Maximus XI Formula DRAM won’t POST with DIMM in Channel B

[toolbear]

TL;DR - cannot POST with DRAM in B channel.

BIOS: 0903 x64
CPU: i9-9900K @ 3.60
DRAM: G.Skill 4x8 @ 3600 quad kit
DRAM Part: F4-3600C17Q-32GTZR


DRAM is a quad kit. I verified sequential serial numbers. Listed in the ASUS QVL as well as the G.Skill QVL.

G.Skill tested specs: https://www.gskill.com/en/product/f4-3600c17q-32gtzr

Physical arrangements tried:

A_2 .................... SUCCESS
B_2 .................... FAIL 

A_2 + B_2 .............. FAIL 

A_1 + A_2 .............. SUCCESS
A_1, A_2, B_1, B_2 ..... FAIL 

B_1 .................... FAIL 

A_1 .................... SUCESS (no stress test performed)


Those arrangements that worked were able to POST, boot Win 10, and pass about 1 hour of AIDA 64.

DIMM serial numbers are:


xxxxxxx6309
xxxxxxx6310
xxxxxxx6311
xxxxxxx6312


For the A_2+B_2 arrangement I even tried ensuring sequential numbers (i.e A_2:6309, B_2:6310). For single arrangements I made sure to try the same DIMM that was stable in channel A (i.e. both A_2:6309 and A_1:6311+A_2:6309 were stable so I tried just B_2:6309 and couldn’t POST).

For completeness, these are baseline BIOS settings. I load these from a user profile before applying other BIOS changes:

(Load Optimized Defaults)
ASUS Armoury Crate:
  Download & Install ... : disabled
Secure Boot:
  OS Type: Windows UEFI mode


I’ve tried the following BIOS settings 

Ai Overclock Tuner: XMP I
  XMP: XMP DDR4-3603 17-18-18-38-1.35V
DRAM Frequency: DDR4-3600MHz
DRAM CAS# Latency: 17
DRAM RAS# to CAS# Delay: 18
DRAM RAS# ACT Time: 38
DRAM Command Rate: Auto
DRAM Current Capability: 130%
DRAM Voltage: 1.35

With the above settings I also tried DRAM Voltages of 1.4v, 1.45v, & 1.5v

With any DIMMS in B channel using any arrangement I get 3-5 POST attempts before it finally gives up and reports code:AC with “Detect Memory” shown on the OLED.

I’m new to RAM overclocking. I relied on this thread for ideas so far:

https://rog.asus.com/forum/showthread.php?109972-Asus-X-hero-G-Skill-F4-4133C19D-16GTZR-problems/pag...

Because the A_1+A_2 arrangement works, but not even a single DIMM in B_2 works, does that indicate a bad mobo? Or is this something I might be able to get working by tweaking more BIOS settings?

[FornaxLacerta]

toolbear wrote:
TL;DR - cannot POST with DRAM in B channel.


I’ve tried the following BIOS settings

Ai Overclock Tuner: XMP I
  XMP: XMP DDR4-3603 17-18-18-38-1.35V
DRAM Frequency: DDR4-3600MHz
DRAM CAS# Latency: 17
DRAM RAS# to CAS# Delay: 18
DRAM RAS# ACT Time: 38
DRAM Command Rate: Auto
DRAM Current Capability: 130%
DRAM Voltage: 1.35

With the above settings I also tried DRAM Voltages of 1.4v, 1.45v, & 1.5v

I have a Formula XI and also couldn't get stable with a similar mem kit at the default XMP values. I increases voltage to 1.4v from 1.35v and manually set timings to 19-19-19-41 to get stable. I'd try incrementing your timings up a bit.

[toolbear]

FornaxLacerta wrote:
I have a Formula XI and also couldn't get stable with a similar mem kit at the default XMP values. I increases voltage to 1.4v from 1.35v and manually set timings to 19-19-19-41 to get stable. I'd try incrementing your timings up a bit.

Thank you for the reply.

Were you able to POST with all 4 DIMMs using default BIOS? I cannot. I have to remove all modules from B channel, boot into BIOS, change timings or voltage or whatever, power off, seat modules in B_1 & B_2, find that it didn’t work, repeat. I cannot even boot when all I have is 1 DIMM installed in either B_1 or B_2.

I’ve since tried 2133 MHz and still cannot POST with any DIMMs in channel B.*

[FornaxLacerta]

toolbear wrote:
Thank you for the reply.

Were you able to POST with all 4 DIMMs using default BIOS? I cannot. I have to remove all modules from B channel, boot into BIOS, change timings or voltage or whatever, power off, seat modules in B_1 & B_2, find that it didn’t work, repeat. I cannot even boot when all I have is 1 DIMM installed in either B_1 or B_2.

I’ve since tried 2133 MHz and still cannot POST with any DIMMs in channel B.*

I actually had issues posting with DIMMs in B until BIOS 0906 dropped. After that I could then post without issue with all 4 DIMMs populated but was still getting occasional memory errors running memtest and BSODs and app crashes in Windows. Then I adjusted timings to get stable. Might be a matter of a BIOS update being required... or one of the other suggestions in the thread. Good luck!

[vitalik_t]

toolbear, you resolved the problem or opened a support ticket?

[Silent_Scone]

Remove the CPU and check for any bent socket pins, reattach the cooler whilst checking for even mounting pressure. If the motherboard is in an enclosure, make sure the stand offs are not overtightened surrounding the CPU and memory array.

[Murraymint]

I had the same problem and found a bent pin under the CPU, chances are that you have the same.

[vitalik_t]

I have a very similar problem which is under Asus engineers investigation. My suggestion, before reseating a CPU: inside BIOS disable DIMM B slots, turn of the PC, occupie DIMM B slots, go to BIOS and check if DIMM B sticks are recognised.

[toolbear]

Silent Scone@ASUS wrote:
make sure the stand offs are not overtightened surrounding the CPU and memory array.

My EKWB water block was overtightened. Loosened it and it POST'ed no problem.

In case it was uneven tightening, I loosened all the posts and tightened them an equal number of turns. I was able to reproduce the failure eventually, so I backed off a few turns on all the threads and called it good.

It's an i9-9900K so no pins to bend. Also liquid cooled loop with rigid tubes so reseating the CPU would be quite an ordeal. I'm glad loosening worked.

Next occasion I have to drain I'll dismantle a bit and check the seating, too.

My temps are a few degrees hotter now, but I think that's explainable by the CPU having to control 4 DIMMs instead of 2.


Thanks for the advice that lead to the fix!

[Silent_Scone]

toolbear wrote:
My EKWB water block was overtightened. Loosened it and it POST'ed no problem.

In case it was uneven tightening, I loosened all the posts and tightened them an equal number of turns. I was able to reproduce the failure eventually, so I backed off a few turns on all the threads and called it good.

It's an i9-9900K so no pins to bend. Also liquid cooled loop with rigid tubes so reseating the CPU would be quite an ordeal. I'm glad loosening worked.

Next occasion I have to drain I'll dismantle a bit and check the seating, too.

My temps are a few degrees hotter now, but I think that's explainable by the CPU having to control 4 DIMMs instead of 2.


Thanks for the advice that lead to the fix!

No problem we've all done it at some stage.