i noticed they have brought out a new heat sink for the new ultra series processor with the same footprint as the other processors, but it has a single what looks to be a bit beefier heat pipe and one that covers the chokes as well. I currently own a ryzen 5 7460 that overheats sometimes and was wondering if i upgraded the cpu to the ultra cpu would it make a difference. Just for confirmation the ryzen 5 is 28w tdp an i7 has a tdp of 55w and the ultra with a stunning tdp of 115watt it must be the best cooling heat sink. unless they manage the power in a better way, i guess. any thoughts???
My guess is that there are tolerance differences between them. With the mission of Framework, I don’t think they would make a different heatsink part/sku for every processor series unless there was a reason. My guess is that differences in the CPU die sizes and heights required different dimensions on the heatsink.
In other words, I suspect there’s a good chance the Intel Core ultra heatsink wouldn’t work for the Ryzen 7040 series. But that’s just speculation.
As for the overheating, have you checked the thermal paste application, or maybe even re-done it?
This is likely false advertising
thats max but from their website!
i guess not but you never know!!
im going to repaste but its a single core that gets nearly double the temp so im guessing its a tad off. ill repaste though.
I guess it’s probably used in a gaming laptop.
The processor, itself, by Intel’s design, is designed to function up to that TDP. *
**Need alignment from board / system / laptop manufacturer.
This is were you will be capped by Framework’s design.
Here:
What do you mean by this? Can you articulate it in terms of symptoms, impact and behaviour?
Just moving to ptm7950 is probably the least cost/risk/effort for reward on the heat angle there and should probably be done before trying something as drastic as a probably not that much better cooler that doesn’t officially fit.
basically it started having hardware crashes, but now its no longer!!! just email to resolve the situation. the last temp that one core got to was 104.7degc the other cores were 50-60degc.
that certainly looks like a defect, may just be bad paste application
i was using grizzly kryonaut but it still over heated plus the thermal paste doesnt make an absolute huge difference as long as you dont put too much and dont smooth it out like a monkey ><.
for direct die you absolutely need to make sure you have full coverage so you do need to smooth it out though preferably more like delicately than your average monkey XD.
Also ptm is not comparable to paste, it’s closer to slightly worse liquid metal without the big drawbacks of lm, at least in the power envelope of the framework.
you definitely DONT smooth it out the pressure of the heatsink even with large dies pushes it out to make an even coating!!! someone hasn’t dont their research there is like a billion and one videos of it on youtube!!!
The heatsink pressure definitely smooths stuff out, especially if you evenly tighten it but on direct die you can’t take chances. if you just put a bead in the middle and slightly unevenly tighten it is quite possible you just squeeze out everything out to one side and end up with voids on the other side, one core being at 105 while the rest are at 50 is very likely a symptom of something like that. Even worse is if you have voids in non core areas you don’t have visible sensors for, then you just get random crashes or instability for no apparent reason if not outright damaging the chip (modern chips tend to be a lot better in regard to shutting down instead of taking damage though).
With paste you should spread it at least somewhat on direct die, with liquid metal you HAVE TO and you should also do the heatsink side.
For stuff with an IHS it really doesn’t matter as long as there is enough spreading there is mostly just unnecessary effort and doesn’t make a difference either way.
well i came to the conclusion that spreading it causes voids because it traps the air in, where as a blob in the middle squeezes the air out. i think i read that nasa do it that way but i could be wrong.
Nasa also have tools and processes to ensure extremely even pressure and the capabilities to test. A bad spread can certainly introduce voids but that’s why you add a lot more paste than absolutely necessary, or just use the superior anyway ptm that comes as a pad XD.
Seriously though, ptm is probably the cheapest extra cooling you can get on this thing. I am still very curious how much better the intel heatsink is but I doubt it’s going to be worth the cost and effort to put it on the amd chips even if possible.
I used to use Kryonaut all the time but dumped it from my laptops after trying PTM.
PTM was far far better.
1 Like
This stuff really sounds over-hyped until you try it XD
I also did some ptm vs stock paste testing on the amd framework 13.
Unfortunately I only figured out how to bypass stapm throttling after I moved to lm so I can’t test how close it gets to lm above 30W but under 30W it was pretty much eqal.