I am absolutely stumped: Possible CPU Failure?

zprichard

Prominent
Feb 8, 2018
7
0
510
Hey guys - this might be a bit long, sorry it's rambling but I am just stumped.

So we have an editing machine:

i7 4790K
GIGABYTE GA-Z97X-UD7 TH REV.1.0 mobo
32 gigs of RAM
GTX 770
etc etc

Apparently, unbeknownst to me, it started acting up a couple of months ago. When you powered on, it would power cycle for about 20 minutes before finally coming on fully. And by power cycle, I mean it would start for a second or two then cut off and try again.

It never even got to POST and the LED on the mobo never even got a chance to light up.

We tried to isolate the situation by:
Removing motherboard from case and putting on cardboard to make sure no shorts were happening.
Removed all RAM and tested each stick.
Removed ALL components and just start up with nothing but the PSU, CPU/Cooler, and once with RAM and once without.

Nothing. Same thing. Bios light clicks on, machine powers off.

We reset CMOS as well.

Then I saw the BIOS switch and switched to the backup. Yay it seems to have worked! It gets into BIOS on the first bootup... no power cycling.

I take this moment to flash the Main BIOS to the latest revision and then bam... back to not working. Now, no matter how long you let it power cycle it will never come on.

Uh oh.

SO we order a new mobo replacement. Unfortunately it's expensive because it's old but we *need* that Thunderbolt capability.

Mobo comes in, replace all of the components... moment of truth.... same thing.

Powers on briefly. BIOS light comes on... *click* rinse repeat. Never gets to post.

We fiddle with some more cables... SURELY it's a short I think (although when we pulled everything out of the case that should have remedied that). Then voila... it comes on!

But we can never repeat the solution. If we reset the computer it's back to the same problem. Find another cable to fidget with and make it turn on? Great! Except it's just never repeatable so we can't say it's a short.

So now we're sitting here with a motherboard with no bent pins, a brand new PSU attached, a CPU sitting there with NO Cooler just to try it (what else do we have to lose) and it's still doing the same thing. Occasionally with this setup (aka CPU and no cooler) and the old PSU, if I pushed the mobo power cable just right it'd turn on and get to BIOS... but again not repeatable.

And this point... going through all of this (and I'm sure there's a lot I left out) it's *got* to be the CPU right? I just don't want to order a $350 part as a "test" and get stuck with it.

I don't know... my feeble brain is exhausted and just out of ideas.

We've also made sure no pins are bent and no thermal paste is accidentally where it shouldn't be.

Also all capacitors on both mobos seem the same.





 
Welcome to Tom's!
Unless abused, (severely overclocked, etc) CPUs usually do not exhibit intermittent operation. They either work or they don't, and once they don't, they never work again. When you replaced the PSU and motherboard, did you also replace the other cables(SATA,etc.), or did you use the existing ones?

What type(brand,make,model) HDDs are in the machine? How old are they? (Are they backed up?) Do they work when placed in another machine?
 
intel has a Processor Diagnostics Tool utility you can download & run to check the cpu
https://www.intel.com/content/www/us/en/support/articles/000005567/processors.html

i've got a similiar unexplainable problem that when i was researching one of the windows downloads (an intel download actually from windows) and found this thread - apparently a lot of folks have been having odd issues, including failing to boot, booting to black screen etc - at the link below be sure to expand the 4+ replies - but apparently folks are rolling back intel drivers

https://answers.microsoft.com/en-us/windows/forum/windows_10-hardware/display-driver-failures-after-windows-update-intel/8c90cba4-4d9d-4d13-8fb4-1d55a1a2801c?auth=1

fwiw
 

Eximo

Titan
Ambassador
Silly question. During all this testing did you ever try a different outlet or circuit?

Ground loops and other wiring problems can cause a PSU to trip over and over on one of its failsafes.

I agree though, if you got a new motherboard and PSU, swap ALL the cables.

I once troubleshot a system for an hour, one of the SATA peripheral connectors had shorted internally, but not enough for a full connection Wasn't hooked up to anything, just enough voltage drop that the system would mostly work, and then fail to load Windows when the CPU started demanding power.
 

zprichard

Prominent
Feb 8, 2018
7
0
510


There are two Samsung 850 Pro SSDs and two WD Blue 2 TB HDDs... they're about 2 or 3 years old I believe.

However I did unplug even all of those from the mobo as well. One of the few times it did manage to boot up, it gave me the "no bootable drive found" as expected.






See above. Yup tried with no drives installed. Didn't fix it. :(



Not silly at all and yup we definitely moved to a different circuit for that very reason! Still didn't help.

We didn't swap the SATA connectors but we did actually remove them all and that didn't work (except for one random time as I mentioned above... probably unrelated).

Right now it's a Corsair 750W PSU. It's modular except for a hardwired mobo and cpu power. So those are the only two plugged in. And there's zero else plugged in (well except for a CPU) and it just won't do anything. Stumped.

I *hate* not being able to figure this out ha!

Thanks for the replies so far everyone!

 

zprichard

Prominent
Feb 8, 2018
7
0
510


Thanks! Unfortunately I can't get the system to even reliably POST so I'm not sure if I can get in to run this, even with the Linux bootable version.

 

zprichard

Prominent
Feb 8, 2018
7
0
510


Nope :( switched to the backup and nada. Still does the same thing. I'm fairly sure the building is just haunted.
 

zprichard

Prominent
Feb 8, 2018
7
0
510


Thanks to you and everyone else's replies!