After replacing bad capacitors, still nothing

[davmax]

Re: After replacing bad capacitors, still nothing

I realised after stating that the PS tester measured only no load voltages that it is possible to load test each supply voltage with low duty cycle switching. That is if each voltage is loaded for 100mS over a cycle period of five seconds and the total load 100mS load is 300 watts then the average wattage is only 6 watts. I wonder if a low cost device is that sophisticated to do the switching this way and apply different loadings eg a heavy load on 5 volts and a light load on 5 Vsb. The data available on the PS-224 is not revealing.

Has somebody got some experience with this tester?

[severach]

Re: After replacing bad capacitors, still nothing

The Coolmax LCD runs a 5W 33ohm resistor on the 5v lines. It's consuming just under a watt.

[gkirby]

Re: After replacing bad capacitors, still nothing

I just ordered a tester. Hopefully that will help when I get it in.
Thanks for all the help.
Gene

[gkirby]

Re: After replacing bad capacitors, still nothing

I received a Coolmax Deluxe ATX Power Supply tester today (LED model) and just finished testing the 21 Dell HP-P2507FWP power supplies which I have replaced swollen capacitors in. These ps's come with a 20 pin connector, and not a 24 pin. They ALL test exactly the same.
+3.3v
-12v
PG
+5VSB
+12v
-5v
+5v

All lights were green when testing all the power supplies except the -5v LED. The -5v led does not come on on any of them. What does the -5v control?? Do I have a problem with all of them??
Thanks,
Gkirby

[999999999]

Re: After replacing bad capacitors, still nothing

Dell like many OEMs now use PSU without -5V rail. That's not a problem as the boards don't need -5V.

These small plug in PSU testers can only find some kinds of faults, you need an appropriate load enough to put higher current demand on the PSU and then measure the rails, trying to power on the PSU by shorting PS-On to ground, or measure that voltage trying to turn the system on with the PSU connected to it. Since the PSU do turn on and run with the tester we can know some parts are working to some extent but this does not confirm proper function running higher load. Maybe it shifts the odds towards the boards being the more likely problem but it would be premature to assume it yet.

Try the questionable PSU with a known working board, and try a known working PSU with a questionable board. Right now, you don't need to be concerned about powering the Dell board with the Dell PSU (assuming these are standard not proprietary PSU as would also be indicated by the result with the PSU tester) as obviously one or the other is not working right, and it might not always be the same part (PSU vs mobo) that is to blame in all the systems.

[gkirby]

Re: After replacing bad capacitors, still nothing

The motherboards I had been having problems with were in GX270's which are knows to have issues. When I get time, I'll try these power supplies in a GX260 which I know is in good working order. Thanks for all the help.
Gkirby


I NEVER WANT TO HEAR THE NAME GX270 AGAIN!!!!!!!!

[severach]

Re: After replacing bad capacitors, still nothing

The -5v rail was dropped from the latest spec. Dell power supplies do not provide it. The LED tester isn't all that useful. The LCD tester doesn't cost much more and is way more accurate.

If it's any consolation I traded every one of my build-your-own Athlon 1700 to Pentium IV 2400 systems for a bunch of GX270s. Mine work great.

[999999999]

Re: After replacing bad capacitors, still nothing

I have a few PSU here that have failed and have no trouble outputting within tolerance on all rails when barely loaded as with these cheap plug in testers, they imply those PSU are ok. The problem is not one of how accurate the voltage reading, it is the scenario in which it has ran.

The LCD tester is of little use at this point, because it has already been established that the PSU will turn on with minimal load by the other tester, but the LCD tester also lacks the ability to fully load it.

Generally speaking, these tiny plug in testers are snake oil. They can't tell you a PSU is good, only that it's bad enough it won't power up at all. We don't care if it's exactly in spec with such a plug in tester either because it is no sign of a problem if it isn't fully regulating with such a tiny load, in fact most PSU manufacturers actually spec a higher minimum load for regulation than these testers can even produce.

[/rant] Sorry, it's just a pet peeve of mine they market junk wholely inadequate for the purpose. A PSU test that determines proper function has to at least test against the scenario in which the PSU will be used, if not the rated specs on the label.

[gkirby]

Re: After replacing bad capacitors, still nothing

After getting away from these power supplies for a few days, I'm ready again to try to figure them out.
After replacing the bad capacitors in 23 Dell power supplies, they all are acting the same.
1. They will not power up a good tower.
2. When checked with a LED tester, all of them tested good.
3. While plugged into the mother board, I can momentarily jump from the green wire to a black and the processor fan starts running. (I then removed the jumper). Then fan continues to run, but the computer still does not try to boot. The 12v-5v leads all show the correct voltage.
4. I can then replace this power supply with a "good" one, and the tower powers up.
What can be causing these not to power up?? Where do I start checking??
thanks,

[starfury1]

Re: After replacing bad capacitors, still nothing

no idea whats in the guts of those psu's or any likely causes of issues with them

Are you saying all 23 have been recapped
did you replace all DC output caps? or only those that were blown?

Iam assuming since the post is 2 pages long you used the right caps for the job and no odd ball pinouts on MB or psu...(probably result in smoke and flames anyway)

if you are saying they are all behaving exactly the same there my be a common failed component (you should be so lucky) although thats seems pretty much a long shot

I don't know what the effects are of loss of PG signal and I don't know how well that meter is going to show if its ok or not.
but it is required to boot the computer up after the rails have settled
it usually goes high I think for 350 mS (not sure on this point thought)

the SB +5 volt section would have been powered when plugged in 24/7 so you need to look carefully at that

Sometimes smaller value caps fail around the controller chip causing instability in regulation...its fine under no load but put a load on it and off it wonders
So there could be a few causes

this would be my first step, below

I would load the supply with some real resistive loads and see what the voltages are sitting at under more real load conditions load up the 3V3 5V and 12V

you will have to build your self a little test jig setup since you got so many
go get some Power resistors use "ohms law" to work out what you need.
(if you can get hold of an MB socket it could help too with building this)
(1 volt 1 amp = 1 ohm ....so 5 Volts 1 Ohm is 5 Amps.... 12V 12 ohms 1Amp get the picture)

you will have to put them in series and parallel to get the wattage up so the dont burn up...wire wound resistors come in 5 W 10 W values ...Volts X Amps = Watts

Series add like 1+2 =3 ohms
parallel "same value" half 4//4 = 2 ohms (its actually 1/X + 1/X then 1/x = ohms)

check that the voltage is in spec under load. load it up for an amp or 2 to start with

one forum member used head light bulbs as loads

you will need a decent Multimeter here
(doesn't have to be a fluke just something a bit more then the $5 DMM)

At this point you really need to do some conclusive testing I think

the following link will not be what you got but will give you an idea of the blocks used in the a switchmode power supply

http://www.pavouk.org/hw/en_atxps.html

http://www.smps.us/computer-power-supply.html

http://www.xbitlabs.com/articles/oth...thodology.html

although a bit on tech side may offer some insight for you on how they work.

Maybe those that fix on an every day basis can point you in the right direction with possible causes in this specific psu which is probably what you really want I know, but I hope this is of some help anyway

Cheers and yeah it sounds like a nightmare

[gkirby]

Re: After replacing bad capacitors, still nothing

I forgot to say that when testing the power supplies in a tower the only things plugged in were the 20 pin and the 4 pin plugs. Nothing else (harddrive, cd-rom, etc.) were attached. So, these won't power up even on a very light load (motherboard only).
thanks,

[starfury1]

Re: After replacing bad capacitors, still nothing

You still need to check what the actual voltages are under reasonable load.
if they are in spec or out of spec with a decent multimeter.

This would be my step 1

barring some weird Dell pinout which would have probably released the magic smoke
so it seems not your case...ATX standard is what you got by the sounds of it


Well not the psu guru here and wet finger in the air

I would see whats happening with the PG line, this signal tells the MB to boot up once the supply has settled down.

The little tester is in no way a conclusive test of the psu...it will give you an idea of it might be ok but does not stress test the psu
(basically go/no go .... or sorting the "Possible" wheat from the chaff, seems in your case its not even doing that)

As I said, I cant see all 23 having the same issue and would have expected after a recap most of them to fire up. (the law of averages says some should have)

You could hope they all do have a common fault, cause once you find the problem the rest will be a breeze.

Unfortunately I have no experience with this psu or even a clue as to the circuit construction. So I am flying blind.

MOTHERBOARD standard pinout ATX

pin 8 grey is pwr OK or PG

What does your little tester say about that line
(bearing in mind you must take it with a grain of salt)

I think and I am not 100% sure on this (Id have to check that) it goes high +5 DC for around 350 milliseconds.

anyway thats just one possibility...

I repaired a couple of PSU's awhile back there is a common problem with a 10uf capacitor used near the controller chip, under no load they would sit close to the specified voltage but under a real load would drift all over the place so this is why you must see what is happening under a real load.

BTW both did have exactly the same issue

Still this is strange 23 recapped and not one fires up...weird maybe you do have a common problem in all of them.

The other thing while you were recapping did you carefully inspected all the solder joints on power transistors diodes power resistors and transformers
expansion and contraction can fracture these and lead to intermittent or total failure. they could under no load be making contact enough that voltage wise things seem right but once current is drawn they act like a high value resistor.

some equipment has design issues that they do have some common types of failure and say 7 out of 10 fail in the same manor

there can be many possibilities

The other thing is start googling the model with different things like no power faulty not working and see what falls out.

This sounds like a terrible situation you have landed in and wish I could give you something more conclusive to try.

hopefully someone is familiar with these psu's and knows the most common likely causes and will post on it for you.

HTH anyway

Cheers