Help me repair a TT850W PSU

[KAMENi]

First of all, I want to say hi to everyone.

I'm from Croatia, Zagreb, and I've got a link to this forum from a regional
PC forum.
I hope You can help me out with a problem I have.


A Thermaltake 850W PSU, Toughpower, and it has a problem.
when on idle, it behaves good, gives good voltage values, but when plugged in a PC it loses it's grip and gives 3.7V on 5V line, and 4.1V when plugged into a PS tester.
everyting else is ok except 5V.

i opened it and replaced an A928A y-h31 transistor (it was gray-ish), only to discover it's used for overvoltage protection.

i've remembered i have a scheme of a 200W PSU and looked into it.
then i suspected Schottky diode malfunction (on this 200w scheme, after schottky is only couple of coils, 2 capacitors and a resistor)

i swapped the 2 schottky diodes (3.3V and 5V are identical) and it didn't solve the problem, so it leads me to the 2 caps on the 5v output

here is the scheme of 200w atx i'm viewing

[KAMENi]

Re: Help me repair a TT850W PSU

is there noone to comment on this problem? i don't expect a straight solution, just a good discussion


i know that i don't have the good scheme, so i'm using the one i have just for referal

please, anyone?

[Gianni]

Re: Help me repair a TT850W PSU

Hi KAMENi

was the transistor you have replaced damaged? If it was, have you checked all components around it?
If it has been damaged by an overvoltage probably other components are damaged too.

when plugged in a PC it loses it's grip and gives 3.7V on 5V line, and 4.1V when plugged into a PS tester.

Have you tested it only on one PC? Was the PC good?
What do you mean as "PS tester"? An active load or something else? Can you give more details on the way you are testing it.

IMHO it is strange that only +5V is wrong if you put under load the PSU.
As you can see from the schematic, usually the output voltages are coupled on an output inductance that help to keep within regulation all outputs. If one fails I expect also other fail, probably +3.3V.
Can you check with a scope the waveforms on the output diodes? Maybe something on the trafo is wrong.
Also failed caps can cause this problem, if they don't work properly (open) you obtain the rectifier voltage and with the DMM you measure the average value that can change at different load level.

Let see if some other member has more suggestions.

Gianni

[i4004]

Re: Help me repair a TT850W PSU

same as here
https://www.badcaps.net/forum/showpo...6&postcount=17

5Vstby is on separate circuit alltogether, though you're not thinking about that, it seems.

also it's not a chatroom so don't expect answers right away...zemljak..

[KAMENi]

Re: Help me repair a TT850W PSU

tranistor is blown. i found it to be Q1 on the scheme.
didn't check the components around it

this is what i refer to as PS tester:
http://www.vemil.hr/opis.aspx?prikaz...&Sifra=7710025

i replaced the damaged transistor and tried the PS in my cpu. it first managed to enter BIOS, and after restart didn't do nothing.
then i plugged my digital multimeter in the +5v and gnd wires of a molex and it said 3.7V
unfortunately i dont have a scope to see the waveform, so i'm pretty limited to DMM and average values

5vsb is good, PG is good, +/- 12v is good, -5v is good and +3.3v is good

i can try to replace the 2 caps on the output line of the +5v, but i don't know how to test them

[Gianni]

Re: Help me repair a TT850W PSU

this is what i refer to as PS tester:
http://www.vemil.hr/opis.aspx?prika...0&Sifra=7710025

Ok Now I understand.
This tester is the same as DMM, the only difference is that it is a "dedicated tester".
I think the better way to test the PSU is to attach some load (power resistor) and check the voltages but to check ripple you need a scope.
You can try to understand if there's high ripple measuring voltages with DMM in AC mode, it is not accurate but if you measured something greater than 0.1~0.2 Vac for sure something is not working properly.
I have done a test on a small DC/DC converter: I have loaded it at 1.5A, with the original good 100uF cap and then I replaced it with a general purpose 2.2uF, just to simulate a failing cap.
With the DMM the measures were:

with 100µ Vout= 4.93V Vout AC=11mV
with 2.2uF Vout= 4.82V Vout AC=470mV

You can see also the attached picture taken with the scope; if you look at the "mean" values their are close to the DMM measured values. It must be underline that in both cases the output measured with DMM was within spec, 5V ±5%, but with the 2.2uF cap the max/min voltage were completely out of spec. This behavior is due to the regulator that try to keep the voltage at its nominal value but since the output cap is too low, the output oscillates and the average voltage is around the set value of 5V.

The "PS tester" doesn't load the PSU and you can't understand if something is wrong, you can only see if the voltages are present and within tolerance.

When you connect the PSU to the MOBO and you try to boot it, can you "feel" if something is heating up?
If only +5V goes down maybe there's something shorted and the PSU are limiting the power/voltage; if it occurs leaving the PSU powered you should identify something heating on the MOBO.

Anyway Q1 is not for overvoltage protection, as you wrote in the first post, it is a power transistor to switch the high voltage on the transformer to transfer energy on the secondary side. You should check all components around it and also the IC/circuit/components that drive it. You must pay attention because this is the primary side with high voltage and you may get hurt; to check if everything is good on the primary you need a scope and knowledge so be careful. You need also a 50Hz power transformer to isolate the PSU otherwise you can have serious problem with the probes on the primary side: the ground is always connected to earth so if you don't isolate the PSU you can short the input mains.
If you have no experience with high voltage it is better for you no to play with it.

I see your PSU has the active PFC, so the schematic is very different from the one you attached.

By the way, if you think the PSU is failing is not a good idea to connect it to good MOBO/HW, you can damage them.

Gianni

[i4004]

Re: Help me repair a TT850W PSU

as gianni said, q1 is main switching transistor, and i don't think psu can work without it at all....

you changing it and fixing nothing kinda proves that point.
as i said, i don't see how can this be a problem of primary side at all, because if that was the case all voltages would be off or no voltages at all on output(more likely)...

remove the output caps on 5v rail and see what happens...if nothing changes, test the resistor.

if still nothing, post some pix of it; like gianni said active-pfc changes a lot.
you could at least shows us what transistor you changed...
and we could see what caps are inside...etc.

[KAMENi]

Re: Help me repair a TT850W PSU

sorry for not responding. i've been busy preparing few exams.
i hope i'll have time this weekend to replace the 2 caps and take some pics for upload

i must admit, i didn't think of measuring AC voltage when it was connected

regarding the replaced Q1. i will leave it the way it is because i don't have equipment to measure what's going on around it, so i'll say it's working fine

i'll focus on repairing the secondary side (with your help, of course)

[KAMENi]

Re: Help me repair a TT850W PSU

i replaced the capacitors. i'm putting the psu back together, and will post results when i try it

here are some pics of the unit

[KAMENi]

Re: Help me repair a TT850W PSU

on AC it measured 0.009v for both 12v and 5v (on 2v AC on DMM)

[KAMENi]

Re: Help me repair a TT850W PSU

again it failed.
it measured 0.07v ac (on 2v DMM)

i tried it in a old socket A mbo. it was running for few seconds than decided to shut down
voltage was normal till that point. i can't force it to work longer for me to make thorough measuring

any more ideas besides the primary side?

[PCBONEZ]

Re: Help me repair a TT850W PSU

Looks bloated.

.

[kc8adu]

Re: Help me repair a TT850W PSU

Looks bloated.

.

looks like it hatched.

[KAMENi]

Re: Help me repair a TT850W PSU

all 3 of them look normal. i can post some more pics...

[PCBONEZ]

Re: Help me repair a TT850W PSU

If you say so.
It looks bloated in pic.
Maybe it's a camera thing.

[KAMENi]

Re: Help me repair a TT850W PSU

there are 3 of them under that bunch of wires, all 3 are 2200uF and 16V and all look the same. none bloated or anything

only, i can't read what manufacturer is it. they are the same as the 2 on the pics above, only those are 6.3V

it appears they are chemi-con

[PCBONEZ]

Re: Help me repair a TT850W PSU

Chemi-con KY series if the same series as caps in photo.

[KAMENi]

Re: Help me repair a TT850W PSU

is there a way i can make the psu stabile so i can make some more measuring?

[KAMENi]

Re: Help me repair a TT850W PSU

could it be that powergood is blown and "thinks" that it's not good?

[Gianni]

Re: Help me repair a TT850W PSU

You should be able to keep the PSU on connecting the green wire to a black wire on the 20 poles connector.
I usually do it in that way just to see if the PSU start.
Most PSU stay on even without load but it happened to me that few unit didn't start at all.
Yesterday I tried 2 Newton Power I received by Ebay: they start for few second and then they shut down; I had the time to measure the voltages only restarting the units 4 or 5 time.

Gianni