This post is more than 5 years old
17 Posts
0
1603
R910 Leaked/Burned component
Good morning,
I have just opened up our R910 for maintenance and have noticed that one of the components beneath the heatsink for CPU 3 has evidence of heat damage as well as a liquid that seems to have leaked onto the motherboard.
The server passes all diagnostic tests and seems to be performing OK.
Could anyone please help identify what this component is (we believe a diode) and what could have caused this damage please? Image attached:
Mtexter
1 Rookie
1 Rookie
•
205 Posts
0
October 6th, 2016 09:00
If it's a capacitor, then that goop that's all over your system board is an acid of some type and you should be VERY concerned. I'm surprised it's still working, and would consider that server to be a ticking time bomb. Replace the system board ASAP. You'll be fine to migrate your CPUs, memory, etc. to the new board. To reiterate, DON'T TRUST THAT BOARD!! That acid is all over a printed wire and it should only be a matter of time before it eats through it and destroys that circuit.
There's a chance it will neutralize and stop eating through before then, but I wouldn't count on it. To mitigate, if you don't have the option to replace, spray some non-conductive contact cleaner on it and try to wipe it up with a lint-free, static-free cloth. Q-tips/cotton swabs are usually fine for this.
skylarking
548 Posts
1
October 7th, 2016 18:00
Elecrolytic surface mount capacitors tend to look like small round tins sitting on top of a small plastic base. Such surface mount electolytic capacitors can indeed leak but there would be visable physical damage at the plastic to tin can junction (where stuff would ooze out). Such ooze can get under the protective mash and down the vias causing complex faults that can be dificult to isolate, or complete failure.
Other types of surface mount capacitors are ceramic and tantulum which as far ar i know can't leak fluid since there is none inside them :emotion-3:
The device in question, C563, does not look like an electolytic capacitor and it it is not electrolytic it can't leak. Only Dell knows exactly what the component is but i'm sure Dell would not want to devulge such 'proprietary information'. Dell would prefer to sell you a new board with warranty which may be a more sensible option for most.
An in house repair may be cheaper than a new board but the repair quality would depend on the repairers knowledge and skill. Correct and thourough cleaning must be undertaken if electolyte has indeed leaked, especially down vias... Outsourcing a repair would likely be costly and not worth it these days :emotion-43:
In any case, most companies don't provide circuit diagrams, component lists or other repair documentation these days so that makes it all that much more difficult for a repair industry to exist :emotion-6:
But, if C563 is indeed a capacitor, then it is most probably not an electrolytic capacitor. However, i'm no expert on components or hadrware repair :emotion-5:
As such, being VERY concerned may be premature and more information from a forum memebr that knows more would be helpful.
But if OP is concerned at suffering an outage and such an event would be costly, the best insurance may be to get a replacement mobo and swap out the gooy mobo. This would provide peace of mind but maybe not the satisfaction in knowing what the component and fluid actually was :emotion-5:
Heck it could still be a little flux from a factory repair of C563...
DDH88
17 Posts
0
October 4th, 2016 02:00
Hi All,
Just following up on this from yesterday. Does anyone have any clues? I'm open to speculation,
Thanks for your time.
skylarking
548 Posts
0
October 4th, 2016 21:00
OK speculation it is :emotion-1:
SMD components are always tricky to identify, especially for non hardware designers like myself. It's compounded as manufacturer can't settle on standard within their own organisation let alone on an industry level (that i am aware of). So it's a hotch potch method of trying to identify a manufacturers mark, which may be very suttle, a package type and smd code markings just to identify the component.
In your case the picture is a fuzzy and not zoomed in closely enough to identify any manufacturers mark while one can't clearly see the sides of the component to determine pin count. But it seems to be a 2 pin package with 330 and d29 markings, which could indicate a zener diode but it could be something else altogether :emotion-42:
However, Dell has also stensiled C563 on the motherboard and as manufacturers generally use C of capacitor, R for resistor, D for diode, etc, i'd speculate the component could be a capacitor.... but only Dell knows what component is really associated with C563 :emotion-1:
As is, it's all working and should not cause you concern. The more someone handles the hardware, the more the risk that it will fail increases. There is an old saying that to keep things working you need a man and a dog; the man to feed the dog and dog to keep the man away from the hardware :emotion-5:
So maybe leaving it alone is a better strategy.
As to what the liquid is, maybe it's the left over ooze from some long dead bug, maybe it's a little leftover flux that didn't get fully cleaned during manufacture, maybe its a touchup repair of c653 where the flux wasn't cleaned. Who knows? Is it so relevant? Your system is working, be happy :emotion-2:
DDH88
17 Posts
0
October 10th, 2016 02:00
Thanks to you both for your detailed input it's much appreciated. I think both of your opinions very much reflect our feelings on the server.
I think we'll conduct some repairs just to prevent future failures.
Thanks very much for your time and detailed feedback!