820-3437 1.4 Kernel Panics/Shut Downs/Freezing Issues

SMMRepair

Member
I've noticed a lot of the 2014 (specifically) 820-3437 boards with shutdown/kernel/freezing issues that we've been attributing to CPU issues (or rather, something we can't seem to figure out). We don't notice this as much on the 1.3 3437 or the 00165 boards. Has anyone noticed this, and if so, is it possibly related to the tantalum caps on S5_HS_COMPUTING_ISNS (C7619/20/21/22)? I'm wondering if one of those caps failing is causing a drop-out of either ppbus or 1v05. Somewhat similar in nature to the 2850/2610 cap failures, maybe?

Any ideas? At first we suspected BIOS or SMC issues (one or two were), but the vast majority seem to be "deeper", i.e. CPU or something else. Just curious, and this is more of a general question than a specific one, sorry. Thanks in advance!
 

smiba

New member
If you want to be sure about the power lines / tantalum capacitors attach an oscilloscope to that line and keep eyes on whats happening as it crashes. (Have a big enough buffer so you have time to manually trigger it)

What does the kernel panic give you? I've yet to come across a 820-3437 board with this issue
 
Last edited:

SMMRepair

Member
We have over a dozen here (3437, 1.4, no liquid, generally the same symptoms--power on fine but either cut-off within a few minutes or freeze/restart with kernel). Confirmed the issue is not SPI ROM or SMC related (at least on the 3 that I tested in-depth). When one came in with those symptoms, we just set it aside intending to come back when we know more. Haven't tested each and every one (they came in with the notes on them), but we have seen tons. Haven't seen it on the 1.3 or newer 00165s yet. Was just wanting to put a feeler out there in case it's a common issue that we shouldn't spend time trying to figure out (these aren't for customers, so time constraints in that regard). I can provide the kernel info today or tomorrow, when I have time to get one thrown into a case. If it's actually a CPU issue, I won't bother with them, etc. Yeah getting them to fail whilst checking with an oscilloscope will be the best bet, I imagine, if that's the culprit; was mainly curious if anyone had come across and solved this issue at all. Having so many boards with the issue, I thought maybe someone had come across it. Thanks for the input, Smiba!
 

dukefawks

Administrator
Not seen this problem. Only sort of common issue I have seen were corroded traces from the clock chip X-tal by left over residue from manufacturing.
 
Top