• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Bad thermal sensor or weird temp issues?

cdawall

where the hell are my stars
Joined
Jul 23, 2006
Messages
27,683 (4.02/day)
Location
Houston
System Name Moving into the mobile space
Processor 7940HS
Motherboard HP trash
Cooling HP trash
Memory 2x8GB
Video Card(s) 4070 mobile
Storage 512GB+2TB NVME
Display(s) some 165hz thing that isn't as nice as it sounded
This is a fresh redux for the in-laws. Reusing their old case, an H50 with a single yate loon highspeed and an A6-7400K. The H50 was keeping an Athlon II X3@3.5ghz under 55C load. Same case, same fan setup, onboard HD4200 with a decent overclock.

Full specs
A6-7400K@4.1-4.6 turbo .012v over stock
H50+Yate loon high speed in pull
Cheap mid tower
300w SFX FSP with an ATX adapter plate

capture.png


Temp 2 should be motherboard socket temp, CPU temp reads well into the 80C range under load. Temps are really jumpy (almost an instant 40-50C jump to 80C then instant drop back). CPU is gaming stable at these temps however.

What does everyone think bad CPU, bad temp sensor or what?
 
could be that the software is just reading wrong. Is the platform is stabile and works, I wouldnt worry.
 
could be that the software is just reading wrong. Is the platform is stabile and works, I wouldnt worry.

I have been playing Carmageddon on it for the past couple of hours off and on so I would assume that means it is stable.

I threw an 80mm in the side panel for shits and giggles, directly above the CPU so the cooler and chip should have some form of fresh cool air, and there was no temperature difference whatsoever. Rear exhaust doesn't even feel warm I am going with botched reading.
 
I find it doesn't read APU's too well as they are a combined unit.
I'd be more concerned about the low voltage readings for the 12V.
Of course the readings can all be thrown out once overclocked, as this affects graphics as well as CPU. Stock clock of 3.5Ghz with turbo of 3.9Ghz indicates overclocking has been done.
 
Last edited:
AMD APUs are all over the place with software reading. Use the AMD Overdrive software that came with it. Its works

On a side note I loved Carmageddon
 
I find it doesn't read APU's too well as they are a combined unit.
I'd be more concerned about the low voltage readings for the 12V.
Of course the readings can all be thrown out once overclocked, as this affects graphics as well as CPU. Stock clock of 3.5Ghz with turbo of 3.9Ghz indicates overclocking has been done.

12v rail is fine. Software can't read it right.

And yes it's over clocked pretty decently I might add. Now that I know temps are fine sounds like it's getting more voltage.
 
I'd be more concerned about the low voltage readings for the 12V.

That's an aberration, I think. HWMonitor has been unable to read 12v temps on two different PC's (well, a PC and a server) for me with different PSU's. It always shows the 12v at 8v.
 
You may thank AMD for doing temperature acquisition back-ass-wards.

I haven't played with their newer Athlons and FX CPUs, but I know that on APUs instead of passing the absolute value of the temperature they are instead calculating a delta to some critical temperature (which is also unknown).
My A6-5400 and A4-5300 all show different temps in different software. Under the same conditions I might even see some ridiculous temps in BIOS: may be between 17°C and 43°C at cold boot! AMD Overdrive was promised to work, and while it does display the delta and behaves logically (more load - lower delta - hgher temp), it is still almost as random as HWMonitor and other software.

I haven't been able to find any software that can effectively monitor temps on APUs, so your options are limited to physical thermal probes, or scientific finger-poking of the heatsink.

At least my voltages were in a "reasonable" range(+/- 0.5V error margin).
 
It looks like I can go off of socket temp and be safe enough. 40C to tDelta and socket temp shows around 40C so that works for me. Needs a decent vcore bump to get over this clockspeed, not sure if it is worth it to me even if temps are "good". I might work on getting the ram clocked up higher instead. 1866 is all this A6 officially supports, I have had this ram up to 2400 on my FX 9370 so I should be able to do the same on here, but can't get it stable.

capture302.png
 
I know the H100 is a good AIO

640 l3 enabled.PNG


Truth be told....unlocking the L3 cache and overclocking it stops this sensor working.

Temps are fine btw, i checked before i unlocked the chip.
 
I know the H100 is a good AIO

View attachment 71426

Truth be told....unlocking the L3 cache and overclocking it stops this sensor working.

Temps are fine btw, i checked before i unlocked the chip.

Yea the Athlon II I had in it before this was the same way. Dead sensors after unlocking. I haven't done anything with the A6 other than overclock it.
 
Yea the Athlon II I had in it before this was the same way. Dead sensors after unlocking. I haven't done anything with the A6 other than overclock it.
That's the way they are. I had a batch of 5 A4-5300B and an extra pair of A6-5400K: all had the same problem. Tested on 3 different boards w/o OC. Had to dig up my thermocouple just to get an approximation.
 
That's the way they are. I had a batch of 5 A4-5300B and an extra pair of A6-5400K: all had the same problem. Tested on 3 different boards w/o OC. Had to dig up my thermocouple just to get an approximation.

Eh I don't care that much I am going to go back with it didn't crash so its fine.
 
That's all that matters :)
 
That's all that matters :)

I really want to get the GPU up higher. 800-1000mhz made a decent jump in games. I guess I will have to push more vcore :shadedshu:
 
Back
Top