Are the DIMM slots on my Z690 defective?

[sinmurv]

I have the following rig: • 12900K • DDR5 by 2x16GB T-Force Delta 6400MHz C40 • Asus Maximus Extreme Z690

*I have run into a snag. I bought the T-Force Delta 6400MHz C40 with the intention of overclocking it, but it has shown some really weird Memtest behavior, and I thought I’d ask the folks here since they seem incredibly knowledgeable about voltages and motherboard quirks. Here are my Memtest results (being run from the Asus BIOS) •

T1: both sticks in the recommended slots; 3 errors… I thought the sticks were faulty but investigated further •
T2: first stick in A2 pass •
T3: first stick in B2 FAIL, three errors from the get go (read on…) •
T4: second stick is in A2 pass •
T5: second stick in B2 pass (??) •
T6: first stick in A2 again pass •
T7: first stick in B2 pass again (??!! Previously this churned out errors) •

T8: both sticks in A2 and B2 churning errors again.

*What is happening here? Is it my motherboard or is it the RAM? Thank you Edit: yes, the RAM is on the QVL and I have XMP 1 turned on. I am still learning about overclocking so haven’t really figured out how to set voltages… could this be just low voltage?

[Silent_Scone]

Hello,

I assume all these aforementioned tests are at XMP?


1) Errors at XMP does not mean the DIMM or any other components are faulty. Overclocking never guarantees stability.

2) 6400 can be tough on 4 DIMM boards, least of all without manual intervention.



Try increasing VDD and VDDQ 50mV over XMP profile and setting MCVDD under Advanced Memory Settings to 1.3v

[sinmurv]

Silent Scone@ROG wrote:
Hello,

I assume all these aforementioned tests are at XMP?


1) Errors at XMP does not mean the DIMM or any other components are faulty. Overclocking never guarantees stability.

2) 6400 can be tough on 4 DIMM boards, least of all without manual intervention.



Try increasing VDD and VDDQ 50mV over XMP profile and setting MCVDD under Advanced Memory Settings to 1.3v

Thank you! I increased VDD and VDDQ but I cannot find MCVDD.

I did find Memory Controller Voltage but it is set to Auto 1.332V and cannot be changed. Is that what you are talking about? How can I change that? Â*

[Silent_Scone]

sinmurv wrote:
Thank you! I increased VDD and VDDQ but I cannot find MCVDD.

I did find Memory Controller Voltage but it is set to Auto 1.332V and cannot be changed. Is that what you are talking about? How can I change that? Â*

Advanced Memory settings should be visible at the very bottom of the Extreme Tweaker Page. Highlight auto and enter a higher value, try up to 1.4.

[sinmurv]

Silent Scone@ROG wrote:
Advanced Memory settings should be visible at the very bottom of the Extreme Tweaker Page. Highlight auto and enter a higher value, try up to 1.4.

thanks for the help! So I have decided to return the Asus Extreme and purchase a 1 DPC boardÂ*

Given how dense DDR5 is, 4 DIMS is nothing but a waste and the more I learn, the more I realize that the Maximus Extreme is pointless. Worse VRM design, worse memory performance, cheaper capacitors… what was I thinking buying the Extreme?

The Apex seems to be the perfect board, however the memory training and compatibility Â*is absolutely terrible compared to the MSI Unify X Z690

If I hadn’t started talking to you, I would never have learnt the severe handicaps the Extreme has. So, thank you!

[Falkentyne]

As Scone said, 6400 mhz on 4 dimm boards is very difficult, and you get into a problem where it's stable on one boot and not on the next.
Sometimes adjusting "slopes" and skews can help this.
Sometimes minor adjustment of voltages.

I haven't tried this, but sometimes, making sure you clean the RAM slots out with Deoxit D5 contact cleaner first can help. This has helped two people in the past, one who thought their memory slot was busted--it was just dirty or tarnished or something (though the question still remains whether his improvement would be long term or not). Another thing that may help is wiping the PCB with Deoxit D5, between the RAM slots and the CPU socket, however I have no idea if you need to do this on the front or the back side (this would require backplate removal, which is tricky with the ribbon and connector cables back there). One thing someone may want to try is wiping a layer of Deoxit D5 on the underside of the CPU surface when it's removed (enough to make a moist layer but not dripping) and then installing the CPU to see if that gives any increased headroom. One person who tried the washer mod (0.7mm) with the Apex found that the washer mod (which reduces socket pressure) caused MRC training failure at 6400 mhz and when he removed the mod, he regained stability, which in that case, meant that it was proper (and improper) contact pressure on the CPU pins and pads responsible for that.