I actually had a faulty Ryzen 1700x I think? 50 to 60c idle temp.

horrorwood · 7 Mar 2017 at 21:16

Hi,

Setup is MSI B350 Tomahawk and Corsair H110i.

Upon first boot it was sat at 55c in the bios, I thought it may be a buggy BIOS. But I booted into Windows (on my old install) and saw it was now at 60c in the AMD Ryzen utlity.

Remounted the H110i twice with new paste, no luck, identical temps. Putting it under load was impossible as it hit 70c+ pretty much instantly. It was also very loud.

At this point I really am scratching my head wondering how the H110i could be faulty, wondered if the pump wasn't working etc. I was thinking maybe I could have mounted it wrong, but it really is very very simple to install on the AM4 socket. I saw other people with the B350 Tomahawk reporting normal temps, so I ruled that out.

So I picked up a Ryzen 1700 with the stock cooler.

Firstly I fitted the stock cooler with the 1700x... 62c to 69c at idle! at this point I knew it had to be the CPU.

Fitted the 1700 with the H110i, 31c idle...

So yeah I am RMA'ing the CPU. I have built 100's if not 1000's of PC's and I've never had a faulty CPU. Has anyone else had a faulty Ryzen CPU?

Hope this helps if anyone else has the same issue with stupidly high temps!

CAT-THE-FIFTH · 7 Mar 2017 at 21:19

Might be the motherboard overvolting the CPU??

Techen · 7 Mar 2017 at 21:46

I know this might not be the case but have you checked

http://www.tomshardware.co.uk/amd-ryzen-am4-motherboard-heatsink-problem,news-54977.html

horrorwood · 7 Mar 2017 at 22:07

Techen said:
I know this might not be the case but have you checked

http://www.tomshardware.co.uk/amd-ryzen-am4-motherboard-heatsink-problem,news-54977.html

Yep... 100% not a heatsink problem, it was very tight to the CPU with both the H110i and the stock cooler. Also the problem disappeared when swapping from the 1700x to the 1700 but with the same cooler.

CAT-THE-FIFTH said:
Might be the motherboard overvolting the CPU??

CPU-z reported normal voltage. I have tried up to 1.45v on the 1700 and I am still nowhere near the temps the 1700x was reporting.

CAT-THE-FIFTH · 7 Mar 2017 at 22:08

horrorwood said:
Yep... 100% not a heatsink problem, it was very tight to the CPU with both the H110i and the stock cooler. Also the problem disappeared when swapping from the 1700x to the 1700 but with the same cooler.

CPU-z reported normal voltage. I have tried up to 1.45v on the 1700 and I am still nowhere near the temps the 1700x was reporting.

The R7 1700X is meant to be somewhat leakier according to some people,so probably is consuming more power at the same voltage as the R7 1700.

eddiew · 7 Mar 2017 at 22:13

CAT-THE-FIFTH said:
The R7 1700X is meant to be somewhat leakier according to some people,so probably is consuming more power at the same voltage as the R7 1700.

Surely not to the tune of 60 degrees at idle though?

OP does sound like he's got a weird one. I guess it's possible to get a badly made unit without enough/any solder twixt the chip and heatspreader or something... sooner or later, someone's gonna get the duffer.

CAT-THE-FIFTH · 7 Mar 2017 at 22:16

eddiew said:
Surely not to the tune of 60 degrees at idle though?

OP does sound like he's got a weird one. I guess it's possible to get a badly made unit without enough/any solder twixt the chip and heatspreader or something... sooner or later, someone's gonna get the duffer.

Yeah,you might have a point.

Yaayuh! · 7 Mar 2017 at 22:20

Very odd.
What does everyone else's BIOS have for the temp? Mine says 38c (1800X).

horrorwood · 7 Mar 2017 at 22:21

eddiew said:
Surely not to the tune of 60 degrees at idle though?

OP does sound like he's got a weird one. I guess it's possible to get a badly made unit without enough/any solder twixt the chip and heatspreader or something... sooner or later, someone's gonna get the duffer.

Sorry yeah I should have added that when I ran the built in CPU-Z benchmark it was hitting over 70c pretty much instantly on the single thread test, I stopped it immediately. That was with the H110i too.

Davedree · 7 Mar 2017 at 22:42

I had this with an old skt 939 4000+ and had to delid the bugger.
Willing to bet the Solder under the Ihs isn't making a great contact between die and Ihs.

horrorwood · 7 Mar 2017 at 22:42

Rossi~ said:
Very odd.
What does everyone else's BIOS have for the temp? Mine says 38c (1800X).

Snap with a 1700...

http://i.imgur.com/dMbdi6B.jpg

But with a 1700x...

http://i.imgur.com/yBrj65I.jpg

Drollic · 7 Mar 2017 at 23:22

Was the notched side on the cooler bracket correct?

horrorwood · 7 Mar 2017 at 23:39

Drollic said:
Was the notched side on the cooler bracket correct?

Yes everything regarding cooler installation was correct, the problem follows the CPU, it doesn't matter which cooler I use.

I can put the 1700x back in and run the CPU-Z bench if required

Rroff · 7 Mar 2017 at 23:50

Would be interesting to know if it was actual CPU temp or faulty calibration - without something like the fluid temperature in the loop or a laser temperature sensor to go on hard to tell though.

I've run into 1-2 Intel CPUs in my time with faulty sensors (mostly LGA775 quads).

Probably as above poor interface with the IHS however.

Limpnoodle · 8 Mar 2017 at 04:38

I think Rroff is right but i would still rma it

But it wouldn't harm it much if u did try to overclock it and ignore temps

Andrei23 · 8 Mar 2017 at 19:28

Before you send back for warranty, have you tried delidding the chip?

Minstadave · 8 Mar 2017 at 19:35

Andrei23 said:
Before you send back for warranty, have you tried delidding the chip?

Why would you try? It'll ruin your warranty and it's a soldered (probably badly in this case).

woppy101 · 8 Mar 2017 at 19:38

Andrei23 said:
Before you send back for warranty, have you tried delidding the chip?

Is that a joke?

woppy101 · 8 Mar 2017 at 20:29

My chip is quite toasty also anything over 1.2v and 3.7ghz and the prime load temp jumps into the mid 70s

koooowweeee · 8 Mar 2017 at 21:15

Rroff said:
Would be interesting to know if it was actual CPU temp or faulty calibration - without something like the fluid temperature in the loop or a laser temperature sensor to go on hard to tell though.

I've run into 1-2 Intel CPUs in my time with faulty sensors (mostly LGA775 quads).

Probably as above poor interface with the IHS however.

I've had a few lga775 chips with stuck sensors. Mainly on 1 core. But never reporting higher temps.