[Solved] X670E-A QVL Memory after setting EXPO, won't reboot

[axsdenied]

Late last year I posted the below post on the forums looking for help.  The thread is locked now so I wanted to post the solution here.

Many folks believed the issue was that I needed to tweak the memory settings, suspecting something changed in the later bios versions.   I argued that is the point of using QVL'd memory, so I don't have to worry about that.  Others believed QVL memory to be more of a guide than anything else.

SOLUTION: After tweaking memory speeds, voltages, etc. and seeing no change in behavior no matter what I found one thread suggesting it was the cpu.  The person had replaced via RMA and it worked as it should.  So that's what I did.  I RMA'd my CPU and behold everything is working as it should.  Running EXPO 6000 on my same board with no problems and it can reboot.

[Copy of Old Post]

Problem:  X670E-A QVL Memory, after setting EXPO, won't post after reboot.  Stuck in memory training with yellow light on motherboard.  Waited over 30 minutes.

Observations:

1. What is weird is AFTER setting EXPO, if I shutdown the machine and remove power (no this does not clear CMOS) and I plug it back in and it powers back up, with EXPO settings and is 100% stable.  Zero issues...until I reboot.  It's only the act of rebooting that causes this phenomenon.
2. This issue did NOT occur on Bios revisions of 2007 or older.  2124, 2204 and 2308 all exhibit the issue.

I've opened a support ticket with Asus and it's supposedly being "escalated" but I honestly don't know how far I'll get.

Anyone else experience this?

[gupsterg]

Reading your past posts you are using F5-6000J3238F16GX2-TZ5NR, 6000MT/s is OC.

EXPO is just a RAM SPD profile, which doesn't mean CPU IMC will work at those settings, as each CPU IMC can vary.

QVL also doesn't mean the CPU IMC you may have will allow you to use said memory at xxxxMT/s.

I have 4 CPUs at present. If used on my TUF X670E with same RAM and evem same setup for RAM profiles, outcomes can be different.

I think SIlent Scone highlighted most of above, link.

If you are on 9000 series (not got 7000 series, so no idea), I would say FW tweaks are still going on for RAM OC (ie>5600MT/s). 

Latest ASUS beta for my board has changes to marked RTT resistances. Both cases ProcODT/Resistances values are [Auto], 2x16GB single rank in use, same CPU, etc. AGESA and CPU SMU FW hasn't changed, there seem to be tweaks to ASUS FW in regard to memory. I also cross referenced ZenTimings with Ryzen Master and they match, so not read out error in my opinion.

 

[axsdenied]

Thats actually super interesting!

So what I think you're saying, in this instance, is that:

Bios update probably DID change something creating a scenario where I was no longer able to OC in part because of the changes in the BIOS in combination with whatever my CPU'S IMC is capable of.

So while changing my CPU and running on the same board, with the same BIOS now works as it used to what potentially allowed this was not that my previous CPU was faulty but that this new one has a variance of the IMC that happens to work?

I think it's more likely, given the fact that I couldn't overclock to ANY speed above 4800 that my previous CPU's IMC either degraded over time do to heat, voltages, or faults that was exasperated by the BIOS updates which is why the replacement is working as intended.

Also remember, my system WAS stable when it booted and the only issue was rebooting the system. Another symptom, that I don't remember if I articulated in my other post was that after it attempted to reboot the fans would spin up to 100%, then drop and spin up again in an infinite loop.  I don't think i've seen that behavior (not saying it couldn't happen) when ram OC is unstable.