that still need to get its firmware chips back). BUT as you read above... its CCFL lamps are getting close to be 'done'. Lamps are for 25.5" monitor. Today its hard to find such lamps (and they are not cheap enough to make you want to fix such monitor). Its an OLD technology. CCFL lamps easily brake in transport. So I was thinking whats easier to fix, LED monitor 'led failure' or CCFL? Maybe you can just change a LED diode on 'led strip' iside led mnonitor and its 'good to go' compared to faulty CCFL lamps (especially for a bigger monitor like this, compared to 22" one). You know.... you trying to fix a moitor which didn't show any picture, and finally once you tought you fixed it, you see that lamps are already gone, so it leave tht 'bitter' taste, if you know what I mean. (its not a question regarding my monitor, but 'in general' what you think, whats more 'gratefull' type of monitor that deserve to be fixed, CCFL or LED, since CCFL is hard to come by for larger monitors, or monitors of that type).
                                  Last edited by twist3r; 01-16-2021, 07:09 AM.

                                  Comment


                                    #77
                                    Re: Samsung 2693HM - Doesn't turn on (no standby)

                                    Originally posted by twist3r View Post
                                    Ok, and finally the interesting part
                                    ...
                                    You guessed most of it (if not even everything) correct.

                                    Ok, I didn't changed the firmware chips (one of the reason being for not being 100% sure which one is the correct one, or should I switch not one but two chips from one board to another.

                                    DVI and HDMI ports work (didn't tested VGA port, but there should be no reason not to work).

                                    BUT...
                                    Once I was in my desktop, brightness apears to 'breath'. Like... I don't know if I'll explain this well...
                                    It was like trying to ...'adjust' itself. Like... switching to brighter, then go darker, then brighter, then more brither, then back to darker, then again brigth, brighter, darker.... etc.
                                    Interesting.

                                    Yeah, looks like the firmware on the monitor programs how PWM0, PWM1, and PWM3 signals from IC300 behave.

                                    If you open up the SM and go to page 81 (last page), take a look in the top-left corner at the circuit diagram for CN601.

                                    I think it's written in a very VERY confusing way (or maybe it's just mislabeled?) But if I'm understanding it right, at least from what I'm given...
                                    For 2693HM:
                                    PWM3 (signal from IC300) ---> A_DIM (Analog DIM) ---> pin 4
                                    PWM0 (signal from IC300) ---> BRT_CTL (PWM Method) ---> pin 3 (?)
                                    PWM1 (signal from IC300) ---> PWM_DIM (PWM_DIM) ---> pin 2 (?)

                                    And based on the table to the right of that...
                                    Status on 2493HM is outputting where Analog_DIM is on 2693HM.
                                    Analog_DIM on 2493HM is outputting where PWM_DIM is on 2693HM.
                                    PWM_DIM on 2493HM is outputting where PWM Method is on 2693HM.

                                    So I think that ^ above might be why you can't control the brightness.

                                    Now, *IF* the above information I posted is correct, I *think* it may be possible to just re-wire the inverter control on the 2493HM board, should you'd prefer to do that instead of swapping the firmware (if you think that would be easier / less likely to break something.)

                                    What I am thinking *might* work...
                                    First, take the wire harness that goes between CNM801 and CN601 - that's the wire harness for the inverter board control (i.e. the smaller cable "bunch" shown at the top between the boards in this picture.) On the side where it plugs into CN601, see if you can remove the wires from the plastic part of the connector. Typically, that can be done by pushing on the metal pin bit in the cutout in the plastic. We need to move the wires that go to pins 2, 3, and 4 only, so remove those wires from the plastic connector. (If you can't, post a picture of all sides of the plastic of the connector.)

                                    Next, plug the wires as follows:
                                    Move wire #2 to pin position 3 on the plastic connector.
                                    Move wire #4 to pin position 2 on the plastic connector.
                                    Move wire #3 to pin position 4 on the plastic connector.

                                    This modification should not damage anything on the inverter board or the logic board, because according to the service manual, pins 2, 3, 4, and 5 are all output pins, and they all operate between 0V and 5V. So bad comes to worse, the backlight brightness will either be stuck at max (100%) again or min (0%).

                                    With that said, if the above modification does not work, disconnect wire #3 (which you modified to plug into pin position 4) and see if that helps. If not, you might have to bust out your multimeter again and take some voltage measurements on CN601 with the monitor On and @ 100% backlight brightness in OSD, then @ 0% backlight brightness in OST. (Hint: for this test, it might be helpful to solder long wires to the pins on CN601 so that you can measure them when the monitors is closed up so that you can access the OSD menu.)

                                    Anyways... again, the way the SM is written in regards to the backlight control is absolutely stupid. I wouldn't be surprised if there is a mistake in there. So when I was doing the above suggested modifications, I was referring to the pictures you provided of your old 2693HM board to make sure we don't make a wiring mistake somewhere. On that note, if you could provide a picture of your new 2493HM logic board, that would be great so we can compare.

                                    But all in all, I think the above experiment should be safe to try with the wire swapping @ CN601. At least that might get your monitor working again without having to swap firmwares.

                                    As for your other questions, I'll answer them in another post (or rather more posts) after this one... and may not be able to get to all of them today (it's almost 2 AM here.)
                                    Last edited by momaka; 01-22-2021, 12:39 AM.

                                    Comment


                                      #78
                                      Re: Samsung 2693HM - Doesn't turn on (no standby)

                                      Originally posted by twist3r View Post
                                      Thanks for the tips. I tried first to solder the cap while first putting solder on both pads. But... once I soldr one side, the other side of the capacitor would be 'up in the sky'.
                                      ...
                                      So then, I useld solder wick to completly remove the solder from one side, and first solder one side of the capacitor. That was 'better' approach, but because of the size of the capacitor, it wasn't easy at all. Capacitor kept being stickied to my tweezers so everytime I tried to hold it in place it ended up being on the tweezers.
                                      If your SMD component (ceramic cap, in this case) gets stuck to the tweezers, then it's not getting soldered to the board.

                                      Clearing the solder on one side with wick can help, but is not necessary.

                                      What I do is I use a sharp chisel or bevel or spoon tip and make sure it has no solder or almost no solder on it, then run it with flux over the pads that have solder on them and collect a little bit of solder from them (but still leave a tiny layer on there.)

                                      After that, I put more flux on both pads again. If there is any burned, hardened flux on the pads from before, make sure to scrape it off before applying more flux. Old, hardened flux on the pads may not allow solder to stick to it. Once the pads are fluxed, place the SMD component and hold it in place (be it with tweezers or finger nail.) Then, using a sharp chisel or bevel tip again (same as the one you cleaned the pads with), heat one side of the SMD component to solder it. Make sure you touch both the PCB pads and the component leads, as this will melt the solder on both and join them instantly. For those very small "sandgrain" caps, they should literally solder in just a second or two when you touch them with the iron's tip.

                                      SMD work just takes a lot of practice and a bit of steady hands - that's all.

                                      Originally posted by twist3r View Post
                                      Can you try to guess whats the package size of that capacitor that I soldered? I just want to know what is my soldering 'achievement'
                                      Hard to say, but either 2012, 1608, or maybe 1005 (all metric sizes, BTW.)
                                      My guess is 1608 metric (that is, a component with length of 1.6 mm and width of 0.8 mm... hence 1608 metric.) But you can measure it with a small ruler and see for yourself.

                                      Either way... good job!

                                      Those small SMDs are very tricky to solder for anyone who's not done SMD before.

                                      Originally posted by twist3r View Post
                                      Before soldering that capacitor I desoldered couple other of simmilar size and put it in the small box. I showed the box (with ceramic capacitors in) to my father and told him 'check the size of the capacitor that I finally soldered on the board'. He checked the inside of the box and said something like 'what capacitor, where?'. I was like 'exactly' After that he asked 'this? those are capacitors? They look like dirt or dust'


                                      Yeah, isn't it crazy how small some of these components are?
                                      It's even worse on phones and tables! That's why I don't work on those.

                                      Originally posted by twist3r View Post
                                      I guess you are used to, after years of practice, to preffer soldering iron over hot air station.
                                      Well, it depends for what.
                                      But for 2-leaded SMD components, I just take a FAT soldering tip and scoop up the SMD in just a second or two - way faster than a crappy hot air station and probably just as fast (if not faster) than a good hot air station.

                                      I even have my techniques down for removing SMD MOSFETs faster with TWO soldering irons than most people can with hot air.

                                      So yeah, I reserve hot air only for large multi-pin ICs that I just can't do with a soldering iron (or two.)

                                      Originally posted by twist3r View Post
                                      I found that chip on aliexpress. Its not 'too expensive'...like $4-$6.... BUT it comes with ~$75-$80 shipping cost to my country, so thats that. End of the line for possible future 'fix' of my old main/logic board.
                                      That's insane!
                                      Why is the shipping to your country so much?

                                      Maybe if you have a relative living somewhere in EU zone, have the chip ship to them, and then have them repack the ship to you in a small postal mail card or something similar. I actually sent a chip this way not too long ago to a Serbian compatriot of yours, and he received it fine without issue. Cost me like $2 or 3 total in mail stamps, I don't remember exactly... but not much, really.

                                      Originally posted by twist3r View Post
                                      btw. If we imagine for a second that I find the same IC300 chip, and I want to solder it on my board instead of the faulty one.... I'll use hot air station but... (two questions here):

                                      1. Do I need to pre-heat it (under the board)?
                                      2. Do I need a really good hot air station (the one that reach 480-500 C degrees), or I can use cheaper ones that reach ~300-350 C degrees? I'm just asking so I can have some info of what kind of hot air station is required?
                                      If you have a good hot air station, you *may* not need a bottom heater... though I usually advise that one be used, as it makes the removal easier and less likely to damage the PCB.
                                      If you have a cheap hot air station, you should definitely use a bottom heater - it will make the job A LOT easier and quicker.

                                      Originally posted by twist3r View Post
                                      Imagine you have a faulty PC GPU card with faulty memory chip on it. Or you have 2 faulty RAM sticks (modules) for computer motherboard....

                                      1. Do I need expensive hot arii station to change memmory chips or cheaper one can do it (without, ofcorse damaging the chips)?
                                      RAM sticks you can do with a cheap hot air with no bottom heater because the size of the PCB is small.

                                      With desktop PC video cards, you may or may not be able to do them just with a cheap hot air. That said, I have a cheapo station, and I have removed RAM and GPU chips even from bigger video cards. But I used a bottom heater (gas stove and on one occasion, also electric stove), which made this a lot easier.

                                      Originally posted by twist3r View Post
                                      2. If one chip on the RAM stick (module) is faulty, will the RAM module be 'fixed' when I solder the good chip from another ram module (of the same tipe), or... these memmory modules has some 'control' IC that wont allow 'alien' chips to be used on a RAM module? I don't know if I explaind this good. I'm asking if the RAM module needs to be 'programmed' (in a way) to accept new 'brother' (new memmory chip) among other memmory chips which were already on the RAM module?
                                      If you're doing memory chip swaps, the new chip must have the same exact part number as the old chip. That means, not only the same memory size, but also same speed and timings. The last 2 digits in the part number typically indicate the chip speed. For example, for a Samsung memory chip with part # K4J52324QC-BC14, the "14" part indicates a 1.4 ns chip. This corresponds to (1 / 1.4 ns ) x 1000 = ~714 MHz. The same chip with "BC16" at the end would be a 1.6 ns chip and clock at ~625 MHz max.

                                      But apart from that, no these chips are not programmed inside, so you can swap them as you please.

                                      Originally posted by twist3r View Post
                                      Ok, I (think) I got it. Thanks! May I conclude from this that everything that measures under 1.0 Ohm is 'short'?
                                      Pretty much yes.

                                      I usually define "short-circuit" as the lowest resistance my multimeter can measure (0.2 to 0.3 Ohms), so 1 Ohm isn't necessarily a short-circuit for my multimeter... though if I measured your monitor with my multimeter, I would have still concluded the same as with your multimeter, because such low resistance on the +3.3V_LIVE of your monitor simply does not make sense. Again, if this was a PC video card and you were measuring the GPU... that might be borderline passable / OK. But a monitor's low power chip, I'd expect to read at least several 10's or 100's of Ohms.

                                      Originally posted by twist3r View Post
                                      Ok, so the resistance is always checked while the board is turned OFF. But I want to ask you about the ^ bold part above. When you say 'CANNOT' does that mean 'I wont be getting correct resistance values if I try to check resistance while the board is turned ON?. ...or I 'CANNOT' use multimeter ot measure resistance because its DANGEROUS?
                                      Both.

                                      In a low-voltage circuit, trying to measure resistance when there is power present may simply give bogus / fictional results.

                                      In a high-voltage circuit, trying to measure resistance when there is power present can damage your multimeter or even melt its wires violently (if sufficiently high current is available from the circuit), especially if it's a very cheap multimeter with marginal protections.

                                      Originally posted by twist3r View Post
                                      When you say that diodes (usually) fail short-circuit, does that mean that faulty diodes will most of the time BRAKE (cut off) the circuit? And that faulty diode wont allow anymore a current to pass through in both directions if its a faulty one? Is that the definition of a faulty diode (most of the time)? Did I understood this correctly?
                                      No.

                                      When a diode fails short-circuit, it acts simply like a piece of wire - any amount of current can pass through it in any direction it likes and the diode won't stop it.

                                      Every once in a while, though, a diode could fail high impedance / high resistance. In a case like that, the diode could act like a resistor, but very uncontrolled. It won't read a short-circuit / low resistance on your multimeter, so it could potentially trick you into thinking it is OK. This actually happened to me while fixing the ATX PSU in this thread:
                                      https://www.badcaps.net/forum/showthread.php?t=90392

                                      And then there's your case here with a physically broken diode, which read open-circuit. This is very rare... but physical damage does tend to cause diodes to go open-circuit. In the case of SMD ceramic capacitors, though, physical damage can often cause them to go short-circuit - so watch out for damaged/chipped ceramic caps.

                                      Comment


                                        #79
                                        Re: Samsung 2693HM - Doesn't turn on (no standby)

                                        More replies...

                                        Originally posted by twist3r View Post
                                        So...
                                        we measure resistance to GROUND when the board is turned off?
                                        and
                                        we measure voltage to GROUND when the board is turned ON?
                                        For the most part, that's correct.

                                        You don't always measure resistance to ground - that's just something we did here on the +3.3V_LIVE and 1.2V_LIVE, since these were more or less the two main voltage rails for the logic board.
                                        In general, you measure resistance to ground on voltage rails (and always with the power OFF) if the device has a "no power" symptom and you want to check if something is possible shorted (very low resistance) to ground.

                                        As for measuring voltage to ground - yes, that applies most of the time, so you can have a solid reference point (ground). Some exceptions do exist. But for the majority of voltage checks, you can assume the voltage should be measured to ground.

                                        Originally posted by twist3r View Post
                                        The only time we measure voltage on some element by putting multimeter probes on both sides of that element is when we want to see if there is a voltage drop? Correct?
                                        Correct.

                                        Originally posted by twist3r View Post
                                        This thing, the relationship between voltage, current and resistance confusing me for my entire life.
                                        And the reason for that is that freaking formula:

                                        V = I * R

                                        So... everytime I think about it, its like "ok, if we increase the resistance, then the resulting voltage can't be lower, but higher"
                                        Because you are ASSUMING current stays the same when resistance increases... and that's not necessarily the case (in fact rarely.)

                                        This formula can be written in several different ways (just algebraic manipulation):
                                        V = I * R
                                        I = V / R
                                        R = V / I

                                        If the voltage is constant across a resistor (for example, connecting a resistor between 3.3V_LIVE and ground or 1.2V_LIVE and ground on your monitor) and you know the resistance, then you use I = V / R to determine the current going through the resistor. So you can see in the above instance that with constant V and higher R, you will get lower I. And conversely, with constant V and very low R, you will get a very high I... which is what was happening with the faulty power lines on IC300 connected to the 3.3V_LIVE - they had very low resistance to ground (1 Ohm or less), making the current I = 3.3V / 1 Ohms = 3.3 Amps.... and that's at the very least, assuming the resistance really was 1 Ohms and not lower.

                                        This might not seem like a lot of current (modern PC CPUs can easily draw 50-100 Amps under load), but consider the size of the chip and that it has no heatsink on it.

                                        In that regard, another useful formula is the one that relates Power (P), Voltage (V), and current (I):
                                        P = V * I

                                        Knowing that the faulty power lines connected to L300-L307 on IC300 could draw possibly at least 3.3 Amps and that they are connected to the 3.3V_LIVE.... we can do a quick (likely inaccurate, but good for "guesstimating") calculation of the power:
                                        P = 3.3V * 3.3 Amps = 10.89 Watts (of heat).

                                        This amount of heat is equivalent to what an old AGP video card with a passive heatsink would draw or motherboard chipset with a heatsink. It's not a lot of power, but if you ever touched the heatsinks on those video cards or motherboards, you know how hot they can get sometimes. Remove the heatsink, and they will likely die quickly. So here, just doing this basic power calculation, we can see that something isn't quite for such a heatsink-less chip to draw so much power... further supporting the notion that the low resistance on those power lines on IC300 don't make sense.

                                        Originally posted by twist3r View Post
                                        What sweat (or water) does to a resistance? Lower it or increase it? My mind can't decide.
                                        Lower.
                                        Sweat contains water and various salts. Saline water is fairly to quite conductive (varying with the level of salinity, of course.) On the other hand, pure -distilled- water does not conduct electricity well at all.

                                        Originally posted by twist3r View Post
                                        When you said that if we put even tiny load (10-100 KOhm reistor) that would probably keep voltage to zero on the capacitor. First I thought 'why would we put (even) the small load, small resistor'. That would increase the voltage. But then I thought... we need some 'load' (some resistance) if we want 'voltage' to 'happen' across some element. Maybe thats the reason you mantioned 'small load'. Then again... if we put higher restor (bigger load), that must increase voltage? Right? (because of V=I*R)
                                        No.

                                        Lower resistance = bigger load (bigger current draw)
                                        Higher resistance = smaller load (smaller current draw)

                                        The only instance when the above do not apply is when you are dealing with a constant-current power source (i.e. a source that always tries to push the same amount of current through a device/component, regardless of its resistance.) This is not something we are dealing with here and in general not something you see on PC hardware. So ignore this for now.

                                        Originally posted by twist3r View Post
                                        Also... (maybe there is no connection here but) I have to ask: what about electrolytic capacitor's ESR? If its internal resistance (which increase over time) is high, will the voltage on that capacitor be lower or higher?
                                        ESR (also often referred to as impedance) is different from regular DC resistance and won't affect the DC voltage across a capacitor.
                                        ESR mainly has to do with AC voltages and frequency. I'll cut the discussion short here, because it's a long topic that's been covered elsewhere in much more detail.

                                        Originally posted by twist3r View Post
                                        Again diodes...
                                        What I thoguht I learned, is that diodes allow current to flow in on direction and not in the opposite direction. (Yes you said its always a leakage current, which is...(I guess) small, so not important).
                                        So when the diode fails, most of the time what that failure does to that 'line'? Does it brake the line so the current cant' flow in any direction, or it short the line, so current flow (like I guess happen when 230v went through my monitor's HDMI port)?
                                        When a diode fails, typically it will allow current to flow in any direction. How much will depend on how "badly" the diode failed.
                                        - In most cases, diodes fail short-circuit / very low resistance, so they act like a piece of wire: letting current flow any direction it likes.
                                        - Occasionally, a diode may fail high "impedance" (high resistance), which will still allow current to flow in any direction, but it will look like a resistor to the circuit.
                                        - Very rarely (typically caused by physical damage or a huge spike in power that blows the diode to bits and pieces), the diode can fail open-circuit - i.e. the diode will look like a cut piece of wire (not letting current go in any direction.)

                                        Originally posted by twist3r View Post
                                        Then again, there were 'protection' diodes, so... I really don't know if they failed by lettiner higher current flow, which means 'line' was not broken, ...OR they let the current flow and then failed by braking the line. I'm on the edge of misery here.
                                        The "protection" diodes in your monitor have a very simple function: the way they are connected, if a voltage "signal" goes over 4V on the HDMI or DVI signal lines (due to voltage spike or whatever other reason), it will get passed through the protection diode(s) on that line to the 3.3V_LIVE rail, so the voltage signal will be capped to 4V MAX. On the other hand, if a voltage "signal" goes less than -0.7V on the signal line, then ground will conduct through the protection diode(s) on that line and keep the signal capped at -0.7V max. This essentially means any signals going to the ICs will be limited between the range of 4V and -0.7V so as to protect the ICs.

                                        The MAIN PROBLEM behind the above implementation of the protection diodes in your monitor is that the 3.3V_LIVE rails feeds into the ICs and there is no over-voltage protection (OVP) on the 3.3V_LIVE to stop it from going above 3.3V. Thus, with no OVP on the 3.3V line, a signal on the HDMI or DVI line that has sufficient current and voltage over 4V can force the 3.3V_LIVE to go much higher than 3.3V... and that's what likely happened in your case when the cable box was connected. So once the 3.3V_LIVE went much higher than 3.3V, it then comes down to which IC is the weakest and breaks down the first from the over-voltage. In your case, that happened to be some of the power lines connected to L300-L307 on IC300. Basically, they took a hit and that's all there was to it - IC300 damaged.

                                        What Samsung SHOULD have done is put something like a Zener diode rated for 3.5 to 3.7V on the 3.3V_LIVE rail. Then, any voltage spike on the DVI or HDMI signal lines higher than 4V will pass through the protection diodes and get shunted to ground through that Zener diode. If that voltage spike has sufficient current, it would blow the Zener diode short-circuit... which will still "break" the logic board (i.e. it will need repair). But the repair would be much easier - just replace the shorted Zener and the monitor will work again. Instead, Samsung didn't do this and now you have IC300 dead. In other words, the addition of a $0.10 component (or less) could have saved your monitor.

                                        Originally posted by twist3r View Post
                                        Ok, one 'quick' question here: How you guys always calculate what voltage will be measured across that BIG FAT filter capacitor? What is the calculation formula?
                                        ...
                                        If 230v AC comes at primary tranformer side, what voltage comes out as DC at its secondary side after bridge rectifier? I was trying to guess that 230v AC was multiplied by 1.41 (square root (of 2), I don't know how it is properly said in english).
                                        Yup, that's correct.
                                        AC input * Sq_root (2) is indeed how it's done.

                                        I don't usually even bother with that, though. I just know that for 110-120V AC mains, expected DC voltage is usually around 160-170V and for 220-240V AC mains, the expected DC voltage is around twice that (i.e. 320-340V.)
                                        Last edited by momaka; 01-26-2021, 05:24 PM.

                                        Comment


                                          #80
                                          Re: Samsung 2693HM - Doesn't turn on (no standby)

                                          Hey momaka. Thanks for your reply! And sorry for this 1 month late respond. To be honest, I felt (and still am) exosted from 2 months of chasing the ghost.

                                          At the time when I bought the spare main/logic board, I felt the pressure to test it. It was because I bought it of an auction site, and the seller was saying 'don't swap the firmware chips' just put the board in the monitor and try if it will work.'. He was saying same what you were told me. But I had to contact you, and thats why I filled your PM box, cause I didn't want to just do what seller tells me before I check with you.
                                          And also, bad thing was (is) that I can't assemble my monitor easy and keep it open. Because Samsung made it in a way, that the 'cradle' that holds power board and logic board is not 'screwed to anything. inside. Its the plastic back cover that holds it in place. Now whats bad is that once I put back cover its not easy to remove it again, without risking for those plastic 'snaps' to broke. I broke one piece 2-3 months ago when we started this journey. So I was under pressure to test the board by completely assembiling my monitor, and be sure that I wont have to open it again.
                                          In the process I forgot to put 4 screws that hold speakers, so I was pizzed that I had to remove the back cover again and risk for the rest of plastic snaps to brake. I managed not to brake another one, but it was close.

                                          Then, I tested the board and made my post with the results I wrote about a month ago. After that I put my monitor aside to wait some better time, when I'll get a hotair station to be able to remove firmware chips from the old main/logic board and transfer them to a new board.


                                          So...

                                          Thansks for your suggestion and instruction how I can 'short' certain connector points with a wire, BUT...I (currently) don't like that idea. I lack the knowledge and honestly I wont be happy to screw something now when I think I (at least) got to some point where everything (kinda) looks like it 'works'. Ofcorse it doesn't work the way I would say its 'good', but... I got the picture on the monitor. Since I'm not fully satissfied with it, I put the monitor aside, until I can simply solder the old main board firmware chips to a new board chips. In other words, I decided to wait a bit, and see if I can buy a hotair station in some (hopefully not distant) future.




                                          Originally posted by momaka View Post
                                          If your SMD component (ceramic cap, in this case) gets stuck to the tweezers, then it's not getting soldered to the board.

                                          Clearing the solder on one side with wick can help, but is not necessary.

                                          What I do is I use a sharp chisel or bevel or spoon tip and make sure it has no solder or almost no solder on it, then run it with flux over the pads that have solder on them and collect a little bit of solder from them (but still leave a tiny layer on there.)

                                          After that, I put more flux on both pads again. If there is any burned, hardened flux on the pads from before, make sure to scrape it off before applying more flux. Old, hardened flux on the pads may not allow solder to stick to it. Once the pads are fluxed, place the SMD component and hold it in place (be it with tweezers or finger nail.) Then, using a sharp chisel or bevel tip again (same as the one you cleaned the pads with), heat one side of the SMD component to solder it. Make sure you touch both the PCB pads and the component leads, as this will melt the solder on both and join them instantly. For those very small "sandgrain" caps, they should literally solder in just a second or two when you touch them with the iron's tip.

                                          SMD work just takes a lot of practice and a bit of steady hands - that's all.
                                          Thanks for the tips!

                                          Well, when I said that 'sandgrain' ceramic cap is getting stickied to my tweezers, is that when I used (no clean) flux paste, the paste itself was sticky. So the cap just got stickied on the tweezers and refused to let me position him on the pads. Once I thought I got the cap in the place, ready to solder it, the moment I move up the tweezers, it lifted the cap along with them. And the way I held that small cap with the tweezers, I coudn't solder it, because tweezers were blocking the soldering point.

                                          Btw. at the time I had some old tweezers. In the meantime, in last 3-4 weeks, I bought some package of 10 ESD tweezers, that were on sale in local market. They were not perfect, but I believe they are much better than the old tweezer I used. So, in the future, maybe it will be a bit easier for me to solder such a tiny sandgrain ceramic caps.


                                          Originally posted by momaka View Post
                                          Yeah, isn't it crazy how small some of these components are?
                                          It's even worse on phones and tables! That's why I don't work on those.
                                          Yes! I guess same goes to those small caps on the PC grapic cards. When I look at them, I lose any interest to find the faulty one. For that I would need a realy nice scope, which I don't own currntly (and don't know if I will, in near future).



                                          Originally posted by momaka View Post
                                          Originally posted by twist3r View Post
                                          I found that chip on aliexpress. Its not 'too expensive'...like $4-$6.... BUT it comes with ~$75-$80 shipping cost to my country, so thats that. End of the line for possible future 'fix' of my old main/logic board.
                                          That's insane!
                                          Why is the shipping to your country so much?
                                          Oh, man... if you really knew the whole truth behind it.
                                          Its a huge topic in my country that lasts probably a few decades already.

                                          I'll try to put here just few major problems so you can guess how everything 'works' in my country.

                                          Lets say I bought a PC mouse of eBay for 40 bucks from the seller which is located in USA. And I paid the shipping cost (lets say) $30.
                                          It arrived to my country and do you know how much fee I have to pay for it?

                                          Every 'sane' law in every normal country woudl say the fee should be calculsted based on the value of the item I purchased, which in this case was 40 bucks. But not here. A value of that mouse is not 40, but 40+30 (shipping), which ends up being 70 bucks.

                                          Now, I have to pay 20% VAT on that ($14), which ends up to 84 bucks. Then at the top of that I have to pay import tax of 10%, which is not calculated on 70 bucks, but on 84 bucks ($8.4), which ends up being 92.4 bucks.

                                          Now, when you think 'is this nightmare finally over?', think again.... because:

                                          1. If (you paid) or custom personel decided that there was a 'spedition service involved', you have to pay 'spedition' to DHL (doesn't matter if you already paid that shipping cost when you order the item). So...(last time I checked) spedition was around 40-45 euros. Lets say, for now its 40 bucks.

                                          Do you know how everyting is now calculated?


                                          $40 - for the mouse
                                          $30 - shipping cost
                                          $40 - spedition
                                          -------------------

                                          $110 (this is what custom personel will use to calculate taxes and fees). so...

                                          10% import fee on 110 is 11 bucks, which raise the value to 121
                                          20% VAT which use $121 as a base, and after calculation (24.2) it raises the value to $145.2

                                          Thats the worst (or at least you would think) value you can pay. BUT...

                                          2. At the top of that, custom people OPEN every package that they decide to open. For that, you also have to pay them because they moved their 'azzes' and did their jobs, so you have ot pay few more bucks for their 'service', even though its in their job description.


                                          3. custom personel may decide that they simply don't trust the value which was written on the package. So for chinese packages its common thing that its written lower price on the package so that the tax cost could be lower. From other countries, like USA or Japan, they might 'believe the price', BUT...they can still decide no to believe the price on the package.
                                          Imagine you bought the mouse during some 'on sale' time, with much lower price. Maybe that mouse that I bought for 40 bucks, has its regular price (before 'on sale' event) $60 ?

                                          OR, custom personal just decide not to trust any price and then they check prices in my country. The whole reason why people order items (which can be bought in my country as well) is that oversea prices are much lower compared to prices in my country. We pay here for things much more compared to others. Its because of corruption, and some companies here, got exclusive rights to import certain items. They distribute those items to different shops (not the same shop brand) but the prices are identical in every shop, with ofcorse much higher price than what you can buy in-person in other country, not to mantion prices from an online shop outside my country.

                                          Thats just one part of a huge problem.


                                          The other part, is that once you order something you get tracking number, so you can track your package and know where it is currently. Thing is... once it reach my country boarder, a lot of time its just like it went into Bermuda Triangle. There is nothing there.

                                          Reason for that is: The shipping task for that packagte, once it enters my country is done by national shipping service. People at custom personel (and people from shipping service) open the package on their own (cause they like to peak and see whats inside). And, if they like the content of the package, they often steal whole package or some parts of it.
                                          For that reason, the person who order it, complains to a seller, if the item didn't arrived or if it lacks some parts in the package. So they ask for refund, and seller has to return the money.

                                          And at the top of all this sh1t here, a lot of people in my country has a state of mind which makes them want to scam someone (in this case seller) so they ask for refund telling him they didn't received the package even if they did actually.

                                          For all these reason, what the rest of us think, sellers from aliexpress raised the prices for shipping cost by letting us to use only DHL or Fedex (or similar shipping service), which deliver the package to our doors (instead of going through national shipping service).


                                          There is also a thought that governamant made some kind of this mess by forcing sellers to increase shipping cost in order to stop people from my country to order items for cheap price, becaues they will be then forced to pay way overpriced values of the same item if they are bought from the shops in our country.

                                          A lot (a whole lot) of aliexpress sellers don't even offer to sell anything anymore to the citizens of my country.

                                          I hope I managed to explain just part of the problem here.


                                          Originally posted by momaka View Post
                                          Maybe if you have a relative living somewhere in EU zone, have the chip ship to them, and then have them repack the ship to you in a small postal mail card or something similar. I actually sent a chip this way not too long ago to a Serbian compatriot of yours, and he received it fine without issue. Cost me like $2 or 3 total in mail stamps, I don't remember exactly... but not much, really.
                                          Oh, I thought about it. Sometimes I can work sometimes it wont. I mean it can work, but depends if it will be cheaper. For certain thing like a ship, it may end up being cheap, thats correct. But for other (more expensive things) it may end up being equal or almost equal to the price, because you would paying 2 shppping costs. One to your relative in EU zone, and then the price they pay to send you from EU zone to you.

                                          Sometimes, some custom offices in my country don't check (or don't open) some small packaged, so they end up being delivered by a post man without any request for your to pay some extra fee for that item. It gets the stamp on the letter 'no custom fee'. But that is rare, and may work only for small items.
                                          Good thing is that in case of that chip, it falls under that cathegory, so maybe I'll try that once, and see how it will end up.
                                          Last edited by twist3r; 02-18-2021, 10:28 AM.

                                          Comment

                                          Working...
                                          X