P775TM1-G and I9 9900K high temperature problem!

That's what I thought, I wondered since receiving my new build from Eurocom en of July 2019. .At first, they were surprised it wasn't there, then that they'll send me one (through contact in Europe as I was already on a long stay in Europe), and, in the end, I was told they could not send it anymore but I could purchase it from then. It was not a huge deal as was told it'd be fine without, I was surprised still, poor customer care.
Thanks for clarifying this up. (-:

 

Thanks for this recap and clarification on the heat-transfer and your approach. Glad to hear your mod has been great for the last three months, since March. DId I understand correctly, your approach does not require a direct die frame then? (I was ordering the De8auer frame, which is getting out of stock in several sites here.)
If so, what is the advantage to use one like @Ole! did in his mode? (I thought it helps with mechanical stress around the CPU (die) while also helping to mate the heatsink onto die interface.)

Good news, I'll concentrate on sanding slightly and polishing the baseplate then.

This is one of the first videos I watched when I first got my system and heard about lapping the i9-9900K, and I was just starting to watch it again when you posted your reply - yes a good video.
I was going to be conservative on the CPU Die lapping. My main goal is not extreme OC but keeping the CPU for heavy load jobs on regular and warm room-temperature days. It is good to have some target values too before starting to sand and lapping away figure-of-eights on wet sand-paper. (It reminds me of prepping samples before microscopy experiments.

While I am waiting for the parts to come in, I'll measure baselines temps. Regarding your "Test Methodology" that you mentioned in this post
http://forum.notebookreview.com/threads/clevo-overclockers-lounge.788975/page-1886#post-10999525 :
FS is Fire Strike stress and benchmark test, but what is ESO, which is used to "kill all background services before starting FS loop"?

Thank you again for taking the time to answer my questions and to explain your elegant mod approach. Very much looking forward to trying it on my system, which I'll document and publish here fo P775TM1G users.

 

It's in the build box sent by Clevo and comes with the machine always.

 

Yup., I've been running the system with a lot of thermal cycles to see the effects of LM on the vapor chamber, basically speeding up the aging. @ole!!! s mod is awesome, in that it's not janky as mine, as you have a frame holding the CPU in place and the vapor chamber is soldered directly to the heatsink. So even less chance of LM leakage. Only downside was the limited thermal transfer, needing to mod the frame and you can't lap the die too thin. But pros are the added safety and longevity, where as mine is more focused on maximum temp reduction.

The purpose of the 10min FireStrike physics loop was to allow the heatsink get saturated enough to get into a steady state. ESO is a small utility that kills any background services, so that you do not get random CPU usage spikes while running the loop. It also went with why I chose the 5min cooldown after reset, so that Windows would have time to start all the process/services that I could then toggle off with the extreme preset.
https://www.sordum.org/8637/easy-service-optimizer-v1-2/

Be sure to reach out to @ole!!! for his perspective and advice if you plan on going direct die frame.

 

Thanks for sharing your thoughts on the two mods and clarifying about the test methodology. ESO looks like a handy small program and takes care of background services more easily than manually and in a consistent manner for testing, maybe even routine use - great.

I will reach out to @ole!!! about his system and the mod he did on the direct die frame in particular.

A thought about chances of LM Leakage at the VC/Heating interface (with base plate), how about a slim band of conforming silicon varnish (coats thickness close to LM layer) or a tiny bead of hight temp silicon or narrow foam barrier as some kind of slim gasket all around the edge of the VC, wouldn't that take care of it in this case (i.e. with base plate)?

jc_denton, I forgot to ask you about the merit of a second CPU (slim) fan as you and Ole discussed?

Now, reaching out to Ole!!!.

Thanks again for the clear and insightful comments.

 

Check out the foam barrier guide by @Falkentyne it's highly compressible and should form a seal around the socket. I would still have kapton tape around the socket itself to prevent anything even having the slightest off chance getting inside.
http://forum.notebookreview.com/thr...r-liquid-metal-safety-insurance-guide.817207/

Think I mentioned it in a different thread, the 2nd 5v fan nets you around 3-5c temp reduction but at a constant 28dB noise, unless you set up a potentiometer to control it. But you'll need to pull 5v off a SATA or USB port, since P7 does not have the extra dedicated 5v fan connector. Check out brother @matyee thread here
http://forum.notebookreview.com/threads/clevo-p870km1-g-upgrade-›-144hz-gsync-display-tm1-cpu-heatsink-4th-fan-2080rtx-vaporchamber.831287/

The fan, perhaps you can find it locally or online somewhere near you.
https://www.elfadistrelec.dk/da/rad...m-5v-69m-sepa-hy60q05ap/p/30037298?track=true

But keep in mind the 2nd fan is mainly for the P870, as it has a 2nd VRM radiator and room for it.

 

And venting for it too.

 

Even the stock vents are quite limited on the P7 IMO, but from what I recall there really isn't room for a 5mm fan, height wise.

 

No extra room on the P7 that I am aware of.

 

it really doesnt look it. the 870tm as well until we kinda forced it a bit. problem would be shared heatsink thats a pain too. 3mm higher on CPU side means tilted and no contact on GPU side thats a no no

 

I order Kafton tape. Thanks for the link I did come across it earlier.

It does look cramped in the P775 for a second fan. I'll look into these when I dig deeper and as a later update, at the moment I popped the back cover and blow the cooling pad directly, it does help.

 

Exactly @ole!!!, this is my concern with P775 unified heatsink, how it will accommodate the 3mmm Vacuum chamber directly between the die and the heats.
Once my 9900k is delidded, I'll know the height differential between CPU with and without IHS and STIM - hopefully close to 3mm.
If it is thicker I wonder if I'll be able to compensate the height on the GPU side.

 

mine is also retail chip, but removed the solder stim so did help lower overall height of chip, even then its nothing near the 3mm you'll be adding onto your heasink.

 

Best mod would be to improve the airflow for the bottom intake, you can test with it and without and see how restrictive it really is.

IHS measurements, stock 9th gen is ~2.3mm tall, this is with without taking in account the STIM. So the 3mm VC comes pretty darn close to it, but as you know even a tiny z-height difference affects mounting pressure and overall contact.
http://forum.notebookreview.com/thr...ounge-phoenix-4.809589/page-336#post-10950232

 

Vapor chamber for the GPU and change the thermal pad thickness

 

Adding a single or dual VC into the P775 might be interesting, though thermalpads on the GPU will be like 4-6mm thick probably.

 

Heatpipes for the VRMs

 

Can I get that as a Sager performance package?

 

Like a lift kit for a truck

 

Now you just need a good marketable name for it.

 

Elite off roading SUV laptop because everyone loves SUVs now

 

Hi,


Thanks for the replies. I have been moving house this past week and just catching up.

I received yesterday the VC (it was delayed) and De8auer direct die place has been returned to sender (GLS shipping and delivery sucks ) and will receive the part next week.
The VC is 3mm, the copper IHS I had from Rockit is also 3mm (inside CPU die side to the top of IHS), so it should be a close match. If there was a slight excess height I could lap the CPU by 0.1 to 0.2mm. Do you think that using a second VC chamber will help with z-offset on GPU in this case or is it about improving heat-transfer on GPU by using a VC?
The wider surface of the VC looks better suited to mate with the heatsink plate, isn't it?

Question about the VC, are the fabrication dimples as noticeable (on bose sides) on yours too? Should I consider smoothing out the surface and lapping the VC?
https://imgur.com/gallery/6dGjTGK

Regarding the foam pad gasket/barrier. I am trying to locate the right density thickness foam here in my hometown (France). I'll ask @Falkentyne about it but do you guys think that for a thin gap, the silicon conformal varnish (two-three layers) would not be enough? On the Foam barrier Guide, they mentioned compressing the foam down to 0.2mm.
http://forum.notebookreview.com/thr...r-liquid-metal-safety-insurance-guide.817207/
Note that I have on hand Kafton and 33+ tapes, Silicon Conforming liquid.

Meaker@Seager: Regarding the vents, I agree it is not the best. It was removed and my P775 has been running without the back cover and directly onto the cooling pad (4 fans) for the past 2.5 weeks. It does help lower temperature a bit but I am still hitting 99-100c on the CPU when running VR applications and the room temp is above 26c.
https://imgur.com/gallery/BiAxYFG

I'd be open to creating more vent slots if it could be done cleanly.
https://imgur.com/gallery/UyJtpyz
II had access to a CNC I create a few more slots parallel to the one already existing, what do you think could be done? Suggestions on improving ventilations?

@Meaker@seager, the master of performance package and SUV Regarding temp, I was wondering if the heat under the center of the keyboard is from the PCH temperature and a sign it could be improved. Is there a PCH mode for the P775 chassis?

 

I measured with a digital caliper the 9th copper IHS (for delided CPU) I got from Rockit as a reference since I haven't taken out my 9900K cpu yet, and it was 3mm from the inside side (contacting die) to the top outside surface for the heatsink. Yes, it is close, promising, isn't it?

Yes, removing the cover helps with temps, and it is how I am running when doing heavy lifting with GPU and CPU, especially on warm days.
https://imgur.com/gallery/BiAxYFG

 

No, quite the opposite actually. Since you're removing the IHS (-3mm) and then adding a VC (+3mm) it remains in level with the GPU. Adding a VC to the GPU will raise it +3mm above the CPU. And since it's a unified heatsink it will create a bunch of issues.

Yes, the side with the indents, is meant for direct contact. Whereas, the uniform side is meant for the heatsink mounting.

Yes, it's the factory standart. I wouldn't want to sand it down, remember the vapor chamber is quite thin and the internal pressures can be quite high. Sanding it down might affect the overall VC integrity.

• Due to the special construction there is no surface curvature / puffing up and therefore no additional stress for the contacted components.

As long as it's highly compressible foam, as it must not interfere with the mounting in any way for proper contact.
https://www.amazon.com/Duck-Replacement-Conditioner-24-Inch-1285234/dp/B002GKC2US

How big a difference in temperature did you see removing the bottom panel vs having it on? If it's greater than 7c, I would mod it.

 

PCH mode?

 

Think he meant mod.

@D_Loa I changed the stock thermalpad to Gelid and saw 5-7c lower PCH temps, pad itself is ~4mm. The P775TM1 does not have a heatspreader like the P870 does, so the heat is sunk into the metal cover above it via the pad.