Swollen Rubycon on Dell Optiplex GX270

[Neo2_000]

I have been recapping a series of Dells with faulty Nichi's (HM's). Each one has about 6 swollen caps around the RAM slots - all 1500uF x 6.3V.

I have just ordered a bunch of replacement Ruby's of the same value when I opened a new PC today and surprise! Where there were all Nichi's, there are now Ruby's and one of them is swollen!

Will wait till I get the rest of the PC's (about 20 all up) and inspect them before making any rash strategic judgments but I am interested to see what your take on this is?

Should I up the ante and replace

1. the cap in this specific location with 1800uF x 10v on each board

2. or all the caps of this value whether they are already Ruby's or not?


Will be interesting to see what the other boards tell! Will keep you posted.

[pentium]

Re: Swollen Rubycon on Dell Optiplex GX270

Something is fishy.
Are rubycons usually blue or dark blue?
That looks black. The gold printing is also suspicious.

[Neo2_000]

Re: Swollen Rubycon on Dell Optiplex GX270

They can be any of several colours.

Not sure what determines colour selection at the factory but MBZs are Blue and MCZs are Chocolate - maybe a clue there

Got a blue ones from Computek before but the last lot I got were all as pictured.

(Blue ink on caps is mine. Pics taken at ISO 80 thanks to Willawake)

[scarfman]

Re: Swollen Rubycon on Dell Optiplex GX270

I've seen this frequently. I would always check the caps on GX270 boards for Rubycons & see if they were starting to look smoothed out, meaning they would probably start bulging or leaking within a few months. I always tried to get ones from Dell with Panasonic ones if I could.

[Neo2_000]

Re: Swollen Rubycon on Dell Optiplex GX270

Thanks for your reply scarfman.

I have now followed your previous posts and see that this issue has been looked at before.

On this specific motherboard, the VRM caps are all Ruby's including the 3 x 1800uF 16v VRM filter caps hiding under the heatsink.

There are however, 4 smaller 820uF 6.3V Nichis around the AGP slot that I have marked with blue ink on the pic above.

Whatever is causing the Ruby to fail, would I be correct in saying that the issue is confined to the solitary cap located between the AGP and RAM slots?

If so, I will just up the replacement specs on that cap only and forego replacing the other Ruby's on this mobo.

[scarfman]

Re: Swollen Rubycon on Dell Optiplex GX270

I'd say probably 90% of the Rubys I've seen fail are those ones between the AGP & RAM slots, yes. However I've seen some of the 1800uf 6.3 8mm ones fail on the rare occasion.

Has anyone had any reports of the Panasonic caps that Dell has lately been putting on their replacement boards failing?

[Topcat]

Re: Swollen Rubycon on Dell Optiplex GX270

I've seen Rubies fail in that position on the GX270 several times. It's caused by a failure further up the line, when the 3x 1800uF 16v Nichicon's under the heatsink fail, they overload this particular cap. The cap itself was not faulty.

[Spacedye69]

Re: Swollen Rubycon on Dell Optiplex GX270

Maybe, just a suggestion, but should we recomment replacing ALL caps 470uf and up? Oh, wait...

[scarfman]

Re: Swollen Rubycon on Dell Optiplex GX270

I've seen Rubies fail in that position on the GX270 several times. It's caused by a failure further up the line, when the 3x 1800uF 16v Nichicon's under the heatsink fail, they overload this particular cap. The cap itself was not faulty.

So from the time that the failed Nichicon 3x 1800uf 16v caps were replaced, it was just a matter of time before that particular Ruby would fail also?

What about boards that never had Nichicons on them at all? I've seen that Ruby next to the AGP fail recently on several boards that had all Ruby's on them when they were brand new from the factory back in 2003, all with 2003 data codes, long before the GX270 bad cap problems was even noticed.

[Neo2_000]

Re: Swollen Rubycon on Dell Optiplex GX270

I've seen Rubies fail in that position on the GX270 several times. It's caused by a failure further up the line, when the 3x 1800uF 16v Nichicon's under the heatsink fail, they overload this particular cap. The cap itself was not faulty.

Thanks for your reply TC.

On this specific motherboard, the VRM caps are all Ruby's including the 3 x 1800uF 16v VRM filter caps hiding under the heatsink.

There are however, 4 smaller 820uF 6.3V Nichis around the AGP slot that I have marked with blue ink on the pic above.

Perhaps these contributed to the Ruby's premature failure?

I am not suggesting that the cap itself is faulty but I am unsure as to the cause of it's failure and how to proceed.

Would you replace any of the other Rubycon caps around the RAM slots or no?

I'd say probably 90% of the Rubys I've seen fail are those ones between the AGP & RAM slots, yes.

Thanks for your reply scarfman. I will up the specs on this cap to be safe.

Have you ever seen any Rubycon caps located at the end of the RAM slots (see pic3) fail or are these always OK?

However I've seen some of the 1800uf 6.3 8mm ones fail on the rare occasion.

This board obviously holds some unique surprises. Were there any Nichicons hiding under the heatsik on these?

Given the remote odds, I would not recap a CPU row of Rubycons unless I found obvious signs of failure.

[bgavin]

Re: Swollen Rubycon on Dell Optiplex GX270

I have client's 270 machines that have died from PSU problems. Don't overlook the PSU as the top level source of the problem.

[Per Hansson]

Re: Swollen Rubycon on Dell Optiplex GX270

3 x 1800uF = 5400uF bulk capacitance

Most likley the bean counters removed one cap too many (or reduced their capacitance)

A P4 CPU is a real power hog, it needs some decent bulk capacitance

Reason for the failures to me sound like poor mobo design choices

[kc8adu]

Re: Swollen Rubycon on Dell Optiplex GX270

i have seen quite a few gx270 psu with bad caps.
these boxes run 24/7 in a plastic bottle preform plant.
dell swapped the mobo's last year.
i bet when you see rubycon or nichicon fail ouside the bad batches the chain reaction failure starts in the psu.
had one blow up and kill a custom made 32 channel logging card that monitors temps,weights,shot size,ect.
the card is out of production and one place that has one in stock wants 2300.00 for it.
they got real lucky i was able to save it.
needless to say i got to recap 85+ of these psu's.

[Neo2_000]

Re: Swollen Rubycon on Dell Optiplex GX270

20 boards finished OK except two.

On three boards, I replaced the 1500uF x 6.3V with 1800uF x 10v since I did not have matching values. Two of these expired after a short while because of VRM failure in the same location on both boards.

Judging by the scorch marks on the PCB, I question whether it worthwhile attempting to replace these as the board has probably sustained damage.

What do you think?

[Rainbow]

Re: Swollen Rubycon on Dell Optiplex GX270

I guess that when you replace the two MOSFETs (only one is smoked but both are probably bad), it should work.

[Neo2_000]

Re: Swollen Rubycon on Dell Optiplex GX270

Thanks, might give it a try just for the hell of it!

Where would be the best place to get these in Australia?

Thinking of placing a small order with Mouser - wonder if they would have them?

[starfury1]

Re: Swollen Rubycon on Dell Optiplex GX270

possibly RS components or farnell,

although after searching I couldn't find that part No.

You are probably going to have to source an equivalent part type

Here (hopefully) is the 13NO3LA pdf

Basic specs
VDs 25V RDs(on) 12.8 mR ID 30A,
N channel
(does show protection diode)

Unless someone here can give an substitute part your likely to source easily in OZ off the top of their head

HTH anyway

[MadCow]

Dell Optiplex GX270 has extended warrantee till end of 2007

Dell May Fix them For Free:

We had a problem with one of our dell GX270, it was actually reporting as a thermal event, while researching I stubbled across this in the dell forum

http://www.dellcommunity.com/support...l=en&cs=&s=gen

So I contacted them, they got me to visually inspect the caps which one was bulging. So they ordered a new MB and sent a tech to swap it out.
No Charge. They supposedly have extended the warrantees for the 270 model to 5 years or end of Jan 2008