Too much Thermal Compound? - picture -

[Nicholas Peyton]

maybe an image of the chassis with the loop in it, just to be sure noting is wrong that we may be able to spot.

What TIM are you using?

Also what's the number of fins per inch (fpi) for that rad?

TIM is 'thermal grizzly kryonaut' -- it actually comes with EK kits now and is one of the best

RADIATOR
Dimensions: 400 x 130 x 38mm (L x W x H)
Liquid Capacity: ~200mL
FPI: 38

PICTURES OF LOOP:

<<normally an 'EVGA 1080 Classified' in there but took it out for leak test
<< half full - is that right?
<<block assembly before filling

Sounds about right to me I could see mine doing similar in prime, hence prime means naught to me , I can game at different clocks to what I can crunch, if I were you I'd worry about in use temps but not prime it uses too much of a processor and nothing bar crunching will use similar performance making it passe I have enjoyed years of gaming with mine clocked at 4.8 but it would not be here today if I'd crunched at that , I folded mostly and have started crunching recently (4.5@)because this CPU is nearly EOL in main use terms for me all temps at 52 ish + - 5.
I have crunched and folded at 4.8 with this system as is but I had 62 degrees soaked heat + - 5 and a room so warm I felt sick so I calmed it a bit.

Okay but my temps haven't improved since upgrading from an AIO Corsair H80i GT which is an £80 ($100) 120mm AOI cooler.
This custom kit is worth over £300!! And no difference in temps :-( I was expecting at least 10c

*EDIT- 29th Oct 2016:
Above statement in bold (last paragraph) no longer true. Had to wait for new thermal paste to cure -- read posts on page 3 for info on my journey and full story

 

[fullinfusion]

TIM is 'thermal grizzly kryonaut' -- it actually comes with EK kits now and is one of the best

RADIATOR
Dimensions: 400 x 130 x 38mm (L x W x H)
Liquid Capacity: ~200mL
FPI: 38

PICTURES OF LOOP:

View attachment 80530
View attachment 80532 << half full - is that right?
View attachment 80533 <<block assembly before filling




Okay but my temps haven't improved since upgrading from an AIO Corsair H80i GT which is an £80 ($100) 120mm AOI cooler.
This custom kit is worth over £300!! And no difference in temps :-( I was expecting at least 10c

From what I can see looking at the grizzly paste you used way to much, about half of what was in the 1st post should be best.

Try that and see how the temps go.

 

[Nicholas Peyton]

From what I can see looking at the grizzly paste you used way to much, about half of what was in the 1st post should be best.

Try that and see how the temps go.

I used about 2/3rds of what was in the original post at the top. Do you think that could still make a 5-10c difference ? (Before I drain the whole system and re-apply).....

My method when I installed the corsair AOI was spreading a "very very" thin layer all over the IHS (which isn't a great method -- as I've now learned)... so with that in mind -- I'd of expected much better temps with this upgrade.

I was running the 4.8ghz daily-everyday on the £80 corsair unit-- really expected to be able to push this to at least 5ghz with this £300 upgrade :-(

I can see a few little bubbles inside the block still (been running for 24hrs) but I heard they don't actually make that big a difference...?

 

[Loosenut]

You can remove the CPU block without draining the system.

 

[Nicholas Peyton]

I used about 2/3rds of what was in the original post at the top. Do you think that could still make a 5-10c difference ? (Before I drain the whole system and re-apply).....

My method when I installed the corsair AOI was spreading a "very very" thin layer all over the IHS (which isn't a great method -- as I've now learned)... so with that in mind -- I'd of expected much better temps with this upgrade.

I was running the 4.8ghz daily-everyday on the £80 corsair unit-- really expected to be able to push this to at least 5ghz with this £300 upgrade :-(

I can see a few little bubbles inside the block still (been running for 24hrs) but I heard they don't actually make that big a difference...?

true -- but a bit dangerous -- if a tube should come loose as it gets twisted/moved around lol -- okay -- i'll give it a go

 

[Nicholas Peyton]

Just a little update before bed people:

Very very interesting actually -- may also be of use to some other guys here.

Ran my AMD CPU hot for a few hours non-stop tonight (60c ish) -- now remember in an intel CPU that would translate to around 80c as AMD reports about 20c lower on the core.

Kept it running at this hot temp for quite some time ). Temps at idle have dropped a little and temps at load seem to be much more in control now.... I also switched off for a while and left to cool before continuing experiment.

Now I don't know if someone more experienced is able to explain the science behind this?

This is my first loop -- but even with my old Corsair AOI I remember temps "settling" down after 2 days or so. So my assumption (maybe an expert can confirm) it can actually take your thermal paste a couple of strong, powerful, hot workouts until it becomes most effective? Maybe it spreads out more as it heats up; or maybe it "gets in and fills up" more microscopic pits on the copper plate/IHS and forms a more "intimate" or "snug" contact after a good, powerful burn-in/workout then cool down.?

Moreover something else I also noticed... if I keep my VRM cool by pointing an extra fan at the BACK of my opened case (pointing at the BACK of where the VRM sits on the rear of the board) - so 2 fans on the VRM front & back - my thermometer reads a massive temp drop on the VRM area -- but wait for it... THE CPU TOO

The heat from the motherboard must conduct along the circuitry from a hot VRM to your CPU. I am now prime95 stable at 4.8ghz and 1.4875v WITHOUT throttling as long as I keep VRM cool.

So there we have it -- it doesn't matter how amazing your CPU waterblock is -- if you don't adequately cool your VRM it doesn't matter how good your loop is you will never reach the CPU temps you dreamed about.

As soon as I removed the 2nd fan from VRM area (back of case) temps started to rise again on CPU.

Anyway guys I'm going to do a bit more testing tomorrow but for now I'm quite happy -- for the first time ever I can now run prime95 or 'fold' WITHOUT thermal throttling at 4.8ghz & 1.4875v on my FX-8350.

Tomorrow night I'm going for 5GHZ

BTW I never got around to changing TIM; not sure I want to now, but then again I am also curious lol -- so not decided yet. Imagine if I done that too?

lol anyways Good night

 

[basco]

thats called "burnintime" !
some paste needs it some not so much.
what else you could look is how tight the screws are-can make a lot of diff- some think tighter is better but try during a stress test to loosen them and watch temp

 

[Folterknecht]

Yep burnintime - just check your screws on the block and retighten them a little bit if necessary.

 

[Nicholas Peyton]

thats called "burnintime" !
some paste needs it some not so much.
what else you could look is how tight the screws are-can make a lot of diff- some think tighter is better but try during a stress test to loosen them and watch temp

Yep burnintime - just check your screws on the block and retighten them a little bit if necessary.

So loosen it a little, burn in again(I assume this helps it spread even better)??

Then tighten again (slightly tighter than before)..

do I tighten again while it's still hot?

 

[CAPSLOCKSTUCK]

No

basco said

some think tighter is better but try during a stress test to loosen them and watch temp

so, loosen them ever so slightly while you are stress testing and see if it makes a difference.

 

[Nicholas Peyton]

No

basco said




so, loosen them ever so slightly while you are stress testing and see if it makes a difference.

I realise what he said but I couldn't see how loosening how tightly the block is pushed down against the IHS could improve temps -- unless the reason was to allow the TIM to "more freely" spread as it heats up before tightening again.

But if it's only to be loosened and temp checked (with aim of possibly seeing a lower temp) how does that work then? What's science behind that?

What am I missing lol?

Hope that clears up intensions of last post.

 

[CAPSLOCKSTUCK]

some think tighter is better

 

[sneekypeet]

I always, and I mean always, use every bit of thread offered with a cooler. In the case of this EK block, I would also use all of the thread in the hardware. It is not like it is an all-thread stud holding the block on. In other words, I screw things down until they stop! Coolers are designed to use "mounting pressure" and threading on hardware is made to give you what they designed it to use. Maybe I shouldn't have said always, as some blocks to come with all-thread, but the better ones tell you to measure the amount of stud left above the mounting nut for correct pressure.

 

[CAPSLOCKSTUCK]

Finger tight with thumb head screws and then an equal turn with a Phillips is what i do.

The important thing is that they are equal and obviously excesive force is never a good thing especially when it comes time to remove them.

 

[cdawall]

I always, and I mean always, use every bit of thread offered with a cooler. In the case of this EK block, I would also use all of the thread in the hardware. It is not like it is an all-thread stud holding the block on. In other works, I screw things down until they stop! Coolers are designed to use "mounting pressure" and threading on hardware is made to give you what they designed it to use. Maybe I shouldn't have said always, as some blocks to come with all-thread, but the better ones tell you to measure the amount of stud left above the mounting nut for correct pressure.

I didn't crank mine all the way down, that's on the 2011 setup . With them maxed out temps seemed consistently a degree or so warmer

 

[sneekypeet]

I didn't crank mine all the way down, that's on the 2011 setup . With them maxed out temps seemed consistently a degree or so warmer

Never really found that issue. LGA775, LGA115X and LGA2011v3 I take them till they stopped. Initially for my testing, I did try messing around, but did not find anything out of margin of error, like 0.25 degrees when i averaged the core temps. Also since this thread is on AMD, we do not have to worry about the LGA at all, his CPU can only go so far into the board.

 

[CAPSLOCKSTUCK]

Also since this thread is on AMD,

thats why i mentioned the clips, its easy for them not to attach correctly especially if the psu is at the top of the case and restricting finger access.

 

[sneekypeet]

thats why i mentioned the clips, its easy for them not to attach correctly especially if the psu is at the top of the case and restricting finger access.

Are you confused with this post? https://www.techpowerup.com/forums/threads/the-indomitable-cooler-help.227236/#post-3545957 Or am i having some sort of mental moment here?

AFAIK that EK block he is using does not have clips. It should both through the motherboard.
https://www.techpowerup.com/forums/threads/too-much-thermal-compound-picture.227186/#post-3545554

 

[CAPSLOCKSTUCK]

AFAIK that EK block he is using does not have clips. It should both through the motherboard.
https://www.techpowerup.com/forums/threads/too-much-thermal-compound-picture.227186/#post-3545554

my mistake, im getting confused with the 8150 thread.......excuse me.

 

[sneekypeet]

my mistake, im getting confused with the 8150 thread.......excuse me.

Thank all that is high and mighty! I was sitting her scratching my head, confused as to what was going on. Reread the thread and got more confused. I was so close to just shutting off the net and taking a nap, it wasnt even funny!

 

[Steevo]

Watercooling with a phase change TIM will result in significantly longer "burn in" or settling times as the temps may not rise to the required phase change point to turn the TIM to a more fluid state and allow the excess to be pushed out, this is especially true with thicker compounds. I clamped off a coolant line to raise the temps of my CPU to accelerate settling.

 

[cdawall]

Never really found that issue. LGA775, LGA115X and LGA2011v3 I take them till they stopped. Initially for my testing, I did try messing around, but did not find anything out of margin of error, like 0.25 degrees when i averaged the core temps. Also since this thread is on AMD, we do not have to worry about the LGA at all, his CPU can only go so far into the board.

In past experience with AMD block you start to bend the waterblock plate and it lifts, no idea where the stops actually are on the EK ones and would hope they don't allow that to happen.

 

[Nicholas Peyton]

4 springs go on first then thumb nuts ontop of the screws. That pic was taken before those went on as I was checking the assembly was correct.

(You have to dismantle and reassemble the EK block with the 'jet plate' and 'insert' that best suits your socket (researched by EK there are 3 different ones in the pack and you refer to a table in the manual to match the optimal ones for your CPU).

Because I'm AMD (minority( I had to change both the 'jet plate' and the 'insert'. Seems a little complicated to begin with and a bit overwhealming considering everything else you have to also do when setting up your first loop. Then learning you even have to dismantle your entire block and re-assemble it perfectly.... but once you do it; its quite rewarding. I emailed a picture to EK support to make sure I'd got the orientation of the 'jet plate', 'insert' and 'mounting plate' all correct and they responded within 30 minutes. They always do Definitely A++ for customer service. Turns out I had in fact assembled it correctly and all EK said was "looks good". I definitely recommend them.

Temps I am getting now have definitely surpassed my expectations

Going for 5ghz tomorrow but I think its going to be quite easy to get it stable

Once I get 5GHZ I will have the exact same performance as AMD's current highest offering. The AMD FX-9590 (a 220w TDP beast).

Afterall - the FX-9590 is simply a better binned chip originally three times the price of the 8350 or even 8370.

Originally the FX-9590 retailed at $999 ~yeah I know you can now pick it up for £200 now but overclocking on AMD is so much fun -- when you consider the original price and it IS "still" there highest offering today

More on my 5ghz adventure tomorrow ...

---------------------------------------------------------

Watercooling with a phase change TIM will result in significantly longer "burn in" or settling times as the temps may not rise to the required phase change point to turn the TIM to a more fluid state and allow the excess to be pushed out, this is especially true with thicker compounds. I clamped off a coolant line to raise the temps of my CPU to accelerate settling.

Phase change TIM?

 

[cdawall]

Originally the FX-9590 retailed at $999 ~yeah I know you can now pick it up for £200 now but overclocking on AMD is so much fun -- when you consider the original price and it IS "still" there highest offering today

Release price was under $500...

 

[Aquinus]

What thermal paste did you use? Arctic Silver 5 needs to undergo several heating and cooling cycles to cure and won't achieve optimal heat transfer until then. You also don't need to use very much as it spreads out under the pressure of the water block being tightened to the motherboard and CPU. Extra compound only inhibits heat transfer and too much can increase temperatures significantly.