Advice needed: Dell 5150 blown ceramic SMD decoupling cap

[chaphal3]

Hi, I blew this cap by touching the heatsink while I was running the machine with no case. Apparently, suddenly changing from ungrounded to grounded while running a cpu can cause ESD. Who knew? The funny thing is I was had it open to investigate why it wouldn't boot, but all it needed was a CMOS battery, so when it started booting I got worried and held my hand "near" it to make sure it wasn't hot enough to melt the table it was on. I bumped it and: "pop", ozone smell, no more booting, lesson learned, I hope. It would be nice to get this running, as with some very cheap upgrades (pentium D to C2D possibly) it could be my best computer.

The SMD that I could see visibly crumbling seems to be a decoupling cap on the 12v power to the CPU. It was cracking, and had already flaked off a bit, just poking at it a little with a wire scratched it to the state in the picture.

I have had trouble finding a schematic for this board (cap is PC82 on board model PWA J8885 rev A-01, also marked as E187242 F2, or model Connolly from a dimension 5150, also possibly 0HJ054 but that's not printed on the board). A few months ago I though I found a couple different datasheets that might have been applicable but they didn't list the same capacitance; I think they were something like 1 and 4.7 microF (maybe I'll find and link them by the time I finish this post).

Does the capacitance need to be an exact replacement if it's just for power supply decoupling? I was thinking, assuming the decoupler is like a high-pass filter, as long as it's lower in uF than what it's replacing and still fairly close it would be OK. Is that wrong? Can you help me determine the exact value it's supposed to be?

Thanks a ton for any help you can give me.

[stj]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

it's probably 0.1uf

now go clean all the dust out with a paint brush.

[chaphal3]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

Thanks, are those pretty much always .1? I just read a paper saying the "bypass" is usually .01-.1 uF. I guess I misremembered the datasheets I somehow found a few months ago and somehow cannot find now.

I always thought a nice patina of dust gives electronics a nice Classic Look and Protection Against Longevity.

[ChaosLegionnaire]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

its just an emi-rfi filtering cap. its function is non-critical and u can just run the system without it. nothing bad will happen. i've seen mobos and video cards run without it. dont fret too much about replacing it or what value to use.

[chaphal3]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

The system doesn't run now, so if it's non-essential would that mean either there's some detection that prevents booting (I think there was a beep code but I'm hesitant to check because of the ozone smell), or another component is bad? Or maybe do I need to remove it so it's not shorting to ground (I thought it had failed open and not closed but I'll check again)?

I also read these need to be SMDs to reduce inductance. Is that just theory and in practice can I just use short leads? I know this question is basically counter to the advice not to fret about it but in case it turns out I need the part.

Thanks

[ChaosLegionnaire]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

check your power supply. junk/failing power supplies can sometimes nuke smd ceramic caps and cause them to crack like in this thread.

[chaphal3]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

It definitely popped exactly when I touched the heat sink it but thanks; I did in fact quickly test the power supply before then. I initially guessed that the PSU was gone due to condensation but all the voltages were perfect.

I realized I couldn't actually test whether or not that cap had failed open or closed because of other devices between 12v and ground. Anyway I've pulled it now; still can't test the cap because it crumbled (I think more of the de-install was done by the tweezers than the soldering iron). Dusting and reassembling for a test run now.

EDIT: Or were you saying I should retest the PSU after the incident? I can't remember if I did that so I'll do it now.

[chaphal3]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

PSU was good so I tested and it boots! Thanks a ton for the comments which got me to remove it and try again. My dad's about to place an order from AllElectronics, should I just grab a 47nF cap to stick in there? In theory, would I want to undershoot on capacitance to make up for the inductance of adding leads (trying to keep the frequency similar), or overshoot (trying to prevent it from becoming overdamped)?

I also threw some TIM in the shopping cart, I hope theirs isn't garbage.

[momaka]

Re: Advice needed: Dell 5150 blown ceramic SMD decoupling cap

I say just run the board without the cap. It's really nothing critical.

As for why that cap failed: it may not necessarily be ESD. Sometimes, these ceramic caps will develop weak spots in their body (either during manufacturing or when the component was installed on the motherboard). Thus, inserting enough pressure can cause the body to crack and the internal interweaved layers to short out to each other, causing the short-circuit you saw. (I had a neighbors home theater system go out when they moved just a mile down the road. Cause was the same: a ceramic cap on one of the power rails had become shorted, very likely due to the move. Removing it made the system work again.)

And yes, ceramic caps will in almost all cases go short-circuit when they crack, so once the cap fails, you need to remove it from the board.