|
Post by Vinster on Jul 21, 2018 15:20:29 GMT -5
I want to test the difference between the AS5 I use and the graphite pads. I'm away this week. hope to do it after next weekend.
Vin
|
|
|
Post by mllrkllr88 on Oct 2, 2018 9:58:52 GMT -5
I want to test the difference between the AS5 I use and the graphite pads. I'm away this week. hope to do it after next weekend. Vin Did you ever do that testing? For me the difference was a few degrees C. The graphite is slightly worse but it has some side benefits for sure, especially for guys like us that change CPU's every 2 minutes. In other news, I have some nice sponsors on board for the next Freezer' Burn...go dig out your red cards
|
|
|
Post by Vinster on Oct 2, 2018 13:10:44 GMT -5
No, I never did. I actually haven't benched since really (Not Even Gamed ). RL is too busy. I've done 1 bench test on my main rig for team cup but that's it. Working a lot and no time to play.
I've read reviews where the average temp loss is 2 degrees at ambient. I have not seen anyone do a test with cold tough. The best I can do is -25 with my chiller (once it's hooked up) I wonder if that 2 degrees grows or shrinks if running anything sub-zero...
Vin
|
|
|
Post by mllrkllr88 on Oct 3, 2018 23:49:15 GMT -5
I wonder if that 2 degrees grows or shrinks if running anything sub-zero... The technical data states that W/mK actually increases as temperature goes down, which would be a good thing. However, real world results on LN2 have been less than desirable for many people and in my case I have seen a loss of 100+mhz with modern CPU's. I have never tested with -25c ish temps, it might be worth checking out. They may be expensive, but its pretty hard to beat the two mainstream extreme thermal products: grizzly kryonaut and KPx blue.
|
|
|
Post by ShrimpBrime on Oct 5, 2018 17:31:10 GMT -5
Its so very interesting the chat about thermal pastes and different temperatures.
While people in mainstream argue effects of most pastes the end result is always the ambient temp which in most cases will fluctuate with the thermal margin of the air conditioning system for example.
So testing pastes even the same day taking 3 houra of time over say 5 processors, the air conditioning system will have run and shut off multiple times.
End result from any past testing and none recently is to stay away from cheaper pastes. Most people can live with a 1c or 2c difference.
With LN2 I found that simply swapping out a power supply can increase a clock speed scuicide run. It only took from 850w to 1000w.
End result of any overclock is not truly reliant on any thermal compound.
If some one wanted a true thermal transfer with meaning.... They would solder their water block to the IHS plate. But alas the issue becomes the retainer arm and socket joint which would need to be removed so that the soldered to block combo could seated into the cpu socket.
Ive always used Antec Formula 7 and 6 diamond compounds. Even on 90% of the naked chips ranging up to 220w.
Remember when AS5 was a thing lol..
|
|
|
Post by Aleslammer on Oct 26, 2018 17:18:35 GMT -5
Thermal paste!!!! Personally use Arctic MX-2 not that it's any better just comes in 65g tubes and cheaper than most, need a little more turn the chiller down a little.
|
|
|
Post by georgekokovinis on Oct 26, 2018 17:37:09 GMT -5
Ambient :
Thermal Grizzly Condactonaut under and over the IHS. Copper water block obligatory.
Sub zero :
Thermal Grizzly Kryonaut under and over the IHS.
Best of all. But expensive.
|
|
|
Post by ShrimpBrime on Oct 26, 2018 17:40:49 GMT -5
Ambient : Thermal Grizzly Condactonaut under and over the IHS. Copper water block obligatory. Sub zero : Thermal Grizzly Kryonaut under and over the IHS. Best of all. But expensive. Thermal Grizzly Kryonaut on die naked only. Don't know why people keep putting the IHS plates back on. Seems pointless to remove it for 1-2c difference in temp when you would likely drop closer to 10c without the plate all together.... ??
|
|
|
Post by Macsbeach98 on Oct 26, 2018 18:50:33 GMT -5
With late model intel a 20c drop can be seen by changing the paste under the IHS the reason for putting the top back on the cpu isnt high enough to clear the top of the socket and even if you take the load plate off completely the wafer is very thin it doesnt take much pressure at all to crack the die. In fact some skylake CPUs have been known to crack cold after having the IHS off then back on if all the glue was removed but that hasnt happened to me.
Oh yea I use Gelid Extreme under the IHS and MX4 on top for water for cold its gelid all round. Would like to try some of the blue Kingpin paste actually.
|
|
|
Post by ozz on Oct 26, 2018 22:44:59 GMT -5
i use this for any testing, subbing in bot , comps i use mx-4, its paste is cheap, comes in a 30g tube, about 2 degrees behind mx-4 , when youre changing blocks and cpus for different platforms or testing any cpu , im probably on my 10th or more tube of it now i know this with intel up to ddr3 anyway, if you remove the cpu retainer thing you have to use a 50 x 50 mm water block, a 60 x 60 mm block wont fit , it hits the sides of the socket bracket when a cpu is delidded, i found that out the hard way , an no i didnt kill anything , it holds the block off the cpu chip and you cant get a connection because the sides are higher and the block wont fit within the retainer bracket, the 50 x 50 mm doe2 fit
|
|
|
Post by ShrimpBrime on Oct 27, 2018 9:21:46 GMT -5
With late model intel a 20c drop can be seen by changing the paste under the IHS the reason for putting the top back on the cpu isnt high enough to clear the top of the socket and even if you take the load plate off completely the wafer is very thin it doesnt take much pressure at all to crack the die. In fact some skylake CPUs have been known to crack cold after having the IHS off then back on if all the glue was removed but that hasnt happened to me. Oh yea I use Gelid Extreme under the IHS and MX4 on top for water for cold its gelid all round. Would like to try some of the blue Kingpin paste actually. I didn't realize those Intel chips are that delicate. I've delidded a couple, but on customer request reseated the IHS plate back on so never tested one naked even though I'd like to have. Thanks for the information Pete.
|
|