Message boards :
Number crunching :
Who uses or has experience with Thermal Grizzly Kryonaut?
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Uggh!@ This is not what I expected or wanted. The problem may be the IHS is CONVEX as indicated in the photo on a Tom's Hardware story about a LN2 overclocking attempt. Lapping In Pursuit Of MHz Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Wiggo Send message Joined: 24 Jan 00 Posts: 34754 Credit: 261,360,520 RAC: 489 |
Move a straight edge around both the IHS and the base of the cooler as you may have pulled off a daily double (it'd be a real bummer if both are totally opposite to each other). Cheers. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I already know the XSPC Raystorm Pro cpu block is flat as a mirror. The question will be what condition the IHS on the new chip is in reality. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
bluestar Send message Joined: 5 Sep 12 Posts: 7031 Credit: 2,084,789 RAC: 3 |
Should it be about the processor itself, or should it rather be the cooling paste, because here not such a thing as any glue either, when it comes to the paste which could be used. I make it some two rotating fans or blades for that of the Noctua DH-15, or the similar, and also that the thick pieces of metal in between, should act as a coolant as well. Also a large roof mounted fan as well, and only guessing that it should not counteract the processor fan for that of possible airflow inside. Therefore it becomes a question of using cooling paste, when it also could be a regular processor fan as well, which should be making it a free flow of air as well inside. Also that the left side panel is still off, and here I do not know if it helps either, or perhaps makes a difference. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
I already know the XSPC Raystorm Pro cpu block is flat as a mirror. The question will be what condition the IHS on the new chip is in reality. I did a lapping job on my Raystorm Pro. It was very slightly convex. I have a nice square foot piece of 1/2" plate glass that I use for a reference and a lapping base. And a piece of hardwood that has been lapped itself for hand work. Start with 200 grit and end up with 5000 grit by the time I am done. My 6950x is slightly convex as well. I shall be lapping that when I take things apart to move the rig into a new larger case soon. I am not brave enough to try delidding it. Meow. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Well I am too impatient. I'm going to pull the block again and look at the IHS for flatness before re-applying paste. Will check the block again but I examined it well before I installed it the first time to see how it was constructed. I checked flatness in both axis with a utility knife blade and my LED flashlight and looked for light escaping under the blade. Didn't see anything worth worrying. I don't know how extreme of measure I am willing to go to lap the IHS if necessary. Never done so previously. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
It's a good block, Keith. I am talking about a very very small bit of convex. The Aquacomputer NEXT blocks will surpass it's performance, but at a much higher price tag. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I just put the system back online again. Tore it down completely again. IHS was flat, almost no light leakage under the utility knife blade. Rechecked the cold plate. Pretty flat in one axis and just some very small convexity in the other axis. So I broke the loop apart and lapped the cold plate. No light leakage under the utility knife blade. Put the loop back together and bled the trapped air out of the system. Put a line of Thermal Grizzly Kryonaut down the midline of the Ryzen core which is the preferred method. Just got the loop back up to thermal equilibrium with cpu/gpu BOINC loading. Still at 70° C. Removing the block again and switching back to my MX-4 paste. Will try to get back to my original temps of 62-64° C. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
I shall be interested to see how the MX-4 does for you in comparison. Are those chips considered difficult to cool? I have had no AMD running here for a long long time. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Yes, they are if they are overclocked. It doesn't help that AMD for some silly reason decided to add a +20° C. offset to the reported internal chip temperature for the "X" models like my 1700X and 1800X. So any monitor that isn't aware of this, like CPU-Z for instance or HWMonitor reports the temps at 95° C. You have to mentally subtract the 20° C. offset to come up with the "real" cpu temperature. Only SIV and HwInfo64 have adjusted the reported temp in a new "alias" temp called CPU-0 Temp in SIV and Tdie in HwInfo64. Well I decided to try the Noctua HT-1 that came with my NH-D15 cooler since it appeared higher in the thermal tests than MX-4 in the link in this thread for S&G. I'm looking at 62-64° C. right now compared to the 70-72° C for the Kryonaut. Looks like Kryonaut is a big loser for me. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Good to hear you got the temps down. I am sure you don't wanna tear it apart again at this point, but if you do try the MX-4 against the HT-1, that would be interesting to know as well. I seem to recall there was a similar problem with temperature reporting when the Intel core architecture was new as well. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I'm not going to tear it down again. Getting old by now. Will try the MX-4 again at some time. The MX-4 was giving me the same 62-64 C. temps as I'm getting with the HT-1. On those test charts, HT-1 was only a few tenths better than MX-4. I would think that was in the margin of error. But I do believe they did multiple applications/removals for each paste with the same application method to try and eliminate any variability. At least, thank goodness, it's been reported that the new Ryzen+ and Threadripper + chips that are coming in April have removed the artificial +20 degree offset for Ryzen and the +27 degree offset for Threadripper. AMD has come to their senses after everyone in the industry called out their stupidity at the Ryzen launch. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Can't say I blame ya. Let er run. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Vipin Palazhi Send message Joined: 29 Feb 08 Posts: 286 Credit: 167,386,578 RAC: 0 |
I am still using just plain old Vaseline. Cheap, doesn't evaporate, just use enough to make the IHS barely sticky, never had a problem on any of my crunchers. Fills those micro hills and valleys and doesn't get in the way of the metal-to-metal contact. Never have had to redo it, either. Cheap (like me), too. I wasn't aware Vaseline had such a use. We learn something new everyday :) ... but I don't think I will use Vaseline over a good thermal paste. |
Cruncher-American Send message Joined: 25 Mar 02 Posts: 1513 Credit: 370,893,186 RAC: 340 |
Well, it really does work for my crunchers - never a problem with temps, and easy to clean up, and non-conducting of electricity. Very nice when playing around with stuff on test benches, too, which I do when testing for selling stuff on eBay or Craigslist. Or when testing stuff bought there. If you do decide to give it a try, get the straight stuff without additives, just the yellow goop in the jar. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
That's an interesting method I'd never heard before. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Al Send message Joined: 3 Apr 99 Posts: 1682 Credit: 477,343,364 RAC: 482 |
Keith, I just got done putting my daughters computer (i7-6700) under water, using a EVGA CLC 240 because using the crappy stock air cooler, it was sitting around ~60* or so when running BOINC. Well, I had picked up a tube of Kryonaut based upon the decent review I saw, but before you started posting your troubles with it. So, I used it on the EVGA cooler, after removing completely the factory crap that was over-globbed on on the CPU, and the stuff that EVGA had used on the waterblock. I put it on, spread it thin and even, then tightened it down little by little, properly. Booted it up, and it is idling at 27-28*, and the coolant temp is around 24. I have the pump cranked up to 100%, and the fans at about 40-50% due to their well documented noise. So, when I fired up BOINC, it shot up to around 50, which I feel is pretty high for a water cooled CPU. Do you have any thoughts/opinions on the situation? Like you, I'm not terribly excited about tearing it down and trying it again, but if it seems out of line in terms of heat, I might just have to do that. I am also ordering some industrial TIM from Germany to try, as it isn't available here in the US. I think I still have some AS-5 sitting around, but not sure exactly what the shelf life of that stuff is, and I know it's at least a year and a half old at this point. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
I ended up ordering noctua NT-H1 from this thread. Cleaned off the thermal paste that I had bought locally applied it, now it's running high 40-61C (there always seems to be that 1 core that runs hotter than the rest) But it's much better than it was so I'm happy with it. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Al, I've tried the stuff two times now and have come away not impressed at all. There shouldn't be such a delta T rise that fast on water. That is the same thing I saw on the Kryonaut myself. When I use the MX-4 or the HT-1. The temp rise is very slow. It takes at least 30 minutes before the coolant temp is heat soaked and reaches equilibrium. I let the Kyonaut "bake" for two days hoping there was a "cure" time like with AS-5. I never saw that. And Internet searches never came up with any definitive comment that it indeed has a cure time. I would either wait and order something different or just use the AS-5. Unless you are planning to get replacement TIM faster than the five days cure time of the AS-5. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I've read in a lot of places, that HT-1 is a very good TIM. In fact one of the review places, but I can't remember who exactly, uses it as their standard test TIM for all their reviews. I've come away impressed with it. At least equal to my MX-4. A viable alternative to MX-4 and much preferred to Kryonaut. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.