Antec SL350P PSU

[Th3_uN1Qu3]

Re: Antec SL350P PSU

red-grey-red = 2.8k ohms -> not a standard value
brown-grey-brown = 180 ohms -> a standard value

The resistor is bad.

Ever thought it could be brown-grey-red which is PRECISELY 1.8kOhms?

[goodpsusearch]

Re: Antec SL350P PSU

Alternatively, you can bypass the whole active PFC section, by temporarily soldering a bridge rectifier to the primary capacitor. Then wire the AC input cord to this bridge rectifier. You can use the series lightbulb here (not only you can, it's recommended that you do). This way you can find out if the power supply itself is alright.

+1

You can remove that add on APFC board and put a bridge rectifier in its place and see if the psu works. Don't worry about the small red cable. It is needed for the APFC board to work, not the power supply.

[goodpsusearch]

Re: Antec SL350P PSU

https://www.badcaps.net/forum/showthread.php?t=8385

https://www.badcaps.net/forum/attach...1&d=1272611235

It is very easy to be bypassed.

https://www.badcaps.net/forum/attach...1&d=1276531416

https://www.badcaps.net/forum/attach...2&d=1276531416

Here you can see the small red cable.

[amartolos]

Re: Antec SL350P PSU

Ever thought it could be brown-grey-red which is PRECISELY 1.8kOhms?

Ah, these resistor colors... I would rather say the first and third band have the same color, but it is not the first time I have a hard time distinguishing between red and brown. A picture with better color balance was shown by goodpsusearch here, and to me it seems that the two bands are also the same. On the other hand, the values are too close for just a coincidence... I will take the resistor and make a poll around here about the colors

[Toasty]

Re: Antec SL350P PSU

Take the board out in the sunlight and examine it closely. TU is correct. It is more than a coincidence that it would be brown-grey-red making it the 1.8kΩ you measured. I thought you were convinced the colors were the same, hence my post about replacing it.

Toast

[amartolos]

Re: Antec SL350P PSU

Well, if I look in sunlight or incandescent light, the two bands do seem to be a tiny bit different, and I mean tiny. So I will leave it like this.

In the mean time I installed a rectifier bridge in the place of the APFC add-on board (it's actually very easy, as godpsusearch pointed out) - some filtering capacitors will be left out but I guess it's no big problem for testing. The PSU behaves in the same way: standby with 5.12 V, but with low output voltages when running with a small load: 7V instead of 12V, 3V instead of 5V. The mains caps charge at 310V in standby and 105V when powered up (slightly higher than with APFC). Also, just as before, the series bulb flashes just briefly when plugging in in standby, but lights like a ~40W bulb with PSU switched on.

I started to check the components in the primary side, and I got to an 8-pin IC situated close to the edge of the PCB, next to the big transformer, right near the red wire that goes to CON101 of the APFC board. It should be visible in the pictures from one of the first posts. Its surface looks partly darker, so I took it out and tried to make some close-up pictures. I put the light source in such a way to enhance the color difference. It's a UC3844B, a PWM controller and it seems to be available here.

Please have a look and tell me what is the next step. Thanks!

[Toasty]

Re: Antec SL350P PSU

I would replace it. But, trace it's power supply back (see datasheet) and see what drives it. It could be a transistor(s) ahead of it that has failed also.

Check the main switchers that drive the 5/12v section on the input side of the big transformer, and the associated dual-Schottky diodes on the output side of same.

The low voltages are from the lack of boost from the APFC and there might be some overload to make the bulb come on so brightly. It needs to have ~320v or so across the mains caps to run correctly.

You might try 2 - 100w bulbs in parallel. Sometimes these units take a big bump of current to fire up and then back down quite a bit when running with no load.

Toast

[Th3_uN1Qu3]

Re: Antec SL350P PSU

You might try 2 - 100w bulbs in parallel. Sometimes these units take a big bump of current to fire up and then back down quite a bit when running with no load.

Yes but this doesn't impact in any way the running of the supply, it simply delays its startup. Power supplies without APFC (or with the APFC bypassed as is the case) should start up even with a 40W bulb.

What load are you trying on it? If it's too high it's entirely normal for the thing to light up the bulb and have low output. But i suspect it's something else, especially since you have mentioned that the primary voltage drops significantly when the supply is started up. This is *not* normal and indicates a partial short. If you'd ask me as to what may be shorted, i'd check the secondary diodes first, then if all test ok look at the primary.

The UC3842 is likely alright because the supply powers up and runs to some degree.

[Toasty]

Re: Antec SL350P PSU

>>Power supplies without APFC (or with the APFC bypassed as is the case) should start up even with a 40W bulb.<<

I disagree as I do this daily and run into it all the time. PFC or not.

The fact that it's just glowing (with no load) and not flashing (as in trying to start) does likely mean something is awry.

Toast

[PCBONEZ]

Re: Antec SL350P PSU - pictures

Hi,

I did not find ESR values for neither TMR nor TNR caps (I looked in datasheets for TN and TM which are listed as general purpose). As for the ripple current, it is comparable, maybe better for the PWs.

Nope.
Only one of those two is GP - and they shouldn't be.
The other has ratings FAR better than PW.

For specs:
TN = TNR
TM = TMR
The R isn't part of the series name.
It's a marketing thing that I think denotes a special ordered size or a corporate client.
.

[amartolos]

Re: Antec SL350P PSU - pictures

Nope.
Only one of those two is GP - and they shouldn't be.
The other has ratings FAR better than PW.

For specs:
TN = TNR
TM = TMR
The R isn't part of the series name.
It's a marketing thing that I think denotes a special ordered size or a corporate client.
.

Then I got it wrong at some point. You're right, only TN is listed as general purpose, TM being advertised as high-ripple, low impedance. They don't list directly the ESR value in the datasheets, although in the same sheets the caps are presented as low-ESR, so I took the impedance at 100 kHz (for a capacitance of 2200 uF, Z≈R at this frequency) and compared it with the impedance of PW's. It is the same, 0.03 Ω at 100 kHz, so I thought it should be safe to put Nichicon PW. But I see now that the dissipation factor for TM is 0.15, compared to 0.2 of PW, both measured at 120 Hz, implying an ESR which is 1.3 higher for PW's.

Could this be the reason for the PSU failure after recapping? I can put back the TMR to see if there's any difference. But I'd still say that it's the specifications at 100 kHz that matter most for a PSU, because it's closer to its working frequency.

Actually, in the end I used Panasonic FK instead of Nichicon PW (size compatibility), which have an impedance of 0.025 Ω at 100 kHz, but the dissipation factor is still 0.2 at 120 Hz, so the problem remains. The ripple rating of the Fuhjyyu's (TM:1945mA@100kHz, TN: 960mA/120Hz) is lower than PW (1945mA/100kHz) and FK (2140mA/100kHz).

[Toasty]

Re: Antec SL350P PSU

Just a thought... Check all the caps for correct polarity please. One backwards could be the cause of all this.

[amartolos]

Re: Antec SL350P PSU

Thanks for the hint! I checked them several times, both against the silkscreen and the original pictures. -sigh- I wish it were that easy...

BTW, noob question: If I check polar caps for shorts with a DMM in diode test mode, and I apply reverse voltage... could it damage them? I would say yes, but in this case, up to what nominal voltage of the cap?

[Th3_uN1Qu3]

Re: Antec SL350P PSU

No, it won't damage them. The current provided by a DMM in diode test mode is usually 1mA. Small enough to not damage anything.

[Toasty]

Re: Antec SL350P PSU

No, not enough voltage or current to damage them from DMM.

This was working before the recap, yes? (Just causing issues with the system.)

Check -carefully- for solder bridges or a bad (cold) solder joint. Easy enough to put a blob of solder in the wrong spot. There is no pic of the underside.

Toast

[amartolos]

Re: Antec SL350P PSU

Yes, it was working before, it only had the 4700uF/10V cap bloated, and the bluescreens were most likely caused by the mainboard (several bloated/leaking caps, meanwhile replaced, it still loads the BIOS screen after this but no time to test it further). I will also post a pic of the solder side later when I get home.

[Toasty]

Re: Antec SL350P PSU

Ok. Once you double (triple) check all the connections, I suppose it's time for the fire test. Take the lightbulb out of the circuit and with a loooong extension cord, stand waaaay back and plug it in to full power. Fireworks at this point are -not- good.

You've wondered what the helmet's for, eh?

Toast

[amartolos]

Re: Antec SL350P PSU

Just some updates, no time yet for the fire test.

I got a new UC3844 and also traced back its supply on the board. It is the same 19V as for the APFC chip, and the PNP transistor that drives it reads OK.

Since my desoldering skills are improving, I also removed and checked the output Schottky diodes, which respond as expected.

The load that I used for testing was 3.9 Ω for 3.3 V, 1.9 Ω for 5 V, 5.1 Ω for 12 V (the sticker specifies minimum loads of 0.5 A, 1.5 A and 0.8 A respectively). This should give a total load of 44 W, but due to the reduced voltages I had it was rather 16 W.

Tomorrow I will put everything back, cross my fingers, get the helmet and... should you notice a boom, complete with mushroom cloud, you know where it's from.

I also took some pictures of the board with the heatsinks removed. I tried not to use the flash, so the lighting is not the best.

[Toasty]

Re: Antec SL350P PSU

Good. Keep your sense of humor about this and it will work out.

Another member had the *BOOM* occur and he was suspected of causing an airliner crash that day.

Nice pics too.

I don't see anything blatantly wrong. But Antec (and others) have this horrible habit of leaving component leads long and then bending them over neighboring traces. A little bump while working on them can result in a short to the next trace over. Take a real close look at all of those on this board. There are dozens.

I use a very sharp pair of lead nippers to remove all of those. Lacking those, a nail clipper will do nicely also.

Please advise what time you will be doing the fire test...
I want to be sure to be looking West to avoid the flash.

Toast

[amartolos]

Re: Antec SL350P PSU

Aaaaand... it works!

It seems that the light bulb in series was at the root of problems after all. After replacing the UC3844 the symptoms were the same as before, with the bulb still connected, but as I removed it, everything went back to normal. It still makes a very faint buzzing sound, but I'll see if inside a computer case it's still audible.

So now, thinking back, the first failure (the 2SK2996 FET from the APFC circuit) was probably caused by my careless soldering, but I still cannot explain why, after replacing the FET, at the first power-up attempts the fans were not moving and the output was 0, then at some undefined point they started to spin.

The next step is to apply the DM311 mod for the 5Vsb, but this should be really easy, monkey see - monkey do, I should just follow step by step the instructions from the thread I linked to. The parts are already here. I will come back when it's done, maybe with the final list of cap replacement, and if I feel diligent, I will even take the PSU to check with a scope.

Thanks to all who helped me throughout this thread!

PS: I had no helmet in the end, but I did wear a pair of goggles. Incidentally, they are laser protection glasses, which I thought would be handy to at least protect me from UV light if sparks should occur...