Temps too High. Scythe Mugen 5.b + 5600x. What's your max PPT for up to 80c load?

njsx2k

n00b
Joined
Sep 17, 2010
Messages
16
I've had to set cpu PPT to 93W so that load temps max out at 80c with Prime95 v30.8b7 (small FFTs).

Motherboard Temps (Asus B550-F): 23c Idle, 26c Load.
CPU idle temps: 34-40c.
CPU fan speed 100% above 67c.

That's 54-57c above case temps on load. I feel like my temperatures are too high for this cooler and that it should handle a higher PPT. This is a week after re-applying arctic silver 5 (4 years old) using the spread method.
It's a single ccd 5600x, so that means smaller area for heat dissipation - I wonder if this is the problem.

Do I need to get a better thermal paste, or is this simply the limits of the cooler not having a copper plate? When installing the mugen 5 cooler I noticed the screws hit its limits without applying much pressure, but still enough for it to feel rigid. The exhust air at the back of the case doesn't feel that warm. I've had this cooler installed on a lga 1366 system before (200w+ cpu) with much higher pressure.

My adjusted manual PBO limits are: PPT=93, TDC=61, EDC=90. Throttle limit=90c. CO on with -15 to -30 per core. No boost override.
 
Last edited:

freeagentt

Limp Gawd
Joined
Dec 5, 2018
Messages
289
It seems a bit warm..

5600XOCCT2000C14.PNG
 

njsx2k

n00b
Joined
Sep 17, 2010
Messages
16
You have so much more headroom with ppt up to 120w, and your case temps are only ~6c cooler. Is that a single ccd?

I think I need to reseat the cooler. I've also ordered some arctic mx-5 thermal paste.

I've adjusted limits to:
ppt 96
tdc 56
edc 83
and +ve 3 to 4 points on some cores for stability.
Bios updated to 2423.

This combo seems to provide the highest all core clocks (4550Mhz avg) and lower voltage while running cinebench multithread at below 78c, and hits tdc limit first for prime95 so max temps are below 75c.
 
Last edited:

freeagentt

Limp Gawd
Joined
Dec 5, 2018
Messages
289
You have so much more headroom with ppt up to 120w, and your case temps are only ~6c cooler. Is that a single ccd?

I think I need to reseat the cooler. I've also ordered some arctic mx-5 thermal paste.

I've adjusted limits to:
ppt 96
tdc 56
edc 83
and +ve 3 to 4 points on some cores for stability.
Bios updated to 2423.

This combo seems to provide the highest all core clocks (4550Mhz avg) and lower voltage while running cinebench multithread at below 78c, and hits tdc limit first for prime95 so max temps are below 75c.
Lol yeah, I have seen the CPU do 155w PPT with a manual clock, she is a beast :D

Yes, single CCD, this chip is from 2020 I think.. from just after the initial Vermeer shortage.

I was running pretty close to my 5900X settings, so roughly 200PPT 140 TDC, 180 EDC.

Why? Because I lost my piece of paper that had my dialed in 5600X settings lol..
 

njsx2k

n00b
Joined
Sep 17, 2010
Messages
16
Lol yeah, I have seen the CPU do 155w PPT with a manual clock, she is a beast :D

Nice.

It was the as5 paste as I had suspected. The CPU was pulled off the socket on both occasions with this paste. It did look a bit too dry - thankfully no bent pins. I don't understand why there's no latch on AM4 to keep it secure like the one on my previous x58 board.

After reapplying with the new arctic mx-5 paste - this time with the X pattern - my cinebench temps are now below 74c at when hitting 100w ppt, and prime95 with small FFTs is about the same now (hits 60 tdc first, 93w ppt in hwinfo). This is with thermal limit set to 95c.

It's sitting at 95, 60, and 90 ppt/tdc/edc for now with per core -ve curve. Peak voltage registers ~1.38v, and ~1.3v on load. I can push for more if I feel I need the extra performance in the future now that I know I have more thermal headroom.
 

njsx2k

n00b
Joined
Sep 17, 2010
Messages
16
just like all amd chips with pin, you warm it up and then twist the heat sink off. it in their instructions...

That's what I did. Rather instinctivley as I didn't want a repeat of the first incident. Was carefull to twist it first but it still came off too easily.
I think the paste was just bad - I've used as5 for years and this time it just seemed a bit dry. It was a relatively new one, bought in 2018.
 

primetime

Supreme [H]ardness
Joined
Aug 17, 2005
Messages
7,301
serious question.....does the cooler ever actually feel hot when stress testing? i remember back in the 1366 days my thermalright ultra would get scolding hot. these new coolers barley get warm for some reason. Are they some how dissipating the heat that fast or what?
 

njsx2k

n00b
Joined
Sep 17, 2010
Messages
16
serious question.....does the cooler ever actually feel hot when stress testing? i remember back in the 1366 days my thermalright ultra would get scolding hot. these new coolers barley get warm for some reason. Are they some how dissipating the heat that fast or what?
It barely feels warm and the same goes for the case exhaust air. It felt a lot hotter with my previous x58 chip. I think it may be due to the smaller die size especially with the single ccd on this.
 

primetime

Supreme [H]ardness
Joined
Aug 17, 2005
Messages
7,301
It barely feels warm and the same goes for the case exhaust air. It felt a lot hotter with my previous x58 chip. I think it may be due to the smaller die size especially with the single ccd on this.
yea agreed....these newer chips are a bitch to cool...in hindsight i wish i had kept my thermalright ultra to at least test with:) it was very clear when it made good contact in those days...anyways my next chip SHOULD be easier to cool.(either a 5900 or 58003D) ought to have more contact area.
 

freeagentt

Limp Gawd
Joined
Dec 5, 2018
Messages
289
Nice.

It was the as5 paste as I had suspected. The CPU was pulled off the socket on both occasions with this paste. It did look a bit too dry - thankfully no bent pins. I don't understand why there's no latch on AM4 to keep it secure like the one on my previous x58 board.

After reapplying with the new arctic mx-5 paste - this time with the X pattern - my cinebench temps are now below 74c at when hitting 100w ppt, and prime95 with small FFTs is about the same now (hits 60 tdc first, 93w ppt in hwinfo). This is with thermal limit set to 95c.

It's sitting at 95, 60, and 90 ppt/tdc/edc for now with per core -ve curve. Peak voltage registers ~1.38v, and ~1.3v on load. I can push for more if I feel I need the extra performance in the future now that I know I have more thermal headroom.
This could help you..

https://www.amazon.ca/Thermalright-mounting-Bracket-Retention-Include/dp/B08S76Q1SK

I push down a bit and give mine a titty twister before I pull it off.
 
Top