Antec "Stupid" Power rebuild

[BigTroll]

Re: Antec "Stupid" Power rebuild

Throughout the years of tearing stuff down i rarely if only a couple of times found a bad teapo.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

I could easily be mistaken but it looks like the according to the datasheet that pin 7 TPS3510 should be connected to 5VSB and the 12V that is generated off the supply for OV protection on 12V. They both are supposed to feed pin 7 across separate diodes. (5.03VSB - .5V diode drop = ~4.5V until the supply starts) I agree that pin 4 PSON seems low voltage wise, but best as I can tell it is only connected to pin 4 through a resistor so the low voltage is caused by the TPS3510. I am sure this supply had a 5VSB over-voltage situation as the capacitor that was supposed to control it was hammer dead. I suppose it could have easily have gone high enough to damage this chip. It gives no response to its' weak PSON signal. I think this chip may be my problem. Based on this information what do think? I already ordered one of each chip from Digikey last week just in case there was a problem with one of them. I had to place an order and they were cheap compared to time and parts already spent on this project. Should I just wait until they arrive then replace the TPS3510, then the KA7500 if the TPS3510 does not solve the problem? Or do I need to investigate more before replacing parts?


I should be able to post pictures, I may need to wait until tomorrow morning so I will have good light. Bad light will equal bad images. Thanks for taking the time out to respond.

Lloyd

[momaka]

Re: Antec "Stupid" Power rebuild

Although the supply voltage on pin 7 for the TPS3510 is still in spec (more than 4V), I find it a bit low. It may be okay after all, but considering that the supply voltage for the KA7500 is 12V, I think you should look into it. Do you think you can trace the circuit that provides power to the TPS3510 on pin 7?

Also, if I'm not mistaken, pin 4 (PSON) should be close to 5V. It's still in spec as well, but not by much (just 0.2V).

Any chance you can post a few pictures (top and bottom side of PCB) of the controller area?
These ICs usually don't go bad, but it's possible that the out-of-control 5VSB pushed the auxiliary supply too high and killed one of them.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

Well I still am not sure as to why this supply is not working. I have done a lot of searching and reading but I have not found very much that is helpful to me. I am suspecting the problem may be the TPS3510 supervisor chip. I am not sure as I am not getting much out of its' datasheet or the one for the KA7500, especially how they interact. I took some voltage readings in standby for both chips then I did the TPS3510 while trying to power on the supply with a load. The only reading that changed on the TPS3510 was the PSON going from 2.6 to ~0V

Here are the readings by pin number:

TPS3510

1 0V
2 0V
3 0V
4 2.6V
5 0V
6 0V
7 4.5V
8 1.5


KA7500

1 0V
2 0V
3 0V
4 0V
5 0V
6 0V
7 0V
8 2.1V
9 0V
10 0V
11 1.6V
12 12.1V
13 0V
14 0V
15 0V
16 0.1V

Other than the Vdd for both chips I do not know what these should be in standby mode. Any input would be much appreciated as I have no idea what to do next besides replacing these two chips and see if that helps.

Thanks,
Lloyd

[LDSisHere]

Re: Antec "Stupid" Power rebuild

Thanks for helping. I have 11.5V on my PWM chip pin 12. I am attaching the pwm datasheet and the supervisor chip datasheet. I figure the problem may be in that area. I have no idea how they work together. I get no fan twitch. Any ideas on what to check next?

Thanks,
Lloyd

[momaka]

Re: Antec "Stupid" Power rebuild

Does the PSU fan twitch when you try to turn the PSU ON? If not, try attaching a bigger load to it - fans and old dead hard drives are really good for this. Or even better, an old spare motherboard. If the power supply still doesn't do anything, start checking the circuits attached the to secondary side of the 5VSB transformer. There's usually 2 circuits - one for providing power to the PWM/supervisory IC on the secondary and one for the 5VSB output. The circuit for providing power to the PWM/supervisory IC might have a fault in it. It usually outputs 12-25V.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

Well I got my caps in, and I have now replaced almost all the old caps with new ones. There were some 16V ones that I missed on my order but I replaced them with scavenged ones that tested good. (Good brand also). I also have not replaced the main caps yet until I am sure I can get it to work.

I have a problem and I do not know where I should start troubleshooting it, I am only used to replacing power supplies. When I apply power I get 5.03VSB (multimeter reading) which is great if my understanding is correct. When I try my ATX tester I get nothing other than a green LED for 5VSB . Pressing the button to turn on the supply does nothing that I can tell. The only other voltage I get is 2.46V on the PS_ON# wire. I do not know what this is even supposed to be so I am not sure if it is correct. I have not tried to force it on using a jumper wire as of yet, I thought it best to get some input before proceeding. How do I need to proceed from this point?

Thanks,
Lloyd

[momaka]

Re: Antec "Stupid" Power rebuild

I fixed his Antec power supply and it is still going good, but the Intel motherboard has been sitting in my basement for about 18 months now.

Probably needs a new Southbridge... unless you got lucky and only had a linear regulator or standby-power MOSFET die somewhere.

I have a few old dead boards as well. I'm saving them to experiment with when I build myself a reflow/rework machine sometime in the future. Would be fun trying to change the chipsets and other BGA chips on those.

The stripe (Cathode) on the diode should go to the +5sb output, and the other end (anode) should go to ground IIRC

Correct.
I've seen this used on a few older power supplies so it's not a new concept.

[c_hegge]

Re: Antec "Stupid" Power rebuild

Cool. Just remember that the zener should be rated for 1W or more, otherwise, it will just melt if a 5vsb overshoot occurs. The stripe (Cathode) on the diode should go to the +5sb output, and the other end (anode) should go to ground IIRC

EDIT: I should probably add that the idea was originally Th3_Un1qu3's

[LDSisHere]

Re: Antec "Stupid" Power rebuild

^
Retiredcaps, you are correct that guy sums it up very well. I was just excited about that capacitor, since I have recently built my ESR tester it is the worst testing / best looking cap I have yet come across. The only others to show open had their tops ruptured

You can mod it if you don't trust the 5vsb circuit. Either you can replace the circuit with one driven by a PWM chip, which wont overvolt, or you can put a 5.5V 1W zenner diode across the output. Then, if the 5vsb goes slightly over 5.5v, it will conduct the excess over to ground. If it goes way over, it will short internally and blow up the switching transistor but the motherboard will survive.

I think I will go for your zener idea, keep it simple for this situation. It is quick, easy and hopefully never tested. I have been wanting some 5.5V zeners on hand for a while so now is my excuse, I will add them to my order. Thank you very much for the idea.

[retiredcaps]

Re: Antec "Stupid" Power rebuild

The little cap that went bad did not show any outward signs of failure.

Jetadm123, a regular in the computer display subforum here, puts it eloquently

https://www.badcaps.net/forum/showpo...96&postcount=8

[c_hegge]

Re: Antec "Stupid" Power rebuild

You can mod it if you don't trust the 5vsb circuit. Either you can replace the circuit with one driven by a PWM chip, which wont overvolt, or you can put a 5.5V 1W zenner diode across the output. Then, if the 5vsb goes slightly over 5.5v, it will conduct the excess over to ground. If it goes way over, it will short internally and blow up the switching transistor but the motherboard will survive.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

The little cap that went bad did not show any outward signs of failure. Once I saw its' readings I figured bad things happened down stream. I had read so posts that discussed this flaw. The board that was attached to this supply when I got it was an ancient (266 DDR max) socket A board that I was not planning to test as I have better ones boxed up with no purpose already. I am now curious if it is fried, so I may have play with it if I ever get at chance. The board I plan on using is an Asus P4 that I recapped a while ago and have been wanting to put into service.

It is a pretty crappy design IMHO that when the cap goes bad, the voltage fails high vs. off. If I ever intend to hook it to an expensive setup I will probably do the mod that was described above, but a P4 board I am not going to loose any sleep over. I am going to do a little more testing of diodes and such before I order my parts but hopefully I have it in before Friday.

[retiredcaps]

Re: Antec "Stupid" Power rebuild

because the reliability of this circuit depends on a small electrolytic cap - usually 10 uF, 22 uF, or 47 uF rated at 25 to 50V. When this cap fails, the 5VSB overshoots.

My friend's Intel socket 775 motherboard paid the price due to an Antec power supply and the above cap reaching 13V DC by the time I got it for testing.

He knew something was wrong for months as the power supply squealed, but kept using it until the entire computer died.

I fixed his Antec power supply and it is still going good, but the Intel motherboard has been sitting in my basement for about 18 months now.

[momaka]

Re: Antec "Stupid" Power rebuild

After work I was able to pull this cap to test it. It was bad in both ESR and capacitance. I am assuming that I should use the same value for a replacement.

Yes keep the capacitance the same and the voltage same or higher.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

On the SP-xxx, it's the circled one (a 22uF 50V), but still replace all of the small caps.

After work I was able to pull this cap to test it. It was bad in both ESR and capacitance. I am assuming that I should use the same value for a replacement. Thanks everyone, for taking the time to help me out with this rebuild.

[LDSisHere]

Re: Antec "Stupid" Power rebuild

Indeed.
In fact, given the PCB discoloration on that power supply, I am betting the 5VSB has already started to go out of control. These power supplies use an older 2-transistor design for the 5VSB, which isn't all that great because the reliability of this circuit depends on a small electrolytic cap - usually 10 uF, 22 uF, or 47 uF rated at 25 to 50V. When this cap fails, the 5VSB overshoots. So it's imperative to replace that cap. Or better yet, do an "everell DM311 mod" if you have experience with building some circuits.
(more can be found here:
https://www.badcaps.net/forum/showthread.php?t=8403)

That looks like something I may want to do in the future since I can etch PCB. I am only hooking it to an old P4 AGP board for now and I have way too many other projects going. The cap C_Hegge circled is a 22uF 50V. I have not had time to remove it but I did clamp my Blue ESR tester to it an I got an open reading. I will pull it as soon as I have time and see what it really reads. The R15 resistor seems to be fine. I also tested the transistor on the other side of the heatsink from the circled cap for shorts but it was not shorted. I am still going to poke around and check a few more things when I get time.

I was able to pull apart my son's supply, but it was already dark so please excuse the poor quality again. About all I can say is that there in nothing popped or burnt looking. I do not recognize the cap makers so I figure that is a bad sign, but I have not known that it mattered for very long. I am not posting the brand, as I am curious if anyone can recognize it by the guts.

Thanks,
Lloyd

[goodpsusearch]

Re: Antec "Stupid" Power rebuild

@c_hegge, http://9gag.com/gag/5391740

[c_hegge]

Re: Antec "Stupid" Power rebuild

On the SP-xxx, it's the circled one (a 22uF 50V), but still replace all of the small caps.

[momaka]

Re: Antec "Stupid" Power rebuild

The small caps should also be replaced. They can cause either annoying high pitched squealing, or worse - a 5vsb overshoot

Indeed.
In fact, given the PCB discoloration on that power supply, I am betting the 5VSB has already started to go out of control. These power supplies use an older 2-transistor design for the 5VSB, which isn't all that great because the reliability of this circuit depends on a small electrolytic cap - usually 10 uF, 22 uF, or 47 uF rated at 25 to 50V. When this cap fails, the 5VSB overshoots. So it's imperative to replace that cap. Or better yet, do an "everell DM311 mod" if you have experience with building some circuits.
(more can be found here:
https://www.badcaps.net/forum/showthread.php?t=8403)

I think the critical cap in your power supply is either one of the caps in this picture:
https://www.badcaps.net/forum/attach...2&d=1347829503
or this picture:
https://www.badcaps.net/forum/attach...9&d=1347829503
Not 100% sure which one, so may as well replace all 3. Other than that, the power supply looks good.