cancel
Showing results for 
Search instead for 
Did you mean: 

3 Bios Bugs with CROSSHAIR VI HERO

Lextra
Level 7
Note: I'm using a Ryzen 3600 so I'm not sure if these bugs only affects Ryzen 3000.

Bug #1: Can be replicated every time; Is related to the CAS memory latency.
It won't ever accept odd numbers, only even ones. No matter if a D.O.C.P. profile was loaded or if a manual value was entered. Disabling the "Gear Down Mode" fixes this. In my opinion, there should be a check in the D.O.C.P. profile, where it automatically switches Gear Down Mode to "Disable" if the CAS value is an odd number.

Bug #2: This could actually be quite dangerous to the CPU. It causes the CPU fan to turn off.
Now, hear me out here for a second, as I couldn't get this bug to trigger with just synthetic benchmarks. It only happened after playing games for a while. It also seems to happen with any kind of fan profile as I had it happen with the fan on auto as well as with a custom fan curve.
I had HWiNFO64 open in the background so I could monitor temps etc. After closing the game, I saw the fan was stuck at ~1400 rpm (Rated for 200-2000) even when the CPU temperatures already settled to 40°c, which I knew from previously monitoring should have been around 500 rpm.

So I started up the CPU-Z benchmark (in stress test mode), to try and get the fan speed to update. The fan speed did indeed change. But to my surprise, it kept actually reducing in speed, until it turned off completely. Temps reached over 70°c very quickly with the fan still being off. (I have an open benchtable as a "case" so I also checked if it was just a display bug of hwinfo, but no, the fan actually stopped spinning). So with the temps still increasing, and the fan not looking like it would turn on again, I stopped the benchmark and all of a sudden the fans turned on again. They went back up to the 1400 rpm they were previously stuck at. Only a PC restart could restore the proper behaviour.

I hope you see how dangerous this bug could be and find a way to fix it.

Bug #3: A cold-boot-bug when CSM is disabled + the LED below the postcode stays on permanently, in bright white, even with all the LEDs in the bios being set to "Stealth Mode".
I'm not really sure if the cold-boot-bug is related to CSM, or if it also happens with CSM enabled, but having CSM turned off reduces reboot times from 16.5s to 14.5s, makes the bios display in the proper 4:3 aspect ratio (instead of being stretched to fill the screen) and fixes the LED below the postcode to be stuck in bright white when booted into windows, so having it turned off is actually the only valid option here for me.

Every time I cold boot my PC (after ~8 hours of being cut off from any electricity source) the first boot of that day will result in windows 10 getting stuck when the spinning animation appears. Simply hitting the restart button on the mainboard makes it boot properly again.

Bug #4: The VDDCR SOC Voltage control inside the Advanced > AMD Overclocking settings does nothing. I set it to as low as 100mv which shouldn't be possible to boot with, but it still worked. AMD Ryzen Master reported the soc voltage as 0.1.
278 Views
16 REPLIES 16

Ravwyn
Level 7
Hmmm.... this is clearly a UEFI bug and has, indeed, nothing to do with anything SW related.

I'm running in issues now myself - with a 3600X on this board. I was on 7201 initially - but my USB drives kept freezing during file transfer (I build an m.2 usb stick with a SATA based m.2 ssd for... well... "larger files"). It took me a while to figure this one out - and it seems related to PCIe bus issues. Windows keept logging weird RAID and disk issues - I'll spare you the madness that ensued. Bottom line was, only a complete bios update solved this issue (for now, further testing follows tomorrow). I found 7304 here on these forums, it is AGESA 1.0.0.3 (no suffix) based so... fingers crossed. D/L link is here if you want to skip the slooooow forum server. 😃

Since I cannot verify the origin of this one - caution is advised. I had issues and therefore had to try it - others: be vigilant.

Ravwyn wrote:
here if you want to skip the slooooow forum server. 😃


Thanks for the link.

Edit: Just got the cold-boot-bug again. Definitely still around. Please, someone look into this. The only settings I've changes are:
CSM > Disabled
LEDs > Stealth Mode
CPU Fan > Custom

After many tries to fix the PWM issue on my own, I have found a little workaround which worked so far without problems.

For this workaround I changed all my PWM fans from PWM mode to DC mode in the BIOS.
When it is changed, the minimum fan speed on DC mode is 60%.
To get the possible options for lower fan speeds, you need to run the Q-Fan Tuning which will take some take some minutes. After it's done, you can change your fan curves to the PWM one and also have the option to turn fans off on lower temperatures via the BIOS.

The only downside is that fans will run slower on lower fan speeds but the maximum fan speed will be the same like the PWM one. Another downside is the effiency but according to the power consumption of case fans, it can be ignored in my opinion.

EDIT: The workaround is still affected by the bug but somehow running the fans in DC mode delays it better

6kbyte
Level 7
I just ordered a ch6 for a ryzen 3600 setup - wat iam reading here sounds horrible (specially the cpu fan bug)! Maybe it was a mistake buying this board?

6kbyte wrote:
I just ordered a ch6 for a ryzen 3600 setup - wat iam reading here sounds horrible (specially the cpu fan bug)! Maybe it was a mistake buying this board?


I don't regret my purchase of this mainboard. The fan bug only happened 3 times in total so far (it has already been used for multiple 100s of hours since I build it), and every single time I had hwinfo64 running aswell. So this bug might only ever happen when hwinfo "calls" the mainboard for the updated statistics. And the cold boot bug is just annoying. And dosen't really break anything.

Lextra wrote:
I don't regret my purchase of this mainboard. The fan bug only happened 3 times in total so far (it has already been used for multiple 100s of hours since I build it), and every single time I had hwinfo64 running aswell. So this bug might only ever happen when hwinfo "calls" the mainboard for the updated statistics. And the cold boot bug is just annoying. And dosen't really break anything.


While on BIOS 7302 (the only one I've used when installing R7 3700x coming from 6301 with my R7 1700) I've seen the CPU fan ramp down to about 400rpm (set to 1200~1500 ish) and chassis fans to about 700~800 rpm (they're set to their max 1500 ish) as if they weren't getting enough power; and on one occasion the CPU fan just went to 0 rpm, stopped spinning completely. Several times the CPU temperature reading (in hwmonitor) has gotten stuck at a certain number too (random degree celcius). All the above requires a reboot of the system to return to usual operation. EVERY SINGLE TIME any of the above happened, I was having hwmonitor running for extended periods of time (hours, more than a day etc).

On my last reboot, I abstained from running hwmonitor or hwinfo anything longer than 5~10 minutes, only to check status then exit. No problems since. Leads me to be pretty convinced hwmonitor/hwinfo/etc utilities calling/polling the motherboard's sensors/controller for these rpm/temperature values are causing the problems with fan stopping/ramping down to terribly low level. I don't think this ever happened once on my R7 1700 on the C6H Bios 6301 so it may be the BIOS itself (or AGESA? if it's possible even).

Hoping Asus will look into this.

Preface: I am in no way arguing that anyone should have to do this, or that Asus does not need to get a handle on this fan bug, but...

Have any of you considered getting a Corsair Commander Pro and just foregoing the Asus fan headers entirely? It has 6 PWM fan headers on it. It also has the benefit of adding an additional USB 2.0 header (it has 2 on it, but it requires 1 from the mainboard, and c6h has only one) as well as up to 4x 10k ohm temp probes (and it comes with the probes as well as 6x PWM fan extension cables) and 2 lighting channels for Corsair Addressable RGB stuff if you are into that sort of thing.

I have one, and once the Commander Pro is set with a fan curve, you don't actually need to run the Corsair iCue or Link software anymore - it remembers your settings. It also works fine with HWInfo monitoring. All it needs is the aforementioned USB 2.0 header from the mainboard and a SATA power plug from your power supply. A lot of Corsair stuff is pretty expensive, but you get a lot for the ~$80 US they ask for the Commander Pro. I have never regretted the purchase.