Announcement

Collapse
No announcement yet.

Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

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

    Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

    This one should make @Pentium4 smile (if he is still reading BCN forums) – I got a new old stock / open box Casing Power MPT-301 PSU on eBay for $4 total.


    Yes, it’s a very generic-looking box, but the manufacturer isn’t (at least not back in the days), which is…

    Macron Power Technology Co. LTD.

    Let’s look at the PSU itself.



    The shell/case also appears quite generic, like many “sold-with-the-PC-case” units. However, the shell is well-formed and has decent steel thickness. In terms of output wires, it’s pretty standard for the era: 1x 20-pin ATX main connector, 1x 4-pin 12V CPU, 4x Molex drive, 2x floppy, and 1x AUX 3.3V power connector. Most of these are 18 AWG size, but the 4-pin 12V CPU and AUX connector use 20 AWG wire (and of course, the floppy is 22 AWG.)

    Next, the label:

    Does this look familiar now? It should. No, it’s not the same PSU as the ADP MPT-301 that I recently recapped. It’s more like this Inno Power PSU I posted a while back.

    Part of the reason why I got the MPT-301 here so cheap was this:

    Not sure if this was caused by a drop or something else, but the grille on the rear fan is pretty mangled… though the PSU can still be mounted in a case and there is no other damage anywhere else.

    The other reason the PSU was so cheap was… well, it’s a bit of a long story, as usual. But in fewer words, the eBay seller I bought it from messaged me to say they couldn’t ship the PSU the same day (which I never really asked for, so no problems ) and instead would likely be shipped either the next day or the day after that. I replied, saying that’s not an issue at all, since the PSU was intended for a retro PC / hobby project and thus wasn’t needed in a hurry. Then the seller messaged me back and curiously asked about my retro PC build. So we exchanged a few messages back and forth about old PC hardware. Turns out the seller also had intended this PSU for a retro PC build. But the motherboard he had seems to have burned out from the previous PSU (and likely from bad caps. ) There’s probably more to follow on this story, but I’ll leave that for another thread/post, as it pertains more to motherboards.

    Anyways, now let’s look inside this particular MPT-301 unit.




    EMI/RFI board:


    Primary Side:



    Secondary Side:



    As you can see, it DOES look very much identical to the Inno Power I linked to above. However, it actually isn’t the same. The PCB on this one says ATX9806b-p whereas the Inno Power (and a Macron Power MPT-401 I also posted) has a PCB marked with ATX9912p. There are minor differences between the two, though the design and layout is nearly the same.

    Some of you may (again) still question this PSU based on the generic looks. But let’s not forget it’s from 2003. Back then, this kind of build quality was actually considered on the decent side. Sure it’s not built like the older Delta and HiPro units with oversized heatsinks and big rectifiers. But it’s also quite far from being a gutless PSU.

    So let’s talk about the good things here:
    - PSU has a complete EMI/RFI input filter
    - connector on the PCB for the AC input wires
    - overall good-sized heatsinks and output toroid inductors
    - space for a few 12.5 mm diameter caps on the output
    - 3-Amp rated 5VSB with its own TO-220 rectifier bolted to the secondary HS.

    Next, the mediocre stuff:
    - half-bridge design with NPN BJTs and secondary-side PWM (reliable but outdated)
    - main transformer uses EI-33 core (instead of the larger EI/ERL-35)
    - protection fuse is a glass one (so it needs to be either heatshrink wrapped or better yet, be a ceramic one)
    - bridge rectifier consists of individual 1n5406 diodes (though they should be adequate for the job with active cooling from the fan)
    - the heatsinks could have been bigger, as there is a lot of empty space left inside the case
    - TO-220 BJTs on the primary (2SC5679 – please anyone tell me if you can find a datasheet for them), which might be cutting it a bit too close for 300 Watts continuous (maybe 300 Watts peak for a short period… though I still wouldn’t feel too comfortable with that, given the heatsinks.)
    - 20 AWG wire on the AC input side
    - 2-transistor 5VSB with feedback - kind of “meh” today, though considered OK back then. At least this one doesn’t have a “critical” cap.

    And for the not-so-good:
    - Three of the five Y2-class caps for the EMI/RFI filter are just regular 1kV ceramic caps (so UL hi-pot testing likely won’t do well here)
    - the capacity of the output caps is a bit on the low side (only 2200 uF for the 3.3V rail)
    - there are no PI inductor coils between caps on the output… though I don’t know if that counts as “not good” here, since the PSU could have been designed without them. The Inno Power also doesn’t have them on the 12V and 5V rails by design, IIRC
    - 12V rectifier is only a F12c20c part – therefore, that “15 Amp” rating for the 12V rail on the label is probably a “peak” value and not continuous (which surprises me, as Macron doesn’t usually over-rate their PSUs.)
    - BAD CAPS! Perhaps hard to see, but take a look again at that picture of the secondary side – one of the 10 mm caps on the lower-left side of the picture is bulging ever so slightly. To be honest, though, I was expecting much worse from this PSU, as Macron Power often used some of the worst garbage cap brands in their PSUs. In this one, it’s mostly CS-logo caps, which IMO are “medium-bad” for these PSUs (with the worst being GL –branded caps, and the “best” being Fuhjyyu and CapXon ) In fact, one of the reasons I bought this PSU was because I was curious how bad the caps inside it must be and just felt like saving another older PSU.

    So my overall thoughts: it’s still a pretty decent PSU for its time, though perhaps a slightly more “cost-oriented” version of the Inno Power I have, given some of the items outlined above. What really bothers me the most are the 1kV ceramic caps in the Y2 filter spots, as I think this may not pass UL tests. Then again, looking at the label, the UL logo looks very “botched” (to say the least) along with many of the other safety marks. Besides the ceramic “Y2” caps, everything else looks like it should pass UL tests easily. The strange part is that the Inno Power PSU also has the same dodgy safety marks, but that one does actually use proper Y2 and X2 caps throughout. I think Macron just didn’t do their label right, because IIRC, back when the UL site still had searchable listings, I do recall seeing the MPT model of PSUs there. Also, the background picture on the green box of the PSU above does show the Y2 caps as being ceramic ones (there’s the same picture on the back of the box, but more clear.) So it’s not like Macron was trying to hide anything – that background picture does appear to be of an actual Macron PSU. On that note, they do show a PSU on there with slightly better heatsinks (and a cut-out on the primary for a PPFC, I think.) But all in all, what’s shown and what’s inside isn’t too far apart. I just don’t understand the part with the ceramic Y2 caps. Definitely a goof up there!

    Components aside, here is a fan shot:

    Unfortunately, the promise on the box of a ball bearing fan was a false claim. This is a classic non-sealed Power Logic *sleeve bearing* fan. At least these are actually very reliable fans, especially with a fan speed control (which this PSU does have.) So I can’t say I’m that disappointed. The fan actually feels very solid.

    And to finish on a positive note, let’s see the solder-side:

    If there is one thing Macron got right, it’s their soldering, IMO. No complaints here whatsoever.

    Needless to say, this PSU will be recapped. I pulled most of the output caps out and checked them on the GM328 tester.


    Basically, only the CS-logo 10V, 2200 uF caps in 10 mm diameter turned out to be bad, as they all showed above-normal (40%+) capacitance, suggesting their electrolyte is starting to break down (and indeed that must be so, judging from the bulged one.) I expected to see the same with all of the other CS-logo caps, but surprisingly they still showed good capacitance and ESR. Time will tell how long those will last, though.

    The primary G-Luxon 470 uF caps are also still good (or at least the one I tested) - but that I expected. And they are not lying about their capacitance either.


    Now for the parts list…

    ICs:
    TL494 (PWM controller), LM339 (quad comparator used as output supervisor)

    Wiring:
    * 600V, 20 AWG input wiring
    * 300V, 18 AWG output wiring (except 4-pin CPU and AUX, which are 20 AWG)
    * Output connectors: 20-pin ATX, 4-pin 12V CPU, 4x Molex drive, 2x floppy, 1x AUX

    Primary Side:
    * Input Filtering: two X2-class caps (1x 0.47 uF, 1x 0.22 uF), two Y2-class caps (2.2 nF) and three 2.2 nF 1kV ceramic caps in Y2 spots, two CM chokes
    * Input protection: 6.3A 250V glass fuse (F-type), SCK 053 NTC inrush current limiter
    * 4x 1N5406 diodes for a bridge rectifier
    * 2x G-Luxon SM, 200V, 470, 22x40 mm, 105°C caps
    * 2SC5027F-R NPN BJT (TO-220) + 2sc945 BJT (TO-92) for 2-transistor 5VSB circuit
    * 2x 2SC5679 NPN BJTs (TO-220) in half-bridge configuration for main PS
    * 2x PCE-TUR, 50V, 10 uF, 5x11 mm caps for BJT drive coupling
    * 1 uF 250V PP metal film cap for H-bridge
    * main PS snubber network: 47-Ohm 3 Watt resistor + 4.7 nF 1kV ceramic cap
    * EI-33 main PS transformer, EE16 BJT driver transformer, EE16 5VSB transformer

    Secondary Side:
    * 5VSB
    *** 2x CS logo, 10V, 2200 uF, 10x20 mm with PI coil (5 mm core, 6 turns, 20 AWG)
    *** F06c20c (TO-220) fast-recovery rectifier

    * 3.3V Rail
    *** 1x CS logo, 16V, 2200 uF, 12.5x20 mm cap without PI coil
    *** 1x SBL1640CT (TO-220) Schottky rectifier
    *** dual mag-amp saturation toroids and no load resistor

    * 5V Rail
    *** 2x CS logo, 10V, 1000 uF, 10x18 mm caps without PI coil
    *** 1x CS logo, 10V, 2200 uF, 10x20 mm cap without PI coil
    *** 1x SBL3040pt (TO-247) Schottky rectifier
    *** 22-Ohm 2 Watt load resistor

    * 12V Rail
    *** 1x CS logo, 16V, 2200 uF, 12.5x20 mm cap without PI coil
    *** 1x F12c20c (TO-220) fast-recovery rectifier

    * -12V Rail
    *** 1x CS logo, 16V, 220 uF, 6.3x11 mm with PI coil before it
    *** 3x FR104 diodes (2x for rectification, 1x as a series voltage dropper)
    *** 620-Ohm 1 Watt load resistor connected to 5V rail (instead of ground)

    * -5V Rail
    *** 1x CS logo, 16V, 220 uF, 6.3x11 mm with PI coil before it
    *** 2x FR104 diodes for rectification
    *** 240-Ohm 1 Watt load resistor connected to 5V rail (instead of ground)
    Attached Files
    Last edited by momaka; 06-02-2021, 10:11 PM.

    #2
    Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - 5VSB circuit

    And here is the schematic for the 2-transistor self-oscillating 5VSB circuit of the Casing Power MPT-301 PSU:


    Unlike 2-transistor 5VSB circuits in other PSUs, this one doesn’t bias the main switch device, QA2, (an 2SC5027 NPN BJT) from the rectified +320V DC bus, but rather from the Live line through a resistor-cap series network (which I’ve never seen done before.) Thus, at first power-up, it looks like the main switch is pulsed ON and OFF a few times with whatever the line frequency is (50 or 60 Hz) until the primary-side auxiliary winding takes over. My guess for why this is so, is perhaps because there is also no negative feedback taken from the main switch’s Emitter resistor, R00. In other 5VSB circuits I’ve seen, there is usually a small resistor that gives negative feedback from that Emitter resistor to the Base of the driver switch device (in this case, QA1 – a 2SC945 NPN BJT) so that the main switch device can turn OFF and self-limit its switching until the primary-side aux. winding runs at normal steady-state. I think the implementation of the Casing Power 5VSB circuit might be more clever, though, because even in the event of an open primary aux. winding, the main switch device may self-limit in a more controlled manner than a constant DC-bias of a resistor, as done in other 5VSB implementations. Or maybe not?

    Also note the 2-transistor circuit of the Casing Power PSU doesn’t have a “critical” / startup electrolytic cap filtering the primary-side auxiliary winding. Therefore, this PSU may be less prone to over-voltage failure over time… or at least not from that component failing. However, I honestly don’t think anymore that any 2-transistor self-oscillating circuit is 100% safe, because as I’ve seen recently in another PSU, bad electrolytic filter caps on the 5VSB output side can still cause the 5VSB to go crazy and nuke itself. So as always, consider these 2-transistor 5VSB circuits somewhat more risky and less likely to go down nicely.

    In terms of load-testing the 5VSB, I didn’t go too much in-depth. With no-load, the 5VSB sits at 4.99-5.00V exactly, and the PSU draws about 1.8 Watts from the wall according to my Kill-A-Watt. That’s not bad for a 2-transistor self-oscillating design. In contrast, the Channel Well –built ADT MPT-301 (note it’s *not* a Macron-made PSU, despite what its label suggests) from this thread pulled 3.3 Watts from the wall with no-load attached – and that’s after the recap and removal of the 10-Ohm load resistor (which was replaced with 2x 100 Ohms in parallel.) Of course, the Casing Power 5VSB circuit does not have any load resistors on its output, so that probably contributed a fair deal as to why it pulled less power.

    With a 0.35 Amp load on the 5VSB, the Casing Power PSU drawing 5.3 Watts / 10.3 VA from the wall. Efficiency-wise, that comes out to a measly 33%... which, again, actually isn’t too terrible for a 2T-circuit. In fact, as a comparison, the recapped and modded CWT ADT PSU drew 5.8 Watts from the wall with the same load… or only 0.5 Watts more than the Macron. Not coincidentally, that extra 0.5 Watts is exactly how much the two parallel 100-Ohm load resistors draw @ 5V. So if we account for that, the 5VSB circuit in the Macron performs identical under load to that of the CWT.

    Next, I tried a 2.2 Amp load on the 5VSB of the Casing Power PSU. With that, the output voltage dropped down a bit to 4.91V, which is well within spec. Nothing inside the PSU got too hot after 5 minutes, except of course, the primary heatsink on which the 5VSB main switch is mounted to. That heatsink always gets quite warm, though (when the PSU is in soft-off mode), regardless if the 5VSB was loaded or not. I forgot to record how much power the PSU drew from the wall with the 2.2A load. But if I remember correctly, the efficiency increased somewhere to around 45-50%. Those are still pretty poor power conversion efficiency numbers, but normal for 2-transistor flyback designs. In contrast, modern PSUs (at least known good brands with 5VSB controller chips) will typically draw less than 1 Watt (even less than 0.3 Watts nowadays) from the wall with no load on the 5VSB and usually achieve no less than 50-70% efficiency at other loads.

    Lastly, it’s worth noting that the above tests were done after recapping the 5VSB circuit. That is, I put one new Rubycon YXJ, 10V, 2200 uF cap in place of the 1st filter cap on the 5VSB output filter and replaced the 2nd filter cap (originally a CS logo 10V, 2200 uF) with the 16V, 2200 uF cap from the 12V rail. This change was needed, because all of the CS 10V, 2200 uF, 10 mm caps were starting to fail, as mentioned above.
    Attached Files
    Last edited by momaka; 06-02-2021, 10:25 PM.

    Comment


      #3
      Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

      Next up: a few load tests…
      Test #1: ~6 Amps on 12V, 2.5 Amps on 5V, 1.6 Amps on 3.3V
      Test #2: ~12 Amps on 12V, 5 Amps on 5V, 1.6 Amps on 3.3V
      Test #3: ~1.7 Amps on 12V, 10 Amps on 5V, ~2.9 Amps on 3.3V
      Test #4: ~1.7 Amps on 12V, 20 Amps on 5V, ~2.9 Amps on 3.3V

      But before that, I had to do a quick recap as it just didn't seem like a good idea to load the PSU when there are bad caps in there. To keep the PSU as original as possible for the load tests, I swapped only a few caps in key spots.

      The recap consisted of the following:
      12V rail: CS-logo 16V, 2200 uF, 12.5x20 mm ---> (used) Nichicon PW with same V and uF.
      5V rail: 1x bulged CS-logo 10V, 2200 uF, 10x20 mm ---> Rubycon YXJ with same V and uF.
      -12V rail: CS-logo 16V, 220 uF, 6.3 mm dia.---> Nichicon PM 16V, 220 uF, 8 mm dia.
      5VSB rail: 2x failed (but not bulging) CS-logo 10V, 2200 uF, 10x20 mm ---> 1x Rubycon YXJ with same V and uF (for 1st filter spot) and 1x CS-logo 16V, 2200 uF cap from the 12V rail (for 2nd filter spot - I just barely managed to squeeze it in there.)


      Only after this I ran my load tests, while also abusing the PSU with a low AC input voltage (through a series heating element) to simulate worst-case AC line input scenario. These were the voltage output results:

      12V-heavy load (test #1)


      12V-heavy load (test #2)


      5V-heavy load (test #4)


      As you can see, the PSU did OK with the 12V-“heavy” load tests, but went out-of-spec on the 2nd 5V-heavy test (20 Amps on the 5V rail), where the 12V rail went up to 12.7V. So despite being an older PSU, it didn't do particularly well with a heavier load (100 Watts) on the 5V rail. However, with a 50 Watt load (10 Amps on the 5V rail), it was fine. Thus, it may be able to handle most retro 5V-heavy systems OK, so long as they don't load the 5V rail too much (or perhaps at least load the 12V rail a little more with several HDDs.)

      Efficiency-wise… this PSU seems to be pretty standard from what one could get with a half-bridge consumer design from the early 2000's – i.e. ranging between low and mid-70's % for most tests. AC input varied between 90-100V AC, depending on the series heating element used (700W or 1400W), but the AC from the wall was a steady 120V.
      Test #1 (with 700W series element on AC input): 91V AC to PSU and ~92.93 Watts on DC side… efficiency: roughly 75-80% efficiency
      Test #2 (with 1400W series element on AC input): 90V AC to PSU and ~177.43 Watts on DC side… efficiency: roughly 68.5-73% efficiency
      Test #3 (with 700W series element on AC input):… didn't record data
      Test #4 (with 1400W series element on AC input): 101V AC to PSU and ~133.05 Watts on DC side… efficiency: roughly 71-75.6% efficiency

      While doing these load and efficiency tests, on #2 and #4 I briefly switched the series heating element from 1400W to 700W, further dropping the AC input to see how the PSU would react. As previously found with my KDM PSU, it seems that these half-bridge PSUs don't actually shut down with a low line AC input, but instead continue to try working. This is obviously not desirable as it probably causes a lot of stress on the primary side (and hence why I run this test only briefly, as I haven't tried to see what would happen if I ran it for an extended period of time… but I will on a junker PSU some day. ) In the case of the Casing Power PSU here, on test #2, the 12V rail dropped to about 10.3 Volts. And dropping the AC input the same way for test #4, the 12V rail dropped from 12.7V to 12.02V – nicely in spec… until one looks at the 5V rail (where the load was), which dropped out of spec from 4.96V to 4.66V. In both of these cases, however, at least the PSU's Power Good logic operated correctly and went low to signal that it wasn't outputting correct voltages. It would be better, of course, if the PSU did that and then shut down. So I guess that's not just a quirk with the KDM PSU, but with all half-bridge PSUs, as I tested a few others, and they all acted exactly like this.

      On a different note, I also decided to pull out my type-K temperature meter and see how hot things got inside this MPT-301. Basically, I had the top cover of the PSU properly installed, but just without the screws. Then, after running each test for ~10-15 minutes, I shut off AC power, pulled the top cover off, and took measurements of various components of interest – namely the primary and secondary heatsinks, 3.3V and 5V/12V output toroids, 3.3V mag-amp saturation coils, and the output caps.

      What I concluded from that is the secondary heatsink was usually the hottest-running part (actually, the NTC thermistor on the AC input got much hotter than anything, but that's normal for these.) Of course, how hot it got depended on the load. For example, in 12V-heavy test #1, it only got up to 50°C. Meanwhile, the primary heatsink barely touched 40°C. And from the output inductors, really only the mag-amp saturation coils for the 3.3V rail got somewhat warmer than the other components, reaching around 45-50°C. As for the 5V/12V main toroid, 3.3V filter toroid, and output caps – nothing really got over 40-45°C (most caps remained cool at around 30-35°C.) For 12V-heavy test #2, the temperatures didn't really change much compared to test #1, except the secondary heatsink – it got considerably hotter (about 60°C) due to pushing the F12C20 rectifier for the 12V rail right down to its maximum rating at 12 Amps. 5V-heavy test #2 produced similar temperatures and the secondary heatsink again hovered in the low 60's °C due to the high current through the 5V rail rectifier. However, by far the worst was a modified version of 12V-heavy test #2, where I rewired my load tester and reduced the current on the 5V rail from 5 Amps to about only 2.5 Amps. Intuition might suggest that things should have ran cooler… but that' wasn't the case for the output toroid, which got mighty-hot due to the difference in cross-load currents and attempting to equalize the output voltages. It easily hit 65°C after only a few minutes of operation. So it's quite possible it could reach 70-75°C in a not-so-well ventilated case and after running for a good few hours. Goes to show what harsh cross-loading can do!

      Although my PSU “load tester” is still not finished (actually, not even past the beginning stage of construction ) and thus not able to load my PSUs more… I have to say I'm overall satisfied with how this PSU performed. All I need now is an oscilloscope to check ripple and noise output… but I don't know if I will ever do that. :\
      … and kind of sad to write all these walls of text and not even have a single scope shot, isn't it? On that note, I hope this isn't considered a complete waste of time and e-space. Recapping info/diagram coming next, so perhaps that may redeem things a bit.
      Attached Files

      Comment


        #4
        Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - recap & mods

        When it comes to recapping of this PSU, I’ll jump straight to the diagram here:

        It’s a very straight-forward PSU to recap and super-easy to work on, thanks to Macron putting connectors for the fan and AC input on the PSU board, making it easy to take apart.

        After doing the load tests above, I did a slightly more complete recap. Namely I replaced the single 3.3V rail CS-logo 16V, 2200 uF cap with a United Chemicon LXV, rated for 6.3V and 3900 uF. I figured the boost in capacitance could benefit the 3.3V rail. On the 5V rail, there are 3 caps originally: 2x 1000 uF and 1x 2200 uF. I had already replaced the 2200 uF CS-logo cap with a Rubycon YXJ. But then I decided to also replace one of the 1000 uF caps with another 2200 uF Ruby YXJ. So the total capacitance now on the 5V rail is 5400 uF (compared to 4200 uF before.) And finally, I also changed the CS-logo 25V, 470 uF cap on the secondary side auxiliary rail with a United Chemicon KZE of same voltage and capacitance ratings. All other caps are still the original ones, as I am currently out of stock for the small caps. On that note, in addition to the electrolytic caps listed in the 1st post, there are 3 more small caps on the secondary side for the logic circuits. These are, I think, 1x 10 uF cap, 1x 4.7 uF cap, and 1x either 1 uF, 2.2 uF, or 4.7 uF cap – all rated for 50V. Of course, 25V or 35V ratings should work too. So here’s a look at how my recap came out:




        As for modding, there are several things I did to this PSU. Starting with the 2nd picture above, I changed the no-load output resistors on the 3.3V and 5V rails. In particular, the 3.3V rail had none and the 5V rail had a 22-Ohm, 2W resistor. I added a 100-Ohm ½ W resistor on the 3.3V rail (since there is a spot for a resistor on the PCB.) Meanwhile for the 5V rail, I removed the 22-Ohm resistor and replaced it with another 100-Ohm ½ W part. So far, the PSU seems to be OK with this change – at least when running it with a load. Perhaps with nothing attached on the PSU outputs, the rails might go a bit “wonky”… but I am intending to use this PSU in a PC, so that shouldn’t be the case here.

        Also notice in the 2nd picture that one of the rectifiers now has a pink thermal pad behind it. No, I didn’t just do a thermal pad “upgrade”. Rather, I swapped the 12V rail’s rectifier (previously an F12C20C) with an STTH1602ct that came from this Delta DPS-300ab-15b PSU (along with the pink thermal pad.) So now this PSU could technically meet its current rating specs on the label for the 12V rail.

        Next on the list was that bent part on the back of the case shell near the fan.
        As posted above, the -before-:
        https://www.badcaps.net/forum/attach...9&d=1622692889

        And the -after-:

        It’s not perfect, but not bad either for the ~10-15 minutes of work I put in it (this was done back in March when it was still cold in the garage, so couldn’t spend too much time there without my hands freezing.)

        Then eventually I also added more ventilation holes to the case shell.


        Last but not least, I’ve always thought about doing this to my Inno Power MPT-301, but never really committed to it:




        It might not seem significant, but it does actually add quite a bit of surface area to the secondary heatsink (almost doubles it.) For good heat transfer, I did couple the two aluminum surfaces with white ceramic thermal compound. The screw on one end ensures that the two surfaces are tightly against one another. But for redundancy, that steel wire piece also adds pressure on the two heatsinks. I then did a 2nd quick re-iteration of the load tests, and the secondary heatsink ran about 3-5C cooler on average. Both the top and bottom part were at the same temperature, so the thermal transfer between the two is working well.

        And that’s all I did to this PSU… so far. It is boxed back up and ready for use. I have several systems in mind that this might go in.
        Attached Files

        Comment


          #5
          Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

          I'd guess the realistic do-it-all-day output power for that at around 200W. Unless the fan is providing a hurricane of cooling air, those heatsinks really are too insubstantial for 300W. The cheap fake Y-caps are glaringly obvious. The 470uF input lytics probably are about right for 250W, but heatsinks and the smallish main transformer core aren't.

          I didn't find a datasheet for the 2SC5679 switch transistors, but did learn that Fuji Electric discontinued them in 2013. Interestingly, the PCB is laid out to accommodate larger TO-3P or TO-247 transistors.

          I noticed that the box advertised that it had a ball bearing fan, but the fan is labelled "Sleeve Bearing".

          With your partial re-cap and some real Y-caps you might have a very decent 200W P/S capable of handling occasional surges to 250W. If its 5VSB really is capable of 3A, that was top end for when it was built and probably is still pretty typical.
          PeteS in CA

          Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
          ****************************
          To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
          ****************************

          Comment


            #6
            Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

            Originally posted by momaka View Post
            All I need now is an oscilloscope to check ripple and noise output… but I don't know if I will ever do that. :\
            … and kind of sad to write all these walls of text and not even have a single scope shot, isn't it?
            Thread fixed
            This is with a 10Ω load on 5VSB, so only 0.5A but still excellent results.
            Oh, never mind the PSU on the right, that is just acting the load as it was the first decent resistor I came across
            And momaka: great writeup, I did read it all because I noticed it is the same PSU that I posted about here, so that is what this scope shot shows
            P.S: Mine has legitimate Y caps and what looks like much larger transformers.

            Attached Files
            "The one who says it cannot be done should never interrupt the one who is doing it."

            Comment


              #7
              Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

              Originally posted by Per Hansson View Post
              Thread fixed
              This is with a 10Ω load on 5VSB, so only 0.5A but still excellent results.
              Thank you for posting this!
              Yeah, your 2x 3300 uF Rubycon MBZ caps will certainly keep the 5VSB circuit nice and quiet on ripple... though even the stock 2x 2200 uF are probably already doing a really good job. Most 5VSB circuits I've seen from Delta, HiPro, and LiteOn that are rated for 2-2.5 Amps typically have 1x 2200 uF and 1x 470-1000 uF. So 2x 2200 uF for 3 Amps on a flyback design seems quite appropriate, if not a bit on the generous side. Of course, the original CS-logo caps (or other crap stock brands that come with this PSU) don't have very good low ESR spec, so that's probably another reason for the higher capacity / bigger can.

              Originally posted by Per Hansson View Post
              Oh, never mind the PSU on the right, that is just acting the load as it was the first decent resistor I came across
              That looks like a BTX -model Deer/L&C... and probably one of their "better" builds, given that it has EMI/RFI input filtering.

              Originally posted by Per Hansson View Post
              And momaka: great writeup, I did read it all because I noticed it is the same PSU that I posted about here, so that is what this scope shot shows
              P.S: Mine has legitimate Y caps and what looks like much larger transformers.
              Nice!
              Going through that thread and the links in there, it seems a lot of other people had this unit or it's slightly updated brother based on the ATX9912p PCB. I guess it was fairly popular PSU back in the day... and not a bad one, save for the crappy cap choices, of course.

              Originally posted by PeteS in CA View Post
              I'd guess the realistic do-it-all-day output power for that at around 200W. Unless the fan is providing a hurricane of cooling air, those heatsinks really are too insubstantial for 300W.
              Agreed!
              I think Macron pretty much printed the same label for all of the MPT-301 units and called it a day, when in reality, some of the MPT-301's built with "lesser" parts like this one really should have been downgraded to 200W continuous / 250W max. I suspect this PSU might provide 300W peak for a few very short moments... but overall consider it risky. At low line voltage (and thus higher transformer currents), I suspect the main transformer may saturate and take out the primary or the primary may just go out by itself even before then.

              As for the fan - it does actually push quite a bit of air. I think it's running around 6-7V in idle and increases slightly when the PSU gets hotter (though I didn't measure how the fan voltage is impacted - perhaps a curious test for another day. )

              Originally posted by PeteS in CA View Post
              I didn't find a datasheet for the 2SC5679 switch transistors, but did learn that Fuji Electric discontinued them in 2013. Interestingly, the PCB is laid out to accommodate larger TO-3P or TO-247 transistors.
              Hmmm... perhaps a datasheet was never released on the internet? That means, if I had to guess, these are somewhat similar in specs to 13007's or a tier lower.

              As for the PCB layout... my other MPT-301 (the Inno Power) and the MPT-401 (400W Macron) both have TO-3P main BJTs. So this Casing Power is definitely a cheaper / lower-end version.

              Originally posted by PeteS in CA View Post
              I noticed that the box advertised that it had a ball bearing fan, but the fan is labelled "Sleeve Bearing".
              Yup.
              That's why I showed the box and the fan.
              At least the PSU wasn't terribly grossly over-rated, like some of the really nasty 100-200W -capable gutless wonders still sold today with the promise of being 600+ Watts. Ugh

              Also, that PowerLogic fan is actually a really decent fan - very quiet for the amount of air it pushes.

              Originally posted by PeteS in CA View Post
              With your partial re-cap and some real Y-caps you might have a very decent 200W P/S capable of handling occasional surges to 250W.
              Oh, I forgot to mention in the above walls of text on the recap - I did actually replace the non-approved ceramic "Y"-caps with real ones... or 2 out of 3, anyways. Basically, I replaced the one between Live and Ground and the one between rectified negative primary bus and Ground, as these are the ones that can present danger to the user in the case of a failed Ground and one of these caps being bad. The only original "Y" cap I didn't replace is the one between Neutral and Ground, since it won't present a safety issue - not unless Live and Neutral are switched... which won't be the case, as we have polarized plugs in the US. In Europe, though, with countries that use the Schuko plug - definitely all non-approved Y-caps need to be swapped. Interestingly enough, the EMI/RFI PCB soldered to the input power connector does contain 2x real approved Y2 caps.

              Originally posted by PeteS in CA View Post
              If its 5VSB really is capable of 3A, that was top end for when it was built and probably is still pretty typical.
              I think I did a quick and "unofficial" test on the 5VSB on my Inno Power PSU long time ago, and indeed it could provide 3 Amps.

              Comment


                #8
                Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

                Originally posted by momaka View Post
                That looks like a BTX -model Deer/L&C... and probably one of their "better" builds, given that it has EMI/RFI input filtering.
                It's actually the PSU from this old thread that I fixed last weekend and posted about here just now

                Originally posted by momaka View Post
                Agreed!
                I think Macron pretty much printed the same label for all of the MPT-301 units and called it a day
                Oh for sure, go back to that thread where I post this:
                "Also if you want to see something EXTREMELY funny look at this old post by me, of a PSU with an identical model number, but higher ratings on all the outputs..."
                "The one who says it cannot be done should never interrupt the one who is doing it."

                Comment


                  #9
                  Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

                  Originally posted by momaka View Post
                  This is obviously not desirable as it probably causes a lot of stress on the primary side (and hence why I run this test only briefly, as
                  I hope that when you this type of testing that you do a complete write up about how you did it

                  Originally posted by momaka View Post
                  I haven't tried to see what would happen if I ran it for an extended period of time… but I will on a junker PSU some day.
                  I would be very interested in the results of this type of testing
                  9 PC LCD Monitor
                  6 LCD Flat Screen TV
                  30 Desk Top Switching Power Supply
                  10 Battery Charger Switching Power Supply for Power Tool
                  6 18v Lithium Battery Power Boards for Tool Battery Packs
                  1 XBox 360 Switching Power Supply and M Board
                  25 Servo Drives 220/460 3 Phase
                  6 De-soldering Station Switching Power Supply 1 Power Supply
                  1 Dell Mother Board
                  15 Computer Power Supply
                  1 HP Printer Supply & Control Board * lighting finished it *


                  These two repairs where found with a ESR meter...> Temp at 50*F then at 90*F the ESR reading more than 10%

                  1 Over Head Crane Current Sensing Board ( VFD Failure Five Years Later )
                  2 Hem Saw Computer Stack Board

                  All of these had CAPs POOF
                  All of the mosfet that are taken out by bad caps

                  Comment


                    #10
                    Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

                    Originally posted by Per Hansson View Post
                    Oh for sure, go back to that thread where I post this:
                    "Also if you want to see something EXTREMELY funny look at this old post by me, of a PSU with an identical model number, but higher ratings on all the outputs..."
                    Hmmm, just stumbled upon this after trying to catch a few details of my recap...

                    So, the Q-TEK ADT-350 in the last link above - that's not actually a PSU made by Macron. It's a CWT ISO series, and probably one of their most gutless ones. See this thread on the CyberLink CWT-320ATX:
                    https://www.badcaps.net/forum/showthread.php?t=39222

                    Originally posted by sam_sam_sam View Post
                    I hope that when you this type of testing that you do a complete write up about how you did it
                    Will do for sure.
                    There's still a lot of things about it that I'm experimenting with, though. So perhaps when I get a solid method (and setup) of doing everything the same way every time, that's when I will post it. Right now, I'm still doing everything with a ton of jumper wires/clips and it just looks like a bird's nest, lol.

                    Comment


                      #11
                      Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

                      Originally posted by Per Hansson View Post
                      higher ratings on all the outputs..."[/I]
                      35a on the 5v rail... Deer/L&C anyone?
                      Main rig:
                      Gigabyte B75M-D3H
                      Core i5-3470 3.60GHz
                      Gigabyte Geforce GTX650 1GB GDDR5
                      16GB DDR3-1600
                      Samsung SH-224AB DVD-RW
                      FSP Bluestorm II 500W (recapped)
                      120GB ADATA + 2x Seagate Barracuda ES.2 ST31000340NS 1TB
                      Delux MG760 case

                      Comment


                        #12
                        Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3] - load testing

                        Originally posted by momaka View Post
                        Hmmm, just stumbled upon this after trying to catch a few details of my recap...

                        So, the Q-TEK ADT-350 in the last link above - that's not actually a PSU made by Macron. It's a CWT ISO series, and probably one of their most gutless ones. See this thread on the CyberLink CWT-320ATX:
                        https://www.badcaps.net/forum/showthread.php?t=39222
                        The fun part if it was not obvious was that these two PSU's had the same manufacturer on the type rating plate, and the worse one was rated higher!
                        "The one who says it cannot be done should never interrupt the one who is doing it."

                        Comment


                          #13
                          Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

                          Too bad they didn't resist the temptation to downgrade the parts inside the unit but still use the label that until now meant a lot to me.

                          BTW check the pdf attached
                          Attached Files

                          Comment


                            #14
                            Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

                            Originally posted by momaka View Post
                            This one should make @Pentium4 smile (if he is still reading BCN forums) – I got a new old stock / open box Casing Power MPT-301 PSU on eBay for $4 total.
                            Ha!!! Yes, I'm back on the forums at last. I cracked a smile at the title before even opening because I knew that a thorough post was awaiting

                            Originally posted by momaka View Post


                            Yes, it’s a very generic-looking box, but the manufacturer isn’t (at least not
                            Is that a Sparkle/FSP unit pictured on their box?

                            I also believe that the Power Logic sleeve bearing fans are surprisingly reliable. I've seen many Power Logic fans fail on video cards but I've seen their 80mm sleeve bearing fans that appear to be bone try but still spin well with no noise

                            In my experience with these units, I've actually seen the CS caps hold up the best compared to Fuhjjyu and GL.

                            Let me know if you need some of those low level 1uF/2.2uF/4.7uF/10uF 50V caps. I have so many that are 1-3 years old and I can throw some in the mail for you The BJT balance caps usually get toasted in these Macrons.

                            Awesome post! I have a couple Macrons in use still. I even got contacted by my old company when they went to upgrade RAM in one of the bench computers I built. I recapped an 11 year old (at the time) Macron back in late 2014/early 2015 and sent me a picture. It's still going to this day! They're excellent units. With fresh caps and an oiled fan, I have pure faith in them that they will just keep on working. They're also very easy to solder on.

                            Do you by chance still have that recapped Vantec (Macron) that I sent you a while back?

                            Originally posted by momaka View Post


                            It might not seem significant, but it does actually add quite a bit of surface area to the secondary heatsink (almost doubles it.) For good heat transfer, I did couple the two aluminum surfaces with white ceramic thermal compound. The screw on one end ensures that the two surfaces are tightly against one another. But for redundancy, that steel wire piece also adds pressure on the two heatsinks. I then did a 2nd quick re-iteration of the load tests, and the secondary heatsink ran about 3-5C cooler on average. Both the top and bottom part were at the same temperature, so the thermal transfer between the two is working well.

                            And that’s all I did to this PSU… so far. It is boxed back up and ready for use. I have several systems in mind that this might go in.
                            This is awesome!! I love it Definitely a plus side of having the rear 80mm exhaust fan, all that extra room in the case.

                            I appreciate all the testing you did with this unit. Definitely instilled more faith than I already had with these things

                            Comment


                              #15
                              Re: Casing Power MPT-301 [PCB ATX9806b-p REV: A3]

                              Originally posted by Pentium4 View Post
                              Ha!!! Yes, I'm back on the forums at last.
                              Pentium4!!!!!!!!!!!!!!!!!
                              Wow, you're back! Man, I was starting to get worried. Hope you're doing well and good to see you more often around here. Your motherboard recap posts are missed!

                              Originally posted by Pentium4 View Post
                              I cracked a smile at the title before even opening because I knew that a thorough post was awaiting
                              Yeah, sorry I can't help it. A PSU post with my name under the tittle is almost always guaranteed to be a super-long post - for the better or for the worst.

                              Originally posted by Pentium4 View Post
                              Is that a Sparkle/FSP unit pictured on their box?
                              That's what I thought at first, but no - that is an actual Macron PSU in the image on the box. Macron just used heatsinks that are very similar to Sparkle/FSP. But there's the same image on the back of the box with the same PSU... but bigger / with more details... and it's more clear from that one that it is a Macron PSU - transformers labeled with "MP" prefixes, the 1kV Y2 caps I mentioned in post 1 above , and the same layout on the PCB for the PWM controller & comparator ICs. I guess Macron just wanted to show how good the PSU can be built... but in my case with this MPT-301, it was a more cost-oriented version. Then again, back in the early 2000's when crappy Deer and PowMax PSUs were all the rage, this would have been considered a good tier above them.

                              Originally posted by Pentium4 View Post
                              I also believe that the Power Logic sleeve bearing fans are surprisingly reliable.
                              They really are - but that's the older 80 mm *open* -type sleeve bearing fans, like this one. On GPUs, I suppose PowerLogic went all "modern" now and builds their fans with a *sealed* sleeve bearing - this is essentially the same thing as a regular sleeve bearing, but you just can't get to the sleeve bearing anymore without drilling the plastic on the back of the fan. And even after that, it's tricky to service those. On top of that, your observation is absolutely correct: their fans fail a lot more on GPUs

                              Originally posted by Pentium4 View Post
                              In my experience with these units, I've actually seen the CS caps hold up the best compared to Fuhjjyu and GL.
                              Would you believe it that the few non-failed Fuhjyyu caps I pulled from the Inno Power MPT-301 PSU are still reading in spec. A few of them I re-used to recap a crappy old Deer, and those are still good too.
                              I suppose anyone's MMV when it comes to crap caps. But you're right, CS-logo caps aren't all that terrible. GL, on the other hand... - they never last. JEE, who else is that bad?

                              Originally posted by Pentium4 View Post
                              Let me know if you need some of those low level 1uF/2.2uF/4.7uF/10uF 50V caps. I have so many that are 1-3 years old and I can throw some in the mail for you
                              Thanks for the offer!
                              However, it looks like I might need to do an order of my own soon, as I may need >100x from some of the common values (like 47 uF 50V). Basically, I have an audio amp and a few CRT monitors that are shaping up to need a full recap. Add to that a dozen or so PSUs and some other electronics that I've been trying to eradicate from my "fix" pile... and, it's probably going to be a pretty big caps+parts order.

                              Originally posted by Pentium4 View Post
                              The BJT balance caps usually get toasted in these Macrons.
                              Noted.
                              I'll keep an eye on mine or replace them next time I open it.
                              IIRC, I pulled one of them back when I was recapping this unit, and it measured in spec, so I put it back in. Then again, this PSU pretty much has 0 hours on the clock.

                              Originally posted by Pentium4 View Post
                              I even got contacted by my old company when they went to upgrade RAM in one of the bench computers I built. I recapped an 11 year old (at the time) Macron back in late 2014/early 2015 and sent me a picture. It's still going to this day!
                              That's so awesome to hear!
                              Was that the Uptek/Uptime Tech. (the exact name eludes my mind at the moment) place? I remember seeing a sticker on one of the motherboards you sent me... this one, which is now sitting in a PC in the guest / spare room in my parent's house. It's a bullet-proof system.
                              As for recapped Macrons... yeah they may not be the newest design on the block or very efficient or offer a ton of features... but they are simple and indeed bullet-proof too, with a recap.

                              Originally posted by Pentium4 View Post
                              Do you by chance still have that recapped Vantec (Macron) that I sent you a while back?
                              You mean the regular (non Vantec?) Macron MPT-401 you sent me a while back?
                              - Yeah, of course I have it! It's sitting just a few inches away from my left foot right at this very moment - installed in a Gateway GT-5656 PC under my desk. I put it in there around 2017, I think. That PC is essentially one of my "for-goofing-on-the-internet" systems and used semi-frequently (typically winter and spring months when it's cold in the house.) I put the Macron in that PC, because its ECS motherboard (particularly GeForce 6150 MCP) was giving me grief with crashing. So to make sure the issues weren't from the original PSU, I figured no other PSU would be more stable than the recapped Macron you sent me.

                              Originally posted by Pentium4 View Post
                              This is awesome!! I love it Definitely a plus side of having the rear 80mm exhaust fan, all that extra room in the case.
                              Yup, and the heatsink addon definitely works. In fact, it works a little too well, because the speed of the fan is based on the heatsink temperature, and I notice it runs slightly slower now. On the minus side, the output toroid gets a little warmer because of that.

                              Originally posted by Pentium4 View Post
                              I appreciate all the testing you did with this unit. Definitely instilled more faith than I already had with these things
                              Me too.
                              For an old group-regulated PSU, these hold up admirably well with a heavy cross-load like I was doing, which is actually the hardest on the output inductor, due to imbalanced 5V & 12V currents. If I balance the currents on 5V and 12V rails, the output inductor runs noticeably cooler, even at higher currents.

                              Comment

                              Working...
                              X