Dell proprietary PSU repair [SOLVED]

[spilihps]

Hello!

Well, i decided to try to better understand how a PC psu is built, and since i had an dell SFF with a bad psu sitting, i thought that would be a good project. Turns out it may not be.

I opened it up, and found that there have been black dust expelled from the W7NK90Z MOSFET (its completely shorted when measured out of circuit), and also that the top of a mysterious IC on a "daughterboard" had blown of. (see picture)
For starters, i cannot find any info about the CL2181 IC, what it does or how it works, neither any replacement, well i found replacements in china though..

I did some poking on the little daugtherboard and found that the large mosfet is connected to the IC:
ICpin1 = source
ICpin3 = gate

Also found that a resistor called R301 (cant see the value, either there were no printed from the start, or it have gone missing in this) is open circuit. The rest seem OK though.

The question is. Can this be repaired? What else to look for? What value of resistance to replace the R301 with? Can the CL2181 be replaced with something else, more widely available?

It would be a fun and a good learning exercise to do this repair, but maybe im in over my head. Also it would save a nice computer, that has less value than a replacement PSU on ebay 

Any help and input in this would be greatly appreciated.

Thank you

[wraper]

Just get one on ebay or Aliexpress.

[spilihps]

Yes i think i will do that then.


But i really need some way of figuring out the value on the resistor though..

[wraper]

You haven't said to what that resistor is connected. How someone can tell the resistance if there is no schematic and unknown where it is used.

[wraper]

All I can find. But in this schematic it is connected to the mosfet through the transformer.

[spilihps]

Hello.

Thanks for your help.

The R301 is connected this way:

CL2181 pin1 -> R301 -> pin3(Source)MOSFET

[raz221]

Maybe I'm missing something, but where exactly is this resistor you're talking about (in the picture you've posted)?
All I see is a transistor, a chip and 3 electrolyte capacitors. Also, it might help if you'd provide the full model number of this PSU.

How someone can tell the resistance if there is no schematic and unknown where it is used.

By the color coding and a simple online resistor calc:

Code: [Select]

http://www.digikey.com/en/resources/conversion-calculators/conversion-calculator-resistor-color-code-4-bandEach resistor has color coding (4, 5 or 6 bands) which specifies it's unique resistance value.
I always use this calculator before measuring resistors (unless I have the schematic), and sometimes get em out of the circuit for verification if I get the wrong value or no reading (if they're connected in parallel with other ones, the reading can be inaccurate). But the above only applies to regular (metal or carbon film) resistors of course. For SMD resistors, some of em have the resistance value written on em (which often requires a magnifier to see), and for others, you really can't tell without having the schematic. But even if you got the correct value (and the resistor turns out to be faulty) sometimes manufacturers use custom-made resistors (with non-standard values) which can make repairing difficult.

[wraper]

I guess that likely is SMD resistor. You can check if the circuity around pin1 and pin 15 is similar and check the values. If R221 on you picture is close in value with your circuit, then there is a good bet you can use 2.7k instead. Also check if the resistors connected to the source are fine, they tend to blow up together with mosfet.

[spilihps]

wraper, thank you again very much for your help.
Yes its a SMD resistor.

Ive done some quick checking on the circuit diagram that you provided, the R221 is the same exact value as on my board. Also, the circuit around those pins seem to be the same. I also checked some of the others, and they seem to be the same also..

Thanks for the tip about the resistor, and this is the resistor for the source, so you are right there also i guess?

[wraper]

By resistor to the source I mean some higher power resistor(s) which passes all of the current from the source to GND (most likely). Not this resistor which is solely for feedback.

[spilihps]

Hello yes, sorry i misunderstood you the first time.
And yes, i have a resistor that is connected to the source that is also OC
it is red, violet,silver gold, so i guess its 0,27ohm?

Will be very interesting to see if this is infact repairable, seem to me to be a rather large "crash?"

The components found bad this far:

IC301 CL2181
R301 (2,7kOhm according to the above circuit)
Q102 MOSFET W7NK90Z
C125 102K cap
R105 Red, violet,silver,gold resistor.

Again, thanks for your help and input.

[wraper]

Yes, it's 0.27R. Be sure to use replacement of not lower power than original is, can be estimated by it's size. Fault happened in the way it usually does. BTW MOSFET model is STW7NK90Z. 102K is 1 nF 10%

[spilihps]

Now i made some research about what resistors to choose. ( I have sourced the CL chip from china, so it will be a loooong wait)

The larger 0.27R have a diameter (on the thickness part) of 3.45, and length of 9.46 and the one that comes closest is a 1/2 watt. Do you think this would be safe?

The SMD resistor, i looked just on the gap between the pads that seem to be correct with 0603, that says 0.9 and i got 1.0 when measuring the board..

I used the charts at resistorguide.com

[wraper]

The larger 0.27R have a diameter (on the thickness part) of 3.45, and length of 9.46 and the one that comes closest is a 1/2 watt. Do you think this would be safe?

1/2 W or 0.6W resistors usually are much smaller. Use 1W or small 2W metal film resistor. Likely you looked on dimensions of some crappy carbon film resistors.

[wraper]

As you can see from the following table, in such size it can be even 2W resistor.

[spilihps]

Oh i see. And yes the chart said "carbon and metal film". So i guess that chart is nothing to trust.
Good catch..

I use a 1W then. Better look for another SMD-chart also to be sure

 

[damn_dirty_ape]

This is probably repairable with the right equipment, which might be worthwhile if you're just trying to learn, but from a practical standpoint you should probably just get a new supply.  Repairing surface mount stuff takes a lot of practice, and I've totally destroyed a couple boards by not using the correct equipment

[spilihps]

In this case, the power supply is not worth getting for the money they are asking. Seems like these supplies fail often. And the time invested is the price for learning
Yes SMD parts can be a hassle, but i have experience with it and the only SMD part in this case is a resistor, and its already out so no drama there.



Thanks for your input also.

[spilihps]

Hello again.
Tried to send you a PM (wraper) But i dont think it worked.

However, tonight im trying to source the parts, and i was just wondering if you have any idea of what wattage and voltage the 0603 SMD resistor should be in this case?

[spilihps]

Late update on this.

So ive manage to source the parts, and meanwhile i did find several more that needed to be replaced, so ive done that now.
Still no go though.

The PSU does now source +5v standby, but nothing happens when grounding the green wire.. Also It seems though that the +5v standby is completely separated from the rest of the psu.
When grounding the green the only thing that happens is that i can hear a very very very faint high pitch sound and that output goes from 0,000v to 0,001v.
There seem to be no voltage over the optos LED pins during this time either..

Any input on what to test and/or measure from now is greatly appreciated.

Parts found bad so far:

IC301 (CL2181)
R301 2,7kOhm smd resistor
R104 0,27R
C125 102K
Q102 W7NK90Z
R51 1R (To the 5v supply)
Fuse
Q104 KSP2222A (transistor for one of the optos) it had E-B fail

[stj]

post a foto of the main area where the fet is.
i bet it has a snubber ceramic and a fuseable resistor to one side.

this issue sounds like a similar problem on a newer psu.

[spilihps]

Hello!
Thanks for your help
Ive attached a picture of it here.

Ive got a tip from another user to measure the voltage over the optocoupler during the moment it tries to turn on, and for a brief second it actually rises.
Dont know if this rules anything out completely?

[wblock]

A failed Dell power supply, with L-tec capacitors (not the worst of the worst, but not the best of the worst, either)...

Even if those L-tecs checked out okay on an ESR meter (they might; it's happened), I'd replace them all.  That involves measuring all their dimensions (height, diameter, lead spacing) and trying to find the spec sheet to get something close to their claimed ESR.  If it was worth it, which it probably isn't, but it sounds like this is a learning experience or quest.

[stj]

if that snubber cap is 1Kv rated, and discoloured then uprate it to 2Kv

there is a newer supply that fails a lot (as in - all of them!!!) because the snubber cap is under-rated and cooks until it fails killing the fet and blowing the control circuits.

[spilihps]

A failed Dell power supply, with L-tec capacitors (not the worst of the worst, but not the best of the worst, either)...

Even if those L-tecs checked out okay on an ESR meter (they might; it's happened), I'd replace them all.  That involves measuring all their dimensions (height, diameter, lead spacing) and trying to find the spec sheet to get something close to their claimed ESR.  If it was worth it, which it probably isn't, but it sounds like this is a learning experience or quest.

Hello and thank you for your reply. The small blue one is to the left of the FET is one that i replaced because the other one was blown, it was a brown one, marked with 102k, and replaced by this that is also a 102k. the exact type i can look up on my order sheet if necesarry. And you are completely right, this is a learning experience

if that snubber cap is 1Kv rated, and discoloured then uprate it to 2Kv

there is a newer supply that fails a lot (as in - all of them!!!) because the snubber cap is under-rated and cooks until it fails killing the fet and blowing the control circuits.

Hello and thanks for your help

The small blue one is a new one that ive replaced, because the other one was blown, it was a 102k.
However the large blue one, checks OK with capacitance measurments, so i left it.

Could that one be causing it not to start up?