Issue with Asrock B450 Steel Legend and Ryzen R5 3600

After having had some major hardware issues with my old components, I finaly managed to build up my new system. First time using windows 10. And I’m having a few issues. The system seems to be unstable. Blue screens, or just getting stuck. (audio clipping, PS/2 keyboard not responding with the numlock key… Event log reports some unknown critical issue, which according to google is commonly due to a bad PSU)

I’m trying to fix this issue, because I don’t think it’s the PSU at all. It’s an 850 watt 80+ gold Zalman PSU, and it works just fine with my old components which had a bigger power draw.

But the main issue I’m wondering about, is the CPU temp as reported in HWmonitor, bios and everywhere else. It will for example jump to 40 degrees C, and then within a few seconds drop back to 30 degrees with 2 degrees intervals. To then just instantly jump back up to 40 degrees. I can not explain this at all, because I’m using an AIO liquid cooler. And even if I had it poorly mounted, it should not jump around with those temps like that. 40, 38, 36, 34, 32, 30, 40 again… Does anyone know what can cause this issue, or is it just a software issue… (it happens in hw monitor and bios/uefi)

Btw, I did try an cinebench test. Got a perfectly normal score there, and no thermal issues, other then it floating around. It would also hit 4200 mhz.

(if you guys/galls need more specs, I can provide them) I’ve also updated my bios to the newest version.

How high is your soc voltage?
When in bios you could try to set it manually to 1.1V.
That might help with stability issues.

However reading trough your posts, i kinda get the feeling that your,
board might make an unwanted contact with the case.
also double check you front usb ports of the case if your case has those,
and being connected to the mobo.
Maybe one of them is damaged, that can also lead to several strange freezes like that.
Or the internal kabel.

What i would do in this case, is to take everything out of your case,
lay the mobo on its cardboard box, install gpu drive etc, and fire it up outside the case.
To see if you are still facing randon freezes.
Just to rule out an earth issue with the case.

Another thing you might wanne double check are the cpu pins.
The cpu should drop in the socket fluently.

Also the psu isnt totally to be rules out here as well.
However if the psu is bad, then generally the system should just shut off.
But still i wouldn’t rule it out that fast.

Hi, and thank you for your reply. In the past, programs like HW monitor never showed me realistic values for power supply stuffs. The numbers they listed were always way off and not realistic at all. (mainly due to me usually using old hardware) So I don’t know muhc about voltage values, nor have I really looked into overclocking. Because of that, I just compiled a screenshot of some values shown in HW monitor.

I do have some propper casemods going on. And I had to weld together a bracket to mount my 240 radiator. It’s pushing onto the USB/fan/RGB headers on my motherboard. But because I was aware this would happen, I made sure there was nothing making any contacts. This isn’t the problem.

But yea. I think I solved this instability issue. My pc has been up and running without a crash for over 15 hours now. Before, it wouldn’t last 40 minutes. As part of the troubleshooting I had played around with drivers. I also auto OC-ed my 1060 3g video card and ran a bunch of heaven benchmarks. This did not cause an crash. I’ve also run a few back-to-back cinebench benchmarks for the CPU, which also didn’t cause a crash. And if there would be an power supply issue, I would have asumed it would crash during “heavy load”. (mind you, an 850 watt PSU for an Ryzen 3600 + 1060 3g is a massive overkill, so the PSU would still remain kinda at idle)

Because I was playing with a fresh windows 10 install, and because I never before played with windows 10 before, I only had my SSD connected. I also had one of my old drivers HDD’s in another pc. To put it back into this rig, I had to remove the memory (clearance issue) to mount it back in. Made sure the memory went back in propperly. At the same time, I also updated my bios/uefi for the B450 Steel Legend from version 2.30 to 2.70. The Asrock site also told me I should install some software to complement this bios/uefi update, which I did. Considering I couldn’t get my pc crash-free for 40 minutes before, I think the instability has been fixed. (writing this, so it might help someone else out in the future)

But my main confuzzlement for initially creating this post remains. The temps for package (node 0) keep fluctuating. Will go from 40, slowly down to 30, to instantly jump back up to 40 degrees C. That kind of behavior can’t really be explained while using an air cooler, and especially not with an AIO. Ambient temps in my room are about 23 degrees C, so I do think I’ve got the AIO propperly connected. (it keeps all the other values low and stable, it’s just the package (node 0) which keeps bouncing. What causes this behavior.

@MisteryAngel

Just ran Heaven 4.0 and Cinebench R20 again, to take propper screengrabs. I used an old driver for my GPU, because of the stability testing. It’s version 388.13, will upgrade after the PC has been stable for at least 2 days. (24/7 usage)

Scores look good, no?

Based on those figues.
I don´t really see anything strange.
The core voltage on full load might be a little bit high,
like 1.424V for a 3600.

I need to dig a little bit deeper into this,
and gather some more info.
However it could also be a bios bug.
The problem is that the Ryzen 3000 cpu´s on previous gen boards,
can still be a bit quirky from time to time.
I don´t really think that there is really much to worry about at this point,
in regards to the package temp jumps as long as it doesn´t go sky rocket.
But atleast keep an eye on it, as long as the system runs stable.

Large core temp fluctuations at light load is normal for ryzen 3000 chips.

The temp reading is the hottest out of lots of sensors all over the die, when the cpu is given any work to do it boosts aggressively, spiking the heat output in one tiny part of the chip briefly.

Because the cores are so much smaller and boost so much harder than intel ones do the temp spike is larger regardless of the cooler used, the bottlneck becomes getting heat from the tiny part of the die into the IHS.

Installing the newer chipset drivers changes the windows power plan to make it less likely to boost when it isn’t needed, but running lots of things in the background on your pc raises the noise floor enough to negate that a bit.

The way I solved the cooler fan speed spiking was to have my fan setup so below 50c is 30% speed, 50C to 70C is 48% and between 70C and 90C it ramps to full speed.

1 Like