ThermalTake Smart M Series M850W – blown primary cap and fuse

[momaka]

I recently acquired a ThermalTake Smart M Series M850W (model SP-850AH3CCB) power supply. It has a blown primary cap, blown fuse, and shorted 5VSB MOSFET. Now, if the primary cap was a cheap Chinese or Taiwanese brand, I wouldn’t have been surprised. But the primary cap is a Panasonic HD 400V, 470 uF cap and the rest of the PSU also has quality caps (mostly Chemicon KY and KZE). The PSU barely has any dust in it as well. It looks almost new, actually.

So, has anyone seen this type of failure before? Did the APFC overshoot and cause the 5VSB FET and fuse to blow? Or was it the other way around, with the 5VSB FET blowing and causing the APFC to overshoot and kill the cap as well?

The APFC controller is a CM6800TX (which is also the PWM controller). PSU is a double-forward design. The APFC MOSFETs (two Infineon IPA6R190E6 in parallel) are NOT shorted. PFC diode (QH08t7600 ?) is also OK. Main supply switching transistors are a pair of Infineon IPA6R190E6, and measure fine too. Bridge rectifier and NTC thermistor seem to be fine too. None of the fusible resistors are blown either…. So in other words, pretty much everything appears to be okay, except for the blown 5VSB FET, main cap, and fuse.

I looked up some reviews on this PSU, and looks like it’s a Channel Well / CWT design. The PSU board says ThermalTake on it, and there is also a model number G08-CPUQ161-PG00 REV:1.2). Here is one review I found for it as well:
https://www.hardocp.com/article/2014...upply_review/2

By the way, this PSU also has a 5-year warranty. But I didn’t really look into the label and tore into it after I saw that it didn’t work So I guess I can’t RMA it for warranty (then again, I don’t have a receipt for this either, as I got it for free from the trash, so maybe I would have never been able to send it for RMA anyways ).

And before I forget… pictures attached at the end of the post.

For starters, I'm thinking to swap the 5VSB FET, replace the input cap. disable the APFC, and seeing if the 5VSB comes up. I guess stay tuned for now. Or if anyone has better suggestions, let me know.

[Wester547]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

That primary capacitor doesn't quite look like it blew up, but it's definitely leaking at the top. Maybe some idiot dropped a heatsink or something on the vent during assembly (before the QA sticker was placed on the vent) and it caused a crack at the seams of the vent which allowed for the electrolytic fluid to leak (at the top) over time, the capacitor to fail, the fuse to blow, and the +5VSB FET to die accordingly. Of course that's just my conjecture. Your troubleshooting methods sound good and par for the course. The +5VSB FET could have died and caused all the aforementioned failures as noted. Or the capacitor could have failed if the PSU had to handle a nasty transient after a long time (a couple years or more in standby, going by the manufacturing datecodes visible in the PSU, it was made in late 2012), or many times over. I know when CapXons have failed silently in the primary APFC filter of Delta PSUs that they've caused the main fuse to blow and the APFC FET to fail subsequently, along with some other components.

Any liquid electrolytic will fail if abused regardless of brand or series. All liquid electrolytics have an finite lifetime, all manufactures have come forward and stated this. Of course I am not stating that all brands and series are created equally. Just that nothing is impervious to failure. On an ending and slightly-less-useful-note, non-APFC PSUs seem less prone to failure which is why I will always prefer them...

[momaka]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

Welp... I played around with it today and got some news.

First off, that primary cap isn't just bulging and leaking - it is a completely SHORT-CIRCUITed!

That said, once I removed it out of the circuit, next I started to wonder how to remove the 5VSB FET. The screw holding down the FET to the heatsink was locked on so tight I could not get it to turn at all. I was literally about to start cutting and chiseling things away, until I decided to measure the primary cap. And sure enough, once I found that the primary cap was shorted, I then proceeded to re-check the FET again. Turns out the 5VSB FET is NOT shorted after all .

So next, I pulled my gutless Raidmax RX380-K remains out of the closet and hooked its primary side to the primary side of the ThermalTake PSU. Next, I did the incandescent bulb trick to the live line going to the Raidmax PSU - that way, if anything shorts out (be it on the crappy Raidmax PSU or the ThermalTake), I wouldn't have to pick out small glass shards from my carpet when the fuse blows too.

Once I did that... NOTHING happened. The 5VSB on the ThermalTake still appeared dead. However, further investigation revealed that the ThermalTake PSU has an AC detection circuit. With no AC detected, the whole PSU was disabled. So rather than wiring the 340V primary bus (from the Raidmax PSU) to the primary side on the ThermalTake, instead I wired it to its AC input (after the fuse, of course, as that was blown).

As they say, third time is the charm. (Third time, because the second time I forgot that I had removed the voltage doubler switch on the Raidmax PSU a long time ago, and it was only outputting about 165V DC.) I proceed to power on everything, and the 5VSB on the ThermalTake PSU came alive.

However, when I tried to run the main supply by shorting PS_ON to ground, I got nothing. So perhaps either the low voltage on the primary side (due to no APFC) or the fact that I didn't have a proper load on the output made it not want to start at all. But I will work on that next. I guess the next step is to find a 400V or higher cap and see what the APFC does... either that or try to disable it and run the PSU on a 120-240V step-up transformer... or have two 200V caps in series with a voltage-doubler circuit.

More to come later. Stay tuned .

[Stefan Payne]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

When you see primary cap blown, always expect some major damage on the primary side...

Even the worst of the worst should last for eternity - except if the line voltage is shitty and you have many brownouts or surges...

But the primary cap is a Panasonic HD 400V, 470 uF cap

THAT's just Bullcrap. An 850W should at least have a 680uF Primary...
How's the holdup time of that thing? 8ms Tops??

[stj]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

check the input rectifier.

[Sparkey55]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

When you see primary cap blown, always expect some major damage on the primary side...

Even the worst of the worst should last for eternity - except if the line voltage is shitty and you have many brownouts or surges...


THAT's just Bullcrap. An 850W should at least have a 680uF Primary...
How's the holdup time of that thing? 8ms Tops??

For that wattage and above I would prefer 1000uf 400v-450v.

[Sparkey55]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

check the input rectifier.

Please do.

[kc8adu]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

i am in the cap was damaged during assembly camp.
overshoot causes massive overheat and violent venting.
i will post pics of what that looks like.
i have caps from a yaskawa vfd that got a massive overvoltage.
what a mess!
the secondary side logic probably blocks startup if the pfc chip reports a fault or pfc is low.uvlo.

[momaka]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

Another update...

Everything on the primary side tested good, so I proceeded to install a 450 V, 500 uF, Nichicon LQ cap temporarily.

A bit of a hack-job, but good enough for testing open-frame


First, I tried running the PSU through a 120-240V step-up transformer (in series with a 115V, 470 Watt heating element as current-limiting), hoping that I could get away using the PSU without APFC... but I got same thing like last time - only 5VSB present, but can't turn ON the PSU. Voltage on primary cap was about 325 Volts DC (as expected, due to drop across heating element, and thus across my step-up traffo.)

So I decided to test APFC and see what happened. To do that, I simply installed back the jumper for the APFC that I had removed before and tested again on 230/240V AC through the transformer. Plugged in and got 325V DC across main cap again. 5VSB came up again. Then I pressed the Power button on my test motherboard that I had hooked up to the PSU and...
SUCCESS!
- The PSU worked just fine. Voltage across primary cap read about 375 Volts.

Next, I tried running the PSU directly on 120V AC (through the series 470 Watt heater, of course, as I still don't have a fuse on it)... and it still worked fine. Primary cap voltage was again about 375V, but it took a bit longer to charge at 120V AC.

On that note, the APFC does seem to make a bit of a tick and hiss noise when the PS_ON is pulled low. But I don't see any overshoot, so I imagine that is normal.

Anyways...

i am in the cap was damaged during assembly camp.

Looks like that might have been the issue indeed.

I guess now I need to hunt down Digikey or Mouser for a large cap... and a fuse. I thought I had bought some 10 Amp ceramic slow-blow fuses from Radio Shack... but I my memory has failed me - they were 5 Amp slow-blow fuses. I might still just use it, though, as I am planning on using that PSU in a PC that will draw no more than 200-250 Watts, if even that much.

overshoot causes massive overheat and violent venting.
i will post pics of what that looks like.
i have caps from a yaskawa vfd that got a massive overvoltage.
what a mess!

Can't wait.

the secondary side logic probably blocks startup if the pfc chip reports a fault or pfc is low.uvlo.

Actually, it seems like the primary side is the one that monitors that, as there is a trace from the (+) primary bus going to the daughterboard that holds the CM6800 chip and also another chip. One of these probably handles the APFC UVLO and other protections.

If the APFC had blown components, I would have probably taken off that daughterboard and tried to disable the APFC monitoring completely. But since everything is working, I didn't want to bother.

THAT's just Bullcrap. An 850W should at least have a 680uF Primary...

Yeah, you can see the board has silkscreen for a much bigger cap. I'll see what I can find and fit on there.

How's the holdup time of that thing? 8ms Tops??

Actually, it's close to none at all. The AC detection circuit kills off the 5VSB and the main supply almost as soon as the AC is disconnected. I reckon the AC detection probably has a pre-set delay before it kills the PSU, so the size of the primary cap would be irrelevant for hold-up, it seems. When I disconnected AC power to the PSU while it was powering my motherboard, the PSU shut off almost instantly. I've seen my 185 Watt Bestec (no PFC) with two 200V, 470 uF primary caps have longer hold-up than that on the same motherboard.

The primary cap on the ThermaTake PSU remained charged to about 290 Volts after I disconnected power above, so I think the PSU could have kept going until the voltage on the primary cap dropped much lower... but it didn't.

In any case, I will try to find a bigger primary cap for it (capacitance-wise). I'm also thinking of getting an MOV or 450V diode to put across the cap, so that if there is over-voltage, it will be shunted and not take my new cap with it.

[Sparkey55]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

My way of thinking is that should have been a 450v rated cap from the start. To me 400v seems just too marginal.

[Stefan Payne]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

Hm, maybe. But 450WV caps are way bigger and more expansive...
But the surge suppression things are before the caps. You have a boost circuit before it...

[kc8adu]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

poof!
i still need to take pics of the exploded cap.
one of these vfd were too far gone as seen from fire damage.
the other one is fixed and back in service.

[stj]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

what a fucking mess.
and the vdr's should have been in heatshrink anyway!

[Quaddro]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

Damage in primary side is a lot easier to handle rather than secondary..

The one causing breaker to trip is always my favorite..

[momaka]

Re: ThermalTake Smart M Series M850W – blown primary cap and fuse

poof!
i still need to take pics of the exploded cap.
one of these vfd were too far gone as seen from fire damage.
the other one is fixed and back in service.

That must have sounded and smelled nice.

We had one of those high-end GPUs (I think a GeForce GTX 1080 Ti) short out and take a dump in the shop. PSU was a high-power one, meant for multi-way SLI, as the system had another 1080 Ti in it too. The bad GPU shot quite a bit of smoke before we cut the power from it. Would have probably kept going too.

Took a lot of fans and air out the place.

[momaka]

ThermalTake Smart M Series M850W – temp repair update 2022

I guess I never posted an update on this one… though, it’s not like I changed anything too much from the original “repair” last time.

Still, just for completion of the thread (or until I ever get a proper replacement for the primary cap)…

First, I added a film cap in parallel with the big cap to help out with the high frequencies from the APFC.
https://www.badcaps.net/forum/attach...1&d=1647553736
Not sure how much it helps, but I noticed some Delta high-end workstation PSUs have a similar arrangement. So I figured why not. And yes, the primary cap is still sitting on those ugly copper wire “legs”… which present some additional “issues”, but I’ll get to that in a second.

Next, the “fuse holder”:
https://www.badcaps.net/forum/attach...1&d=1647553736
OK, let’s not comment on it too much. All I will tell you, it’s not made from premium materials and it helped me clean up some scraps from my workbench. In all honesty, it’s probably not too safe, so I will replace it someday when I open the PSU again. I do have proper fuse holders from other scrap electronics now. As for the fuse: it’s a slow-blow 5 Amp glass type, IIRC – not quite enough for the rated power on the label of the PSU, but enough for my needs currently.

And now back to those long leads on the primary cap I was talking about… yeah, I couldn’t put the fan inside the PSU’s case because of this.
https://www.badcaps.net/forum/attach...1&d=1647553736
Not that I would have been able to do it anyways. Even if I shortened the leads completely, the Nichicon LQ 450V 500 uF cap I used is still a few mm longer than what can fit in the PSU. So it’s really just not the right cap for this PSU. But that’s OK, as I don’t really have any PC builds that require a power supply with this kind of power output. In fact, I’ve been using this PSU mostly just for temporary testing of high power GPUs, which there hasn’t been a lot of on my end (very few, actually) – especially in the last 2-3 years, given the craziness around the GPU market (if we can even call it that.)

So the “temporary repair” seen above has been lingering around for 5 years now (since 2017)… and no specific future plans for it to change. I regularly see used 500-700 Watt PSUs of similar caliber (EVGA, TT, PowerSpec, and etc.) on my local classified for no more than $20-30 usually. So spending close to $10 on a proper primary cap just to fix up this PSU (as it should be) isn’t overly-worthwhile IMO. I’d rather keep it as it is currently… and when the need arises to fix it up properly to put it in a PC/case, then I will do that. So case closed on this one for now – figuratively speaking, of course.