ROG Crosshair VIII Formula + R7 5800X3D - Severe Throttling (573MHz) and "Detect HDD" Freezes (1650% Power Reporting Deviation)
Posted by Elmormon@reddit | buildapc | View on Reddit | 15 comments
Greetings everyone, before getting new parts or doing something stupid, I want to ask first.
TL;DR: My Ryzen 5800X3D is getting locked at 573 MHz due to a massive 1,650% Power Reporting Deviation on my Crosshair VIII Formula. The system eventually freezes, slows down or crashes with a "Detect HDD" message on the LiveDash OLED and Q-Code AA, even with no HDDs connected.
Full Specs:
- CPU: AMD Ryzen 7 5800X3D
- Mobo: ASUS ROG Crosshair VIII Formula (BIOS 5302)
- GPU: ASUS ROG RTX 4080 (using 3x independent 8-pin cables with the adapter)
- PSU: ASUS ROG Thor 1200W (V1, non-ATX 3.0)
- RAM: G.Skill TridentZ Royal 3600MHz 64 GB (32x2) XMP Enabled
- Storage: 2x Samsung 980 Pro 1TB (Latest Firmware), 1x Kingston A400 SSD, 2x WD Blue HDD (disconnected for testing)
- Cooling: ROG Ryujin II 360
The Problem: I am experiencing two types of crashes:
Extreme Slowdown: The CPU clocks drop and lock at 573.4 MHz. HWiNFO64 shows Power Reporting Deviation at \~1,650% and TDC/EDC limits at 100%+ despite temps being around 40°C.
- System becomes extremely sluggish (typing delayed, windows lag, videos freeze), but mouse movement remains responsive.
- Restarting from Windows does not fix it, only a full cold shutdown does.
- Hard Freeze: Total system lockup. The motherboard OLED displays Q-Code "Code AA: Detect HDD" or "Code 30: Check CPU" upon crashing/rebooting, which is strange as I’ve tested the system with only M.2 drives connected, that means no HDD, right? Power button from the case doesn't work at all so I have to shut it down from the PSU switch.
Troubleshooting Steps Taken:
- Fresh Windows Install: Reinstalled everything from scratch, updated all drivers. Symptoms persist.
- Replaced Cables: Replaced the cables for the CPU connection on the motherboard, thinking it could be a faulty "sleeved" with the ones that came with the PSU.
- BIOS Settings:
- Disabled Global C-States and PBO.
- Set Power Supply Idle Control to "Typical Current Idle".
- Forced PCIe Gen 3 (instead of Gen 4/Auto).
- Performed Clear CMOS multiple times (removing the battery and using the dedicated button on the motherboard).
- Hardware Checks:
- Slow Mode Switch: Verified it is physically OFF.
- fTPM: Updated/Toggled for testing.
- Power: Verified all EPS 8-pin and GPU cables are seated correctly.
- Isolation: Disconnected all SATA drives; problem still occurs with only NVMe.
- RAM: Tested with OCCT and y-cruncher, no issues.
Current Diagnosis / Theory
This strongly looks like a hardware-level telemetry failure on the motherboard (VRM controller / power reporting / chipset I/O).
The board appears to be feeding false power data to the CPU, triggering a permanent PROCHOT-style emergency throttling state, followed by chipset instability (I/O collapse -> “Detect HDD”).
The issue is intermittent:
- Some days the PC works perfectly.
- Other days it happens multiple times (e.g. twice within 20 minutes).
- Happens under load and idle (the system often freezes while idle for remote access.)
At this point I’m considering replacing the motherboard, 'cause I can't take it to someone to fix it since it's hard to replicate, it could work for him for days and not crash but anyways, I’d really appreciate a sanity check before doing so since changing it means getting a potential DDR5 one and with the price of the rams... Any insight is welcome.
MrTimmyTam@reddit
I have been having this same issue for the last few months. I have a very similar rig to yours. I have..
AMD Ryzen 7 5800X3D CPU
Asus ROG Crosshair VIII Hero (Wifi6) Motherboard
Asus ROG Strix GeForce RTX 4080 Super OC
Asus ROG Thor 850W Platinum PSU
Asus ROG Ryujin 240mm Liquid CPU cooler with colour OLED
G.Skill Trident Z Neo F4-3600C18D-32GTZN 32GB (2x16GB) DDR4
Samsung NVMe M.2 980 Pro 1TB
Samsung NVMe M.2 970 EVO Plus 2TB
Bios is 5302, amd chipset is the latest 7.11.26.2142 I’m at my wits end trying to figure out / fix it. Mainly happens when I game. I’ve tried everything you’ve done, CMOS multiple times. I’ve even tried undervolting… temps are great but out of nowhere it just decides it’s had enough.
https://imgur.com/gallery/BWj0kp5
This is a screenshot I took about an hour ago. Check that power reporting Deviation!!! I had xmp applied, PPT=105, TDC=70, EDC=130, all core -30 curve, c-states disabled, CPPC enabled, preferred cores disabled. Ran a 1 hour stress test in OCCT, everything was fine, bumped it up to extreme 30 mins in this is what happened. I’ve tested after cmos fully stock, same thing. For the last month or so it just happens randomly when gaming. Could be CS2, battlefield 6, definitely don’t like ARC raiders. I can just be sitting in a menu doing nothing. But yeah…. I’m almost ready to just move to AM5.
Straight_Loan8271@reddit
same problem on a crosshair 8 dark hero with 5800x3d. sometimes the soc current goes to 90a without the core current reading going crazy but yeah
MrTimmyTam@reddit
Just have to assume it’s Asus boards. Gonna go find me a MSI MPG X570S Carbon Max and see how that goes.
Straight_Loan8271@reddit
i think i have actually figured out the problem at least in my case. if you have multiple pieces of software trying to access the embedded controller at once, the EC can shit itself and start reporting bad data until the next power cycle. i had hwinfo, aida64, fancontrol and aquasuite running most of the time so i went through all of those and deselected asus controllers as a sensor source. been working fine for the 2 weeks since then
i've only seen this reported as an issue on asus x570 boards so switching vendors would probably fix it too. apparently it doesn't happen in linux either because it reads sensors differently, but if you play games with kernel level anticheat that'll be a problem there
MrTimmyTam@reddit
I’ll have to give that a go when I get the chance. Never had a problem until I played battlefield 6, that’s when it all started. Sometimes I can play for a week and it’s fine, then the next week it shits itself every few games. I usually have one of either Ryzen master, MSI afterburner or hwinfo64 running each session, even done it when I’ve had none of them running. But I’ll try flicking off the Asus sensors and see how that goes for a few weeks.
xthelord2@reddit
power reporting deviation means CPU thinks its getting 16.50x more times of EDC, TDC and PPT than what AMD recommends
in ASUS settings there is a specific setting called core current telemetry which is what HWINFO64 taps into to get that power reporting deviation value, first ensure it is at default value
+ means positive offset, meaning your CPU will think it is seeing higher values
- means negative offset, meaning your CPU will think it is seeing lower values
set it to the negative offset and in field below you type in number, start with 2500 and do increments of 1000 till you get it below 90% than roll back 100 at a time till it resides in 90-110% range
apply new setting, restart and test your changes by running cinebench R15 all core render pass and during that time check power reporting deviation %, it should stay within 90%-110% range
if it doesn't change you got issues with motherboard sensors giving wrong data to the CPU
Elmormon@reddit (OP)
Thanks for the answer, I applied the -2500 value and the during the test it went up to 96.5% according to hwinfo, right now at idle (with just web browser open) it moves around 40 to 70%
xthelord2@reddit
how does it look when you do a full load test?
do clocks match the advertised speeds?
do TDC, EDC and PPT match advertised values?
Elmormon@reddit (OP)
Alright, this was after a second test since I didn't screenshot on the first one. https://i.imgur.com/JX62Juc.png
As for the curve, I'd rather not mess around with stuff that could cause more inestability for now, I just want to make it work properly lol
xthelord2@reddit
effective clocks look in line with whats expected
deviation is in line
this CPU works just fine, time to play with curve optimizer if you want to
TheKitler@reddit
Check the AIO just to rule it out. Your CPU is at 40 degrees but it could only be that low because the clock speed is so low.
Make sure your pump is plugged in, make sure you can see the RPM readout in bios, and make sure there's fluid running through the tubes (touch them - one will be warm and the other cold).
Elmormon@reddit (OP)
Hi, yes, I checked it the usb connector for the pump, it's connected and working properly, Bios and Armoury Crate shows RPM for Pump too.
TheKitler@reddit
Do you feel fluid moving through the tubes? If there's a clog, the pump will appear to be working but no/limited fluid will be moving.
Mounting pressure would also be good to check.
Elmormon@reddit (OP)
Thanks again for the answer, yes, indeed I can feel the fluid moving through, as for the mounting pressure, I just used my fingers even though I could have used a screwdriver for the thumbscrews
TheKitler@reddit
I think swapping the motherboard for something cheap just to test may be the right move then. It still could be the CPU but boards fail more often.
Because the issue is intermittent, have you tried stress testing with OCCT or similar to see if you can easily replicate the issue?