ASUS KRPA-U16 Q-code 90 on boot (not posting)

[boomheadshot]

Hi guys.

So I have had an engineering 2S1404E2VJUG5 sample running on a KRPA-U16 board, and I was mining RTM, until it just stopped working all of a sudden.

Now, when I start it up, I get Q-code 90 and this is what's on the screen:


At the bottom, it says "System Initializing" on the left side and "62" on the right side of the progress bar (62/100)

I've tried reflashing the BIOS with a programmer, clearing the CMOS (and doing a hard reset without power and the battery taken out for a while), trying other RAM, trying 1 stick of RAM, but to no avail.

My CPU temps never went above 63C, I was running @ 3 GHz with 1.05V on the core (which isn't on the limit). There were 2 huge (~30cm) fans blowing on the test bench from 2 angles,

There is very limited info on the internet about this problem.

I've tried turning off BMC via the jumper, and it still gets stuck at the same 62/100, but now it says "BMC DM_LAN1 IP Address: 192.168.1.111" on the right side beside "62"




Can anyone enlighten me?

Thanks in advance!

Edit: just tried booting without 8pin power for the CPU, same result.
I guess I'm screwed ...or am I?

 

[RageBone]

maybe you are screwed.

Postcodes are more of a guessing game at this time and both, code 90 and code 62 don't point to anything in your control.
90 May be paraphrased as " Fatal Error, Link between CCDs not Trained."
62 as "Error Mapping SMN Fuse to Register"


My recommendation is to give the CPU a visual inspection, there likely won't be anything visible but it is a good thing to do anyway.
Since its a rather "desperate" situation, freezing it for a night might recover it or might not do anything.


A recommendation i got from someone was to try a good CPU, then set the Bios to disable most Cores and CCDs, swap CPUs and hope that it works that way.

I tried that myself on a H11SSL and had no luck, so i assume i killed the IOD on mine.

 

[boomheadshot]

maybe you are screwed.

Postcodes are more of a guessing game at this time and both, code 90 and code 62 don't point to anything in your control.
90 May be paraphrased as " Fatal Error, Link between CCDs not Trained."
62 as "Error Mapping SMN Fuse to Register"


My recommendation is to give the CPU a visual inspection, there likely won't be anything visible but it is a good thing to do anyway.
Since its a rather "desperate" situation, freezing it for a night might recover it or might not do anything.


A recommendation i got from someone was to try a good CPU, then set the Bios to disable most Cores and CCDs, swap CPUs and hope that it works that way.

I tried that myself on a H11SSL and had no luck, so i assume i killed the IOD on mine.

I've inspected it, visually the contact pads were completely fine

 

[RolloZ170]

i do expect your CPU is dead.
2S1404E2VJUG5 and all other 2S are lab samples that was not intented to work a long time.
if soldered at all then unclean, more does not have to be because it is ES.

 

[boomheadshot]

Got a good deal on a 64 core ZS and took the gamble, it paid off. System back up again after a CPU switch.

TLDR: Switching the CPU after Q code 90 worked!