Replacing OST caps on Microstar KT6 Delta motherboard

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Gosh, i really need to sharpen up, and watch out for dyslexia (which, IMO, is really not just paying enough attention)
Not sure if it happened while reading it, or during the transition to writing it down.
Or perhaps i got distracted by that buzzing noise, and it was past bed time.
Note to self : Next time you see 2_5 and get a 'reading' of something 2.15, recheck it.
Sorry 'bout that

Rechecked VSUS2_5, and it was 2.51V - not 2.15V
So looks like it's meant to be on.

The voltages at that op-amp U12 are:
Pin (3 & 5) (VREF) = 1.25V
Pin 8 (9VSB) = 9.67V
Pin 2 = 1.26V which goes to resistor R245, and other side of this resistor has 2.51V (VSUS2_5)

I can see, however, that when the PC is running, that "9VSB" becomes 12V because of diode D11.

Does that mean that it becomes a 12V serial bus, or is it still a 9V serial bus with 12V on it?

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

This is still sitting on the desk, so thought to have another crack at it - since all that other testing appeared to be ok
The power switch is activating a transistor Q6 (page 43) that would send PS_ON to ground (except that resistor R44 is missing, and was never there in the first place)
The transistor is working, as the voltage at the collector? drops from 3.39V to 0.78 or 0.6 when the power switch is pushed, and the base? jumps from 0V to 0.079V
Because this resistor is not present, i want to find out what component sends PS_ON to ground, to start the PSU.
PS_ON (and 5VSB via a resistor) is connected to PS_OUT# which goes to an Over Temperature Protection Controller Chip

PS_ON is also connected to PW_OK via a resistor , a transistor and a schottkey rectifier.

Any more clues?

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Well, it just dawned on me... but maybe try to see what happens if you short the green wire on the PSU ATX connector to ground while the PSU is plugged in the motherboard (along with the 4-pin CPU plug).

Because this resistor is not present, i want to find out what component sends PS_ON to ground, to start the PSU.

Well, let's start at the front panel power button. In the schematic, it is connected between ground and PWBTIN#. PWBTIN# goes to pin 1 on U1 (ATTP1/2??? not sure what this IC is) - see page 41. My guess is that U1 then controls signals PS_IN# and PS_OUT#. But I am not sure what that U1 IC is. C an you post what part number is written on it?

PS_ON is also connected to PW_OK via a resistor , a transistor and a schottkey rectifier.

You can ignore that, because PW_OK only matters *after* the PSU has been turned ON via PS_ON signal.

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Looks like a thermal protection IC which will shutdown the system if the CPU overheats, see THERMDN and THERMDP lines on Page 5.

It intercepts the PS_ON signal from Q6 to the PSU. Since we know Q6 works, problem could be U1 or one of its support components.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Shorting the green wire (on the PSU) to ground, causes the CPU fan to spin, but no BIOS Beep, or video signal.
The HDD and power LEDs turn on.

The part number written on the U1 chip is Attansic ATTP1.

Looks like a thermal protection IC which will shutdown the system if the CPU overheats, see THERMDN and THERMDP lines on Page 5.
It intercepts the PS_ON signal from Q6 to the PSU. Since we know Q6 works, problem could be U1 or one of its support components.

So, does that mean that the thermal protection IC connects the terminals either side of where resistor R44 would have being?
Would it be a good idea to put a resistor in that gap and see what happens?
Or is there a more sensible/safer approach?

The Voltages on that chip U1 are as follows:
1 PW_DN: 5.01
2 GND: to Ground
3 PS_IN#: 0.61
4 TMP-D+: 0.79
5 TMP-D-: 4.46
6 PS_OUT#: 4.93
7 VREF: 4.75
8 5VSB: 5.01

Pushing the power button once, and then rechecking, did not change any of these values.

[Sparkey55]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Have you tried to start the computer with no keyboard connected?

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Yes, i have tried that - No keyboard has being connected throughout this testing

[Sparkey55]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

In post #53, I think, you said that the CPU got hot. How hot did it get and what did you use to determine the heat; your fingers on the heatsink or something else?

[Sparkey55]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

I think your Southbridge chip is working ok based on what is being indicated by the leds for the hardware. Have you checked to see if the CPU heatsink is flat on the CPU? Does it still get hot?

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

So, does that mean that the thermal protection IC connects the terminals either side of where resistor R44 would have being?
Would it be a good idea to put a resistor in that gap and see what happens?
Or is there a more sensible/safer approach?

Yes it connects where R44 would go. The schematic shows R44 as "X_0" - the X likely means "not stuffed" (stuffed means 'put on the board') and 0 probably means it would be 0 ohms normally, which makes sense.

You could put a resistor in for testing purposes (You would of course lose CPU overtemp protection) but you would need to disconnect at least the output pin of U1 to avoid backdriving the output and damaging it (U1 might be perfectly OK, we don't know yet)

The Voltages on that chip U1 are as follows:
1 PW_DN: 5.01
2 GND: to Ground
3 PS_IN#: 0.61
4 TMP-D+: 0.79
5 TMP-D-: 4.46
6 PS_OUT#: 4.93
7 VREF: 4.75
8 5VSB: 5.01

Pushing the power button once, and then rechecking, did not change any of these values.

Sorry, I meant check voltages between when the button is not pressed and when it is held down, not when it is pressed and then released.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

In post #53, I think, you said that the CPU got hot. How hot did it get and what did you use to determine the heat; your fingers on the heatsink or something else?

Yes, fingers on the heatsink

More carefully rechecked the voltages on that U1 chip (was mixing the "on" reading with the "off" readings - Next time i'll wait till the morning before doing this stuff)

The power button does talk to it

1 PW_DN: 5.01
2 GND: to Ground
3 PS_IN#: 3.5
4 TMP-D+: 0.89
5 TMP-D-: 3.88
6 PS_OUT#: 4.93
7 VREF: 4.42
8 5VSB: 5.01

When i push the power button and release it quickly, the voltage on pins 3, 4, and 5 change:
3 goes to 0.51 or 0.75
4 goes to 0.75 or 0.83
5 goes to 3.93 or 4.22

(and these 3 voltages are not exactly the same each time i measured - variance is within 0.1 of a volt)

Then pushing the button down again (and holding it down) resets these voltages back to their original values

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

You could put a resistor in for testing purposes (You would of course lose CPU overtemp protection) but you would need to disconnect at least the output pin of U1 to avoid backdriving the output and damaging it (U1 might be perfectly OK, we don't know yet)

Well I found this interesting bit from the ATTP1 datasheet (page 3):
When the temperature reaches the preset trigger value (Over Temperature Protection, OTP, event occurred), PSON_OUT is pulled up to turn the system power off, and a siren signal is sent out. PSON_OUT will keep logic high until AC/5VSB power loss, or PWRBTN pressed.

Since PS_IN# is not changing when the power button is pressed, then I don't think there is a point to install that resistor.

Yes it connects where R44 would go. The schematic shows R44 as "X_0" - the X likely means "not stuffed" (stuffed means 'put on the board') and 0 probably means it would be 0 ohms normally, which makes sense.

Interesting observation.
However, I see lots of other resistors with X_Y (where Y is the resistance) on the schematic. That includes R62 (X_0) on page 41 going to pin 1 of U1 as well as R68 (X 10k) on page 41 again, which partially controls OTP TRIGER through Q11. If these resistors are not there... then PWBTIN# wouldn't be going anywhere. And that shouldn't be the case, because PWBTIN# is what comes out of the motherboard's power button on front panel

I traced the PS_ON signal in the schematic, and here is how I think it flows:
> PWRSW is the actual header on the front panel that needs to be pulled LOW (i.e. 0 V) to start the board. PWRSW is connected to PWBTIN# through R412 (page 44).
> PWBTIN# is pulled high (3.3 Volts) to 3VDUAL voltage through a 4.7 KOhm resistor (RN107, page 45).
> PWBTIN# partially controls OTP TRIGER (page 41) (OTP TRIGER state can be changed both by PWBTIN# and HI TRIP signals).
> OTP TRIGER goes to the CLOCK input of a D flip-flop (74HCT74, U4b on page 41 again.), and this D flip-flop controls signal THRM_PSON.
> THRM_PSON is directly connected to the base (B) of transistor Q6 (page 43). Transistor Q6 directly controls signal PS_IN#.
> PS_IN# goes to pin 3 of U1, ATTP1 (page 41 again).
> U1 then controls signal PS_OUT# (pin 6) based on PS_IN# and other signals (THERMDP, THERMDN, VTIN2).
> PS_OUT# is directly connected to PS_ON on the ATX connector.

Pressing the power button: what happens...
PWRSW/PWBTIN# are initially pulled HIGH (3.3 Volts). With PWBTIN# HIGH, transistor Q11 is closed across C-E terminals. Therefore, the voltage on the anode of D2 is LOW (0 V or close). At this point, this means that OTP TRIGER is only controlled by HI TRIP signal.

Pressing the front panel PWR switch pulls PWRSW/PWBTIN# LOW. This makes Q11 open across C-E terminals, thus allowing a voltage from 5VSB to go through R59, reaching the anode of D2, going through D2, and out to OTP TRIGER, making it HIGH (about 5 V). This will CLOCK the 74HCT74 D flip-flop, U4b, and output a HIGH/LOW signal (not sure which) for THRM_PSON, which will make the C-E terminals of transistor Q6 closed/open. With Q6 open, PS_IN# will be pulled from LOW (about 0 V) to HIGH (about 5 V), signaling U1 (ATTP1) to pull PS_OUT# low as well, provided that signals THERMDP, THERMDN, and VTIN2 are also what they should be (I haven't traced what those should be). With PS_OUT# set to LOW, PS_ON on the ATX connector also becomes LOW, thus enabling the PSU.

Now the fact that the voltage on PS_IN#, pin 3 of U1 does not change (0.51 V -> 0.75 V is not really significant IMO) suggests that something before U1 is preventing the PS_IN# signal from coming in. Since PS_IN# is controlled by Q6, you should first check the voltage on the base (B) of Q6 before the power button is pressed and while it is pressed to see if there is any change. If not, you go back further - and that would mean follow signal THRM_PSON backwards. THRM_PSON is controlled by Q20 and R82 (page 41), which are both controlled by pin 9 of U4.

But let me know what you get so far and we will continue from there.

Also, another interesting observation from the datasheet: when you see a signal, you might also see a number or numbers (separated by comma) next to it. At first, I thought this indicated pin numbers on ICs. But it turns out, this indicates the page # where you will see that signal used again. So just a heads up to make it easier for you when looking through the datasheet. It sure made it a lot easier for me now when I look at signals.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thanks for the superb (as well as fascinating) explanation

Now the fact that the voltage on PS_IN#, pin 3 of U1 does not change (0.51 V -> 0.75 V is not really significant IMO) suggests that something before U1 is preventing the PS_IN# signal from coming in. Since PS_IN# is controlled by Q6, you should first check the voltage on the base (B) of Q6 before the power button is pressed and while it is pressed to see if there is any change.

Yes, it does change - it goes from 3.5V -> (0.51V or 0.75V)

Here they are anyway:

Voltages at transistor Q6:

BASE:
initial - 0V
Button pressed - no change
Button released - 0.8V

Collector:
Initial - 3.62V
Pressed - No change
Released - Either 0.36V or 0.70V (appears to be random each time)
Also, does some voltage still remain (instead of dropping to 0V), because the transistor C-E has some resistance?

So what i'm seeing (please correct me if i get this wrong) when the power button is released, is that the collector-emitter of Q6 conducts, and therefore, pin 3 of U1 (PS_IN#) is pulled low (from 3.5V to 0.51V).

With PS_OUT# set to LOW, PS_ON on the ATX connector also becomes LOW, thus enabling the PSU.

Testing with the meter shows that PS_OUT# is not changing from 4.93V when the button is pressed and released - it's not going LOW,, and PW_DN (SIREN# on datasheet) is HIGH - so the PSU is not starting.

provided that signals THERMDP, THERMDN, and VTIN2 are also what they should be

So, on to those now?


Just a quick, 'stupid', side, question:
There is 3.29V on the power switch pin (PWSW+) before its pressed - Where is this voltage coming from? (i searched for "PWSW" in the document, and no other pages came up).
Surely it's not coming from the 5VSB and back through transistor Q11 (top of page 41)

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Haven't being able to find a proper datasheet for this ATTP1 IC yet

Found this tidbit from another forum
http://www.aoaforums.com/forum/amd-m...t-thermal.html

I connect Pin 4 & 5 Attansic ATTP1 working great.

Tempted to try it, but will wait and see what you have to say.

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Found this tidbit from another forum
http://www.aoaforums.com/forum/amd-m...t-thermal.html

Tempted to try it, but will wait and see what you have to say.

Okay, I'll give it an in-depth reading tomorrow and let you know. 1 AM here .

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Just a quick, 'stupid', side, question:
There is 3.29V on the power switch pin (PWSW+) before its pressed - Where is this voltage coming from? (i searched for "PWSW" in the document, and no other pages came up).

The signal for the front panel power switch is called PWRSW (page 44, pin 6 of JFP1), but the pin *name* is called PWSW+. I think this is what confused you.
Then, if you follow PWRSW, you will see that it is connected to signal PWBTIN# through R412, a 22-Ohm resistor (page 44).
And then, if you follow PWBTIN#, you will see that on page 45, PWBTIN# is pulled high to 3.3V (by 3VDUAL) through a 4.7-KOhm chip resistor RN107.
^ That, and R63 (page 41) is also helping to pull it high (it is a 100 Ohm resistor going to 5VSB.)

Surely it's not coming from the 5VSB and back through transistor Q11 (top of page 41)

That's right, it can't.
Q11 is a NPN trnasistor, so the B-C (Base-Collector) junction appears like a reverse diode, and current cannot flow through that.

Yes, it does change - it goes from 3.5V -> (0.51V or 0.75V)

Ah okay, I didn't get that from the previous posts. So it looks like signal is getting through to PS_IN#.

Voltages at transistor Q6:

BASE:
initial - 0V
Button pressed - no change
Button released - 0.8V

Collector:
Initial - 3.62V
Pressed - No change
Released - Either 0.36V or 0.70V (appears to be random each time)
Also, does some voltage still remain (instead of dropping to 0V), because the transistor C-E has some resistance?

Base and Collector voltages look good.

Yes, you can think of it like that - i.e. there being a certain resistance between C-E and thus the voltage not being able to drop down to 0 V. The more technical explanation is that in a NPN transistor, the Collector and Emitter silicon materials (both N-type) are separated by the Base (P-type) - hence the NPN name. The junction between Base and Emitter is thus PN - exactly the same as a diode in forward bias. Therefore, the B-E (Base-Emitter) voltage drop will always be that of a diode or higher (so typically around 0.6-0.8 V). Thus, when the BJT is conducting, the C-E voltage drop cannot be any lower than that of the B-E voltage drop.

So what i'm seeing (please correct me if i get this wrong) when the power button is released, is that the collector-emitter of Q6 conducts, and therefore, pin 3 of U1 (PS_IN#) is pulled low (from 3.5V to 0.51V).

Correct.

Testing with the meter shows that PS_OUT# is not changing from 4.93V when the button is pressed and released - it's not going LOW,, and PW_DN (SIREN# on datasheet) is HIGH - so the PSU is not starting.

A-ha.
So that means either ATTP1 chip is getting a fault condition from the other signals (THERMDP, THERMDN, and VTIN2) or ATTP1 is bad. So this is where we start to dig in below...

But first, I address this:

Found this tidbit from another forum
http://www.aoaforums.com/forum/amd-m...t-thermal.html

"I connect Pin 4 & 5 Attansic ATTP1 working great."

Tempted to try it, but will wait and see what you have to say.

Sounds like a plausible thing to try. However, what I would suggest you do first is check the voltage on pin 2 (ground, GND), of the ATTP1 chip. If it is *not* LOW/0 V after you press the power switch, then there wouldn't be any way for pin 6, PS_OUT#, to be pulled LOW either.
However, if it *is* LOW/0 V, then that leaves only the signals on pins 4, 5, and 7 as the possible culprits.

According to your post #65 above, then the only possibly bad signal is on pin 5, signal THERMDP (or pin name TMP-D-, if you like). I am thinking this needs to be closer to the voltage on pin 4 of the ATTP1 IC...
So I guess, YES, you can jumper pins 4 and 5 together on the ATTP1 IC. Either that, or can solder a resistor across RT2 (see page 41). I think 100 to 1000 Ohms would work for that. If my memory serves me correctly, those thermal resistors (like RT2) are PTC type, so that means their resistance increases with rising temperature. Thus, lowering the resistance (by installing a resistor in parallel/across RT2 terminals) should make the ATTP1 IC think the temperature is lower.

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Just beware that if you jumper those pins via R44, the output will be pulled low even though it's trying to output a high, which would be like shorting it to ground. It would be better if you desoldered\lifted the output pin first, before installing that resistor.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

So that stupid question was actually stupid - at least i knew it was.
Yes, missed seeing the "R" in the middle. Thanks for pointing it out.
At least i didn't mess up the meter readings this time.

Sounds like a plausible thing to try. However, what I would suggest you do first is check the voltage on pin 2 (ground, GND), of the ATTP1 chip. If it is *not* LOW/0 V after you press the power switch, then there wouldn't be any way for pin 6, PS_OUT#, to be pulled LOW either.
However, if it *is* LOW/0 V, then that leaves only the signals on pins 4, 5, and 7 as the possible culprits.

Yes, pin 2 remains at 0V after the power button is pressed and released.

According to your post #65 above

Refer to post #71 instead (the measurements in post #65 are not all correct)

So used a breadboard jumper wire to connect pins 4 & 5 with my hand
Still no sign of life after pushing and releasing power button
Made sure that i placed them on the same spots where i placed the meter probes (because on the sides of the pins there must be some non-conductive coating)

Then accidentally touched pin 6 instead of pin 5 (so pin 4 is jumpered to pin 6), and the CPU fan started to spin (and kept spinning)
This happens regardless of whether the power button has being pressed or not.
Green power LED lights up and keyboard LEDs flash once - but no BIOS beep or video.
Since pin 4 has 0.89V, and i jumper it to pin 6 (which has 4.92V), does that cause pin 6 (PS_OUT#) to be pulled LOW?

The thermal resistor RT2, located below the CPU, is not present and looks like it was never installed

PS

According to your post #65 above, then the only possibly bad signal is on pin 5, signal THERMDP (or pin name TMP-D-, if you like)

How did that eliminate pins 4 and 7 as culprits?

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Made sure that i placed them on the same spots where i placed the meter probes (because on the sides of the pins there must be some non-conductive coating)

That non-conductive coating is actually just a thin layer of flux leftover from soldering from the factory (they apply flux over the whole board, then run it through a wave solder station.) You can usually remove it either by cleaning with IPA (Isopropyl Alcohol) or by scratching with something metal (like a small screw driver).

Yes, pin 2 remains at 0V after the power button is pressed and released.

Good.

Refer to post #71 instead (the measurements in post #65 are not all correct)
So used a breadboard jumper wire to connect pins 4 & 5 with my hand
Still no sign of life after pushing and releasing power button

Noted.
I guess since pin 4 is low, it also pulls pin 5 low. If pins 4 and 5 are to be high or perhaps equal to VREF, then I can see why this doesn't work. But that's a big *IF* here.

Then accidentally touched pin 6 instead of pin 5 (so pin 4 is jumpered to pin 6), and the CPU fan started to spin (and kept spinning)
This happens regardless of whether the power button has being pressed or not.
Green power LED lights up and keyboard LEDs flash once - but no BIOS beep or video.
Since pin 4 has 0.89V, and i jumper it to pin 6 (which has 4.92V), does that cause pin 6 (PS_OUT#) to be pulled LOW?

Yes.
Pulling PS_OUT# low is essentially the same as grounding signal PS_ON on the ATX connector.

The thermal resistor RT2, located below the CPU, is not present and looks like it was never installed

And yet the schematic shows it there as 10 KOhm PTC.
I hate it when that happens! (Schematic not agreeing with the circuit, that is.)

How did that eliminate pins 4 and 7 as culprits?

You're right. The more I dig into this, the more I wonder too.

Some thinking out loud here...
The above makes me wonder what should be the voltages on pins 4 (TMP-D+), 5 (TMP-D-), and 7 (VREF). The scanty ATTP1 datasheet we have barely mentions anything useful. The biggest question I have is about pin 7, VREF. Normally, IC VREF pins are used as a voltage reference to compare other signals to, and it's usually a standard voltage (like 1.25 V, 1.5 V, 2.5V, 3,3V, etc.). That makes me believe that if pins 4 and 5 have a voltage close to VREF, then perhaps that would indicate "normal" temperature. Of course, that might not be the case.

And then there is this question: is the voltage on VREF supplied externally (i.e. from another IC) or internally (generated inside the ATTP1 IC)? I'm thinking it's the former, because VREF is connected to THERMDP through R34 (4.32 KOhms, page 41), and THERMDP connects to 3VDUAl through resistor R39 (10 KOhms, page 41 again). Now, I am not sure what CP2 or CP14 are, but if they are something like jumpers, then that means that THERMDP is likely going to measure at least 1.25V. And if not, then THERMDP should possibly be closer to 3.3V.
Either way, I see that in post #71, you indicate that it is 0.89 V.
Perhaps that is the problem.

Now, can you check if your board has resistors R99 and RT1 installed on it (see page 35 - they should both be 10 KOhm resistors, with RT1 being an SMD thermal resistor.) Also check if R29, R30, and R34 (page 41) are also there and what resistance values they have when measured with a multimeter (remove power before doing that, of course). And lastly, check resistance of THERMDP to ground (again, with power removed).

Let me know what you get.

I think we are going to get to the bottom of this soon. Or perhaps at least figure a way to trick the board to start when the power button is pressed.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thanks for thinking out loud

VREF = 4.42V, so how can that come from 3VDUAL?
Can 3VDUAL have more than 3V on it?

R99 = 6.47K
RT1 = 6.59k
(both located under the CPU)

R29 = Not installed
R30 = 15K
R34 = 3.60K

THERMDP to ground = 57K

Disconnected BIOS battery this time

(Glad that you didn't ask about C13, C14, CP2, CP14, R39 or U3, because i can't find them in the vicinity of the CPU or U1, and perhaps they are not on this board - because this board is version 2.0
U3 looks like it's a 14 pin IC, and the only 14 pin IC that i can find is the HIP6602 - which is already accounted for)