F5F Stay Refreshed Hardware Desktop ASROCK A300 MINI PC has stopped working due to overheating.

ASROCK A300 MINI PC has stopped working due to overheating.

ASROCK A300 MINI PC has stopped working due to overheating.

H
Hanshb3
Member
132
04-19-2016, 07:28 AM
#1
my mom's asrock a300 pc has started stopping unexpectedly recently. THE HARDWARE MODE IS RYZEN 3200G WITH STANDARD SETTINGS, COOLER THAT FITS, 16GB RAM at 3000MHZ DDR4, and a 1TB M.2 drive. I can recreate the crashes using a power test on occt, where the 'CPUTIN' jumps to 95°C while the real CPU core stays cooler. If I run a full CPU workload on occt, temperatures stay normal. When I load a 3D game on occt, the 'CPUTIN' also rises quickly. I dusted the PC and replaced the thermal paste recently, but it didn’t help. I’m planning to reapply the paste to check if the issue was missed. I think the 'CPUTIN' might be coming from somewhere on the motherboard, since it’s significantly higher than other chip readings, but I don’t have a way to confirm yet.
H
Hanshb3
04-19-2016, 07:28 AM #1

my mom's asrock a300 pc has started stopping unexpectedly recently. THE HARDWARE MODE IS RYZEN 3200G WITH STANDARD SETTINGS, COOLER THAT FITS, 16GB RAM at 3000MHZ DDR4, and a 1TB M.2 drive. I can recreate the crashes using a power test on occt, where the 'CPUTIN' jumps to 95°C while the real CPU core stays cooler. If I run a full CPU workload on occt, temperatures stay normal. When I load a 3D game on occt, the 'CPUTIN' also rises quickly. I dusted the PC and replaced the thermal paste recently, but it didn’t help. I’m planning to reapply the paste to check if the issue was missed. I think the 'CPUTIN' might be coming from somewhere on the motherboard, since it’s significantly higher than other chip readings, but I don’t have a way to confirm yet.

D
dawgal
Member
126
04-25-2016, 12:46 AM
#2
Refers to adjustments made for a custom fit. Did you modify the mounting system yourself?
D
dawgal
04-25-2016, 12:46 AM #2

Refers to adjustments made for a custom fit. Did you modify the mounting system yourself?

G
Gravewalker21
Member
64
05-06-2016, 08:48 AM
#3
Remove the ring around the top of the fan indicating AMD or adjust for the stock cooler that’s too tall. It only uses the standard mount with four screws.
G
Gravewalker21
05-06-2016, 08:48 AM #3

Remove the ring around the top of the fan indicating AMD or adjust for the stock cooler that’s too tall. It only uses the standard mount with four screws.

K
KeeezZ
Junior Member
10
05-06-2016, 09:29 AM
#4
It seems the mod should work as expected. I might need to verify if the backplate came loose unintentionally.
K
KeeezZ
05-06-2016, 09:29 AM #4

It seems the mod should work as expected. I might need to verify if the backplate came loose unintentionally.

D
Disco_Master
Member
163
05-06-2016, 11:20 PM
#5
That's a good observation. I did see the backplate shifting slightly when the cooler wasn't running.
D
Disco_Master
05-06-2016, 11:20 PM #5

That's a good observation. I did see the backplate shifting slightly when the cooler wasn't running.

C
Commando__
Senior Member
744
05-07-2016, 04:58 AM
#6
previous mount and paste spread performed well; I realized my version of OCCT was quite old and updated it. currently, the system doesn't display the "cputin" sensor, but CPU and GPU temperatures on die/package remain within normal limits of 70-80°C during load. however, when using an older OCCT version, the CPUTIN sensor stays between 94°C to 98°C with either CPU+GPU or GPU-only states. this makes me suspect it's a VRM temperature sensor. I disabled the overtemp protection to monitor the sensors more closely. setting my CPU package voltage manually to 1.275V instead of auto (which reached up to 1.44V) appears to limit the CPUTIN reading to 91-92°C under load. adjusting the voltage doesn't significantly affect CPU/GPU die temps, only the CPUTIN value. it seems stable, so I'll proceed with this approach for now.
C
Commando__
05-07-2016, 04:58 AM #6

previous mount and paste spread performed well; I realized my version of OCCT was quite old and updated it. currently, the system doesn't display the "cputin" sensor, but CPU and GPU temperatures on die/package remain within normal limits of 70-80°C during load. however, when using an older OCCT version, the CPUTIN sensor stays between 94°C to 98°C with either CPU+GPU or GPU-only states. this makes me suspect it's a VRM temperature sensor. I disabled the overtemp protection to monitor the sensors more closely. setting my CPU package voltage manually to 1.275V instead of auto (which reached up to 1.44V) appears to limit the CPUTIN reading to 91-92°C under load. adjusting the voltage doesn't significantly affect CPU/GPU die temps, only the CPUTIN value. it seems stable, so I'll proceed with this approach for now.