Max XI F: BIOS 1902/1903 issues with XMP

[BizSAR]

After flashing BIOS 1902 or 1903 on my Maximus XI Formula, if I set it to utilize XMPI or II, the platform makes a number of attempts to boot, then boots into BIOS safe mode. I'm able to reproduce this even with all other BIOS settings at default.
Flashing back to 1802, I have no issues.
CPU: i9-9900KF
Memory: Crucial Ballistix Tactical Tracer RGB 3000 MHz, 4 x 8GB CL15 BLT8G4D30AET4K
Can provide more stats if needed.

[BizSAR]

vvoid wrote:
Well ok, I was assuming all the Z390 boards have this setting, Strix Z390-F here.
For me it's under PCH-FW Configuration in Advanced settings...

All the Maximus XI Formula has is a PTT setting and you can either enable it or disable it.
The issue we are seeing occurs regardless of the PTT setting.

ASUS is asking to RMA my board, but if it works perfectly fine using the 1802 BIOS, I see no point.

Anyone else have any luck with them?

[Silent_Scone]

BizSAR wrote:
After flashing BIOS 1902 or 1903 on my Maximus XI Formula, if I set it to utilize XMPI or II, the platform makes a number of attempts to boot, then boots into BIOS safe mode. I'm able to reproduce this even with all other BIOS settings at default.
Flashing back to 1802, I have no issues.
CPU: i9-9900KF
Memory: Crucial Ballistix Tactical Tracer RGB 3000 MHz, 4 x 8GB CL15 BLT8G4D30AET4K
Can provide more stats if needed.

The board is failing training because the applied memory presets are unstable. What Q-Code does the board hang on?

[X-Terminus]

Hello all, I've been writing this in 3 different threads. I hope this is the right thread.

Silent Scone, please take a look at this issue.

I have Maximus XI Extreme, Intel i9-9900K. I can confirm that XMP doesn't work on this BIOS 1903, whether it's XMP I, or II. I use Trident Z Neo F4-3600C18Q-128GTZN. XMP used to work just fine in 1704, which is the version I used before I updated to 1903. Whenever I tried using the XMP, it always POSTs in safe mode, instead of a normal POST.

This RAM that I have, while not on QVL in my mobo's QVL list, it's QVL in GSkill's website. Also, the fact that in BIOS 1704, both the XMP I and II profile work perfectly with no issues at all, shows that this BIOS 1903 is the problem.

This is GSkill's website that lists it as QVL:
https://www.gskill.com/qvl/165/326/1582266217/F4-3600C18Q-128GTZN-Qvl

[Silent_Scone]

Silent Scone@ROG wrote:
The board is failing training because the applied memory presets are unstable. What Q-Code does the board hang on?

Hello,

Without basic information like this replication will be harder. Make sure you provide the Q-Code. If the board is failing POST stress tests, regardless of the values that are causing signal margins to encroach on one another, the overclock is unstable - that's the short answer.

[BizSAR]

Silent Scone@ROG wrote:
The board is failing training because the applied memory presets are unstable. What Q-Code does the board hang on?

When the issue is occurring, the system will start to boot, the orange light stays on, a click is heard and the system tries to POST again. This repeats for 3(4?) times and eventually starts in BIOS safe mode. When the click is heard and the system starts/tries to POST again, the Q-codes go by so quick it is hard to tell. They don't stop long enough to see and it's hard to tell upon which exact code it would be.

[Silent_Scone]

BizSAR wrote:
When the issue is occurring, the system will start to boot, the orange light stays on, a click is heard and the system tries to POST again. This repeats for 3(4?) times and eventually starts in BIOS safe mode. When the click is heard and the system starts/tries to POST again, the Q-codes go by so quick it is hard to tell. They don't stop long enough to see and it's hard to tell upon which exact code it would be.

The LED should stop on the last code momentarily before reattempting training, which will be 00. I don't have the board here to test, unfortunately.

[X-Terminus]

Silent Scone@ROG wrote:
The LED should stop on the last code momentarily before reattempting training, which will be 00. I don't have the board here to test, unfortunately.

Hello Silent Scone. I don't remember the code exactly, but I thought I saw S5 or 5C. Maybe 5C, because somehow in my head, when it happened, I had a feeling that it might have been "short circuit" lol. So, right now, my computer is running my old CMO BIOS settings from when I was still using 1704. I'll try to reproduce the errors here:

1. Start from scratch, default settings here.
2. Without touching anything else in the BIOS, I go directly to that XMP profile selection. I choose either XMP I or II (the outcome doesn't matter, same error).
3. Without touching anything else after changing to XMP I or II, I restarted the system, saving changes.
4. Then memory training begins, but this time, it takes much longer time than the usual XMP training. The number on the board keeps on changing, and sometimes I see the light keeps on switching between orange, to red, though the majority of the time, it's in orange color. Then, I see that S5 or 5C code (again, I might be wrong here). The reason why I remember that S5 or 5C (if I do remember it correctly), is because they are the number that stayed the longest, among all other numbers.
5. And then, the motherboard POSTs, but this time, it says that it's running on safe mode. And if I check the BIOS settings, if I remember correctly, while some of the XMP values are there (like frequency and some timing, I don't quite remember this part, TBH), the VCCIO and VCCSA are at around 0.9V, which is impossible, because I know that I needed at least 1.1V on each of these for my system to be stable (I run MemTest 86 3 cycles, and passed at 1.1V of VCCIO and VCCSA).

So, in desperate attempt, I used my old version 1704 CMO file, loaded it on the BIOS from USB, my computer works. But still, something is VERY wrong if you can't use XMP, especially the RAM maker, GSkill at this point listed my mobo as QVL for this RAM.

I wish I can help you further here, by reproducing this error. But, I already felt that I'm lucky to be able to use my BIOS settings (CMO) from 1704. I'm afraid if I mess with my BIOS again in an attempt to reproduce this issue, I might get in trouble.

I hope that helps. If you have any questions, please let me know. Thanks for looking into this!

[X-Terminus]

To clarify, I use the CMO file from 1704 on the currently installed 1903 BIOS, and it works. My CMO file is based on the XMP II profile of my RAM, except that I lowered the timing further. My RAM's XMP is 3600 MHz Cl 18-22-22-42, 1.35V. In my saved 1704 CMO, while the frequency stays the same, I lowered the timings to 16-20-20-40 at 1.4V. Surprisingly, while the XMP I and II didn't work in this BIOS 1903, my CMO which I took from 1704 works fine. This is what I'm using now. After applying this 1704 CMO on this 1903, I did a full run of MemTest 86, and it passed.

[BizSAR]

Silent Scone@ROG wrote:
The LED should stop on the last code momentarily before reattempting training, which will be 00. I don't have the board here to test, unfortunately.

BTW, I do see that the last code seems to be "69", and I'm not joking. 🙂

[BizSAR]

@X-Terminus

That's the exact issue I and others are having. I submitted a support ticket for the issue and am working with the support folks on reproduction, etc. I'd suggest submitting a support ticket, that way ASUS can see how many people are having the same issue and help us. They have been really responsive and helpful, after a few mis-steps.

I'll post updates on the issue as I have them.