Z440 973 Global Reset: DRAM Corruption Error May Be Observed

[LightGuy48]

I ran memory test for several days and it completed successfully with zero errors.

After that turned up negative is when I started Googl'ing the issue which is what turned up the HP KB article that was part of my original post.

[LightGuy48]

Any other suggestions?   I hate to just start replacing RAM when I can't prove that the RAM is bad, so far tests have provided no failures and there is this KB article saying it was supposed to be fixed but doesn't appear to be fixed.

Is it possible the Intel Mgmt firmware didn't update properly?  

RAM is not cheap for these workstations so without being able to prove there is bad RAM feels like a serious waste of money.

[DGroves]

before you buy ram test with just one stick,

if the test does not trigger the error, add ram back one stick at a time

you might want to also consider reinstalling windows and test without installing software

if the intel AMT did not update correctly, you would know it as the system would not allow you to enter the AMT setup, and other strange boot related errors would appear with win 10

[LightGuy48]

I ended up buying some new Kingston RAM for the machine and it fixed the issue.  I never was able to nail down which stick of RAM was causing the issue because it would go days without happening and when running single sticks it never seemed to happen.  

[DaveP12]

I'm currently going through exactly the same process with my Z440, repeated 973 errors. I'm brought the BIOS up to date, installed the latest ME Softpaq but it still does it, sometimes not for days, sometimes 3 times in an hour, with no apparent usage pattern triggering it.

Annoyingly it's a refurbished unit with a RTB warranty, and I'd rather be able to collect some sort of evidence of what's going on to try and solve it remotely than be without the unit for weeks. Any suggestions for a tool that could test and log any defects?

(it actually shut down on me whilst typing this, but thankfully google chrome remembered 🙂 )

[DGroves]

if the system shuts down every 30 min, the "ME" firmware is corrupt otherwise look at this bulletin from HP

https://support.hp.com/us-en/document/c04926792

the intel AMT should be disabled in a non domain environment as it's a known security risk, and has no useful function outside of a domain setting for most users

[Mis66]

Hi.

I hope it's allright to "hi-jack" the thread a little bit. I've got the same problem with a Z440. Wanted to update from 4-core to 8-core, and stayed within the official specs; E5-1630v3 to E5-1680v3.

It was working allright with an E5-1630v3 and 128GB ECC RAM and with up to date WIN10 Pro.

I upgraded to a used E5-1680v3, and now most of the time it doesn't go through POST at all. It starts all over again.

Sometimes it manages to get through POST to WIN10. It does not stay up for long though, it freezes after a minute, two or sometimes a little bit later. I could inspect HW-monitor readings though. All showed up as they should be, all cores visible and thread too. Memory allright, temps allright and so on.

Mostly then I power up, it gives the ME 973 -error and after that it stays in a power loop, restarting continuously.

I have read suppoert document: https://support.hp.com/us-en/document/c04926792 and acted accordingly.

I have tried another win10 boot source (exact clone actually).

I have tried different RAM (16GB non ECC, two sticks).

I have reseated the processor 3 times now.

I have installed the E5-1630v3 back couple of times, and everything goes as normal, really stable and a fully working system.

I have the latest BIOS M60 02.50

I have the latest ME FW 9.1.45.300

I have a bootblock date 07/02/2014

I have run all BIOS tests to 1680v3 system, and every one passes. I have run processor stress tests also from BIOS, temps show 72-73 °C max. When I am within BIOS machine works fine, no freeze or hicups. It reports everything as it should.  The seller of the processor is sure it's a working piece. He will get it back though if I decide so. Actually return is allready accepted. I just wouldn't want to return it if it's ok. These (E5-1680 v3) are not very easily found for a reasonable price today.

I know it looks like a bad processor at first, but all the BIOS tests are good though?

About disabling ME as suggested by DGroves:

I have read about ME problems, and it should be disabled. But I have read too that it can't be disabled, at least not entirely. And the disabling itself might be more of a "hack"-like procedure.

ps. I have read this helpful forum a lot lately, a couple of years. A lot of really helpful people around here. Thank You.

[DGroves]

973- Global Reset Cause: ME Initiated Power Button Override

in your case this is most likely ram related,.....

either the cpu memory controller is the cause or you have a Dimm issue

try removing the cpu, cleaning the bottom pads with 98% pure alcohol and do the same with the dimm modules

then only insert one dimm (in the correct slot!!) and power on see if you still have the error if so replace the original cpu

if the error goes away the issue is with your new cpu 

[Mis66]

It's just that with E5-1630 RAM works ok like everything else. So I have to think like you, it' should be the memory controller of E5-1680 which is faulty. BIOS memory test do pass though with E5-1680 too, so who knows for sure (the test takes a very long time to pass).

I believe I've got the right chemicals needed, thermal paste removal and IPA alcohol for last touch up too. At least I believe the processor seating is allright. I will carefully clean all the RAM modules too as suggested, and try only one module in slot 1. I've got an 8 GB module left over from earlier upgrade, so I can make it like the official Z440 quickspecs table suggests on page 19: in slot 1 one 8 GB module.

https://www8.hp.com/h20195/v2/GetPDF.aspx/c04400038.pdf

It does not show a 32GB alone in slot 1. I've got 4*32GB now filled according to specs (slots 1,3,6,8).

I will report later,

Thank You !

[Mis66]

Unfortunately this either did not help:

- Cleaned the 8GB, known as good, RAM module with IPA.

- Inspected the memory slot no.1 with magnifier glass, looks clean

- Installed 8GB RAM in slot 1.

BIOS RAM tests passed again, same for the processor and other tests too. Managed to boot to windows 10 Pro. Got CPU-Z open, and everything looks fine and dandy, the voltages around 1V, speed around 3200MHz. After a while (few minutes) windows hang. Soon after it gave me a BSOD saying:

Your PC ran into a problem and needs to restart .. ..0% complete

Stop Code: CLOCK_WATCHDOG_TIMEOUT

Computer never did the restart though, so I forced the shutdown.

www.windows.com/stopcode 

Could not find the exact stopcode from there.