Viper22A 5VSB circuit

[прямо]

So I have a cheap non working ATX PSU that I was learning to repair a decade ago. At the time, it blew the main fuse, bridger rectifier, NTC, and primary 9A 900V MOSFET. Replaced all except the MOSFET. 5VSB came back online. Then I poked around in it so much, measuring components one by one to a point I accidentally made the 5VSB circuit primary side went bang. Blown the AP8022 (Viper22A) PWM chip, along with a low resistance resistor and the PC817 opto isolator. I replaced them all.

In the process of poking around, I also lost a zener diode that stabilize the voltage coming from pin 4 of the 5VSB opto isolator, and this voltage happens to be a part of the main signal PWM switching circuit based on TL3845P. I have no idea what zener it was.

I stopped working on it. The board sat on my shelf for 12 years, lol.

Anyway, what I want to understand is, without the Viper22A PWM chip, I get -165V DC on its feedback and supply pins. Is this correct? Shouldn't I be getting like 15V DC, or is this only happening when the chip is installed?

Here are some close up photos of the circuit.

I understand that I can just bypass the entire circuit by de-populating it, and use a 5V 3A charger module tapping directly into the high voltage instead to get 5VSB.

[temyth]

without the zener the voltage can't be controlled so it goes too high. most part of zeners in the primary are 1N4148, try putting one and see if it changes.

[temyth]

without the zener the voltage can't be controlled so it goes too high. most part of zeners in the primary are 1N4148, try putting one and see if it changes.

[прямо]

Cool! I did what you suggested, and 5VSB returns.




Thank you!

[прямо]

A side note,

I have been repairing SMPS with a single primary cap for too long that I subconciously treating these ATX PSUs with two primary caps like it only has one cap. I did the most stupid thing: I put my negative probe at the junction of the caps!

which is why I got a false reading of -165V DC when in fact, without the missing zener installed, it already has the correct voltage for the PWM chip, albeit not stabilized.

But when I edited the photo, I knew exactly where the correct hot ground is, ahahahaha.

[прямо]

Had to disable the fault protection to find out why it can't start.

12V @ 16.7V
5V @ 7.2V
3.3V @ 3.3V

I guess I have to check the feedback now, but at least the PSU is working again, just out of specs.

EDIT:
Above is with the light bulb current limiter.
Without it, I get:

12V @ 37V
5V @ 15V
3.3V @ 4.2V

[temyth]

It happens to everyone XD. Sometimes, I'm repairing computers and I forget to plug in the power cable, and so on, hahaha. Could you send a photo of the entire PSU for me to see the complete circuit topology? Regarding the unstable voltage on the output, try testing the comparators on the secondary—they often go shorted and activate the protection in the PSU. They are mostly TL431.

[CapLeaker]

Had to disable the fault protection to find out why it can't start.

12V @ 16.7V
5V @ 7.2V
3.3V @ 3.3V

I guess I have to check the feedback now, but at least the PSU is working again, just out of specs.

EDIT:
Above is with the light bulb current limiter.
Without it, I get:

12V @ 37V
5V @ 15V
3.3V @ 4.2V

Waaah!? You get 37V out of the 12V rail? Either your DMM is fubar or measured on the wrong point. You probably couldn't get the PSU that high if you would wanted to by disabling the feedback and let it go full bore without regulation. I've seen many things, but tripling the output voltage from what it is supposed to be. You should have been wearing the caps as they are not rated for that voltage. This doesn't make any sense whatsoever.

[temyth]

Waaah!? You get 37V out of the 12V rail? Either your DMM is fubar or measured on the wrong point. You probably couldn't get the PSU that high if you would wanted to by disabling the feedback and let it go full bore without regulation. I've seen many things, but tripling the output voltage from what it is supposed to be. You should have been wearing the caps as they are not rated for that voltage. This doesn't make any sense whatsoever.

thats a good quote, getting 37v on the 12v rail would be hard and it would blow the caps because they're commonly rated to 16v

[прямо]

It did blew the caps, so I put in 50V caps for now until I can find out what is really wrong.

The only semiconductors that I had to change because they were toast when I got the PSU are:

Primary N channel mosfet FSW09N90AG to FNK09N90.

Primary PWM TL3845P to UC3845B.

I'm somewhat suspicious it maybe caused by UC3845B? maybe it has a different duty cycle?

If I removed the fan, I get 23V - 24V on the 12V rail, and 9V - 10V on the 5V rail.

DMM is ok and no I am not measuring at the wrong point.

Only 3.3V is using a TL431 reference. I did a quick check by removing it, and I got 4.5V on the output, putting it back I get 3.3V.

All measurements taken with the bulb current limiter on.

I'm using 40W bulb. Mains are 225V.

[прямо]

With the fan attached. Fan only runs for a minute. Probably because the sensor thermistor for the fan is missing.

[прямо]

Without the bulb current limiter, and without fan attached.
Only able to take photo of the 12V rail.
5V rail immediately died because a 27 ohm power resistor overheated and burnt.

[прямо]

Oh well, I guess this is a no fix. First the caps blew followed by the supervisor IC because VCC (aka 12V rail) was over the max rating, now the power resistor died, and God knows what else.

[прямо]

temyth ,

here are the board photos.
IC3, IC4, IC5 are AZ431A (TL431A)

IC3 = 3.3V rail FB reference.

IC4 = it made no difference to the 12V and 5V rails when I took it out even though cathode pin is connected directly to pin 2 of the main switching opto coupler. Cathode 1.8V, reference 2.6V

IC5 = 5VSB rail FB reference.

[прямо]

I have ordered TL3845P, a 10K sensor thermistor and the burnt out 27 ohm power resistor. There's another power resistor that makes up the -12V from 12V rail that got overheated as well and I lifted one of its leg for now. I will replace this too when all outputs are back to normal.

[temyth]

I have ordered TL3845P, a 10K sensor thermistor and the burnt out 27 ohm power resistor. There's another power resistor that makes up the -12V from 12V rail that got overheated as well and I lifted one of its leg for now. I will replace this too when all outputs are back to normal.

yeah, maybe the TL3845 is driving the mosfets crazy. there is always a fix, don't worry about it. are you sure you only have 1 comparator? i see a lot of TO-92 transistors in the board, they may not be exactly TL431 but they may be acting as voltage comparators. these half bridge PSU's often get a bad voltage comparator and the voltage goes just crazy.

[прямо]

There are two opto couplers going to the main PWM chip, one is connected to the 3.3V TL431 and one is connected to the other TL431 which has 1.8V on the cathode pin and 2.6V on the reference pin. I think this TL431 IS the feedback for 12V (maybe combined with 5V), except I don't know exactly how it works and what to change to make it compatible with UC3845B instead of the blew up TL3845P that the PSU came with.

Yes, there are a few transistors on board.

Near the main PWM chip is KSA928A (Q2)
Near the fan connector is 2SC1384 (Q6), part of the fan controller.
Near the supervisor is STC945 (Q7)
Near the 3.3V TL431 is STA733 (Q3)

[temyth]

try getting out the TL431 of 12v and 5v rails and test in a component tester. since all of this symptom is pretty strange i would also test the optocouplers with a multimetertry getting out the TL431 of 12v and 5v rails and test in a component tester. since all of this sympton is pretty strange i would also test the optocouplers with a multimeter

[прямо]

Actually, I have already replaced the 12V/5V TL431 and that doesn't change anything. In fact, even without it, I still get the same voltage outputs. Which is so strange and it's also why I suspect that even though UC3845B is pin to pin compatible, maybe it is not a direct replacement for TL3845P that was originally used in this PSU. Maybe it has a much higher duty cycle that can drive the MOSFET and the transformer to put out these higher voltages. Then again, it doesn't make any sense either because the feedback circuit is supposedly to keep them in check.

So I will do what you suggested.
I will replace the opto coupler.

[temyth]

uuuuuh now everything makes sense. you're right! they aren't the same thing, the feedback voltage of the TL3845P can go up to 6.3v, and the UC3844 is limitted to 2.58v, the PSU is probably overvolting the feedback pin and that's why you're getting so high voltages!!

[прямо]

I'm using UC3845B not UC3844.

I replaced the opto coupler, and..... et voila! 12V is now 12.5V and 5V is 5V!!! but now 3.3V has dropped to 1.68V