Innovation Cooling's Diamond 7 TIM test results

[Bail_w]

the red imprinted one on the left

Looks like you have great contact all across most of the paper

Thats what i was wondering about, confused at first why you want me to sent the right one lol.

Edit: So, should i try to use TIM with the Print and Donors? Apply TIM then Print and Donor?

[Martijn]

Thats what i was wondering about, confused at first why you want me to sent the right one lol.

Edit: So, should i try to use TIM with the Print and Donors? Apply TIM then Print and Donor?

I wouldn't try that. It's a pressure test, remember.

Oh well, you might as well test it to see if the TIM spreads the pressure?

[tastymannatees]

Thats what i was wondering about, confused at first why you want me to sent the right one lol.

Edit: So, should i try to use TIM with the Print and Donors? Apply TIM then Print and Donor?

Not what we are looking for right now, On a second go around we can target specific points of interest.

On this test we are looking for contact and pressure and will compare that to our thermal result. The Test with the TIM will spread the pressure some but might obscure problem areas.

Also this is kind of a complicated project on my end coordinating 4 different forum groups with 60 different tests so one thing at a time helps me maintain some kind of focus on the whole thing.

[David_L6]

I did manage to get the Diamond 7 in another computer today. It replaced AS5. The AS5 had been in there for several months (can't remember how long right now). The fresh application of Diamond 7 dropped the temps of the Q6600 B3 2°C..... Here's the kicker though. Ambient temp was 3° higher with the Diamond 7 than with the AS5. Once the house cools down tonight I'll check it again.

[littleowl]

Dang bail that is a nice print. Mine sucks I will have to lap mine one day

[tastymannatees]

we will have to give away a prize for the best print

[tastymannatees]

Andrew, I think you messed the 2 envellopes up. The actual pressure film (the left one on Bail_w's pictures) was in the donor envellope here as well

Does not surprise me..... After cutting up 300 1.5 squares I was a little cross eyed. Silly me I thought it was going to be a 10 min job, send out some samples to a few people

[STEvil]

Would be interesting to do the pressure film thing with a lapped cpu and lapped cooler.

Very nice thermocouple by the way

[Bail_w]

Fyi, my d-tek waterblock and IHS was not lapped.

[littleowl]

Raw image is the first image. second is statistics, next we have histogram, line scan, 2D and 3D models.

I found this info very interesting! seem how my fuzion is only putting about 42lb of pressure!

[littleowl]

I wanted to tell everyone I am sorry about the big images I didn't have time to sit down and adjust every one! it was a pain as is. Also here it is 10pm and I worked 12 hr today and I have to get up and do it again tomorrow.

These were my temps when I started these tests with D7
Test 4
room temp down just a little to 29.4
compound IDC7

Idle:
Core 0:38
Core 1:38
Core 2:35
Core 3:32

Load:
Core 0:47
Core 1:47
Core 2:45
Core 3:43

But after the pressure test and remount I am not getting something a little different. I know it is my fault but my numbers look more like this.

Load:
Core 0: 48
Core 1: 46
Core 2: 38
Core 3: 35

[Bail_w]

@ littleowl, the image size are fine.

[tastymannatees]

What's interesting to me is that the paper has a threshold of 28 psi.

And on the line scan in the center where the die is it looks to be barely enough over the lower threshold to get a reading.

And on the contour image and pseudocolor shows 35 psi and less where the die would be.

On the histogram you have about 50% of the IHS under 40 psi and above 40 psi most pf the pressure seems to be focused on the edges.

Anybody else have a comment or observation?

[Hazaro]

Images are fine, that's a nice chuck of data you get from that

Got mine in today, will test a while later

[Martijn]

Hmm, I just tested the Intel stock cooler, and it only applies pressure on the edges of the core. It is properly red, though

I think I'm going to apply some TIM here as well, and then do a pressure test. I'd love to know if pressure is then spread properly.

[tastymannatees]

Hmm, I just tested the Intel stock cooler, and it only applies pressure on the edges of the core. It is properly red, though

I think I'm going to apply some TIM here as well, and then do a pressure test. I'd love to know if pressure is then spread properly.

I would use as thin a paste as you have ICD7 is kind of thick and takes time to spread cold and probably would skew the result.

One of the reasons to do it dry is that it reveals the imperfections. whetted with grease they fill in and in those void bondline areas the paste can be 2X or 3X thicker losing a good portion of performance. you will still have pretty much the overall psi that you had when it was dry but you lose the more important contact info.

[123bob]

Some results from my problem child, Farm-09.... I did two test runs per Tasty's request. (BTW, I have 5 squares to test with. There were six donors but only 5 print sides.)

Farm-09 thermals are in post # 202. It is a Q6600 running on an IP35 Pro, with a stock TRUE. It is currently stock clocked, but will go up tonight. I was waiting to do this test.

Here's the pic.



Pretty self evident that this CPU has a serious hill in the middle. One would think that should be good since it's right over the die, but I don't know. The tests are fairly consistent with each other, with some minor variances. I tried to put the screws in the same way, in the same order both times.

I take this to mean that the TRUE could stand to have more pressure, or the geometry of the CPU hill is preventing this. It makes me wonder how my Zalman 9700, Farm-07 test will come out.....That is even whimpier at pressure than this.

For the three samples I have left, I plan the following. (Tasty, let me know if you want different.) I'll burn two samples on Farm-07 with its 9700. The last sample I will lap Farm-09, then pressure test, and do a final thermal test to see the lapped results. I think that would be interesting.

On a side note, I've taken to DDT's scheme on machines. Here is Farm-09 on a rack mount tray. MUCH easier to work on this way. I doubt I'll be using rack cases anymore. I may as well "go industrial"...



Thx Victor!

Regards,
Bob

[tastymannatees]

123 Bob What's interesting to me is that when thermalright came out with it's first heat pipe cooler I tested it on a 1cm synthetic die and performance was good but not great. When I put it on my 1 in die (whichs gives me pretty much what you see in system within about 10&#37 Temps dropped like 6 C. It had two inner heat pipes and two outer heat pipes I theorized at the time that the heat was just not able to spread to the outside pipes on a small die. This was a sink designed for an IHS.

You have just about 1 cm on center and it looks like the edges are taking the brunt of the pressure. I would think just from a casual look at things that even when these IHS are lapped they are lapped in a relaxed state and the central bubble is intact, in OWL's case the center is concave, but in any event the bends on the edges on both act as a reinforcement or stiffener to the IHS against pressure.

Just a thought, rather than lapping the IHS a solution might be to just file the edges to relieve the high points. Note the 77lb on Owl's edges my guess is your result will show the bulk of the pressure on edge also.

*****added note - Owl's and your number 2 both seem Biased to one side which would increase the gap on the opposite side.

The last mile is the hardest

[123bob]

123 Bob What's interesting to me is that when thermalright came out with it's first heat pipe cooler I tested it on a 1cm synthetic die and performance was good but not great. When I put it on my 1 in die (whichs gives me pretty much what you see in system within about 10%) Temps dropped like 6 C. It had two inner heat pipes and two outer heat pipes I theorized at the time that the heat was just not able to spread to the outside pipes on a small die. This was a sink designed for an IHS.

You have just about 1 cm on center and it looks like the edges are taking the brunt of the pressure. I would think just from a casual look at things that even when these IHS are lapped they are lapped in a relaxed state and the central bubble is intact, in OWL's case the center is concave, but in any event the bends on the edges on both act as a reinforcement or stiffener to the IHS against pressure.

Just a thought, rather than lapping the IHS a solution might be to just file the edges to relieve the high points. Note the 77lb on Owl's edges my guess is your result will show the bulk of the pressure on edge also.

That is a good thought. I may try it. I'm wondering how much of the edge print is due to mounting the sink though. Do you know how quickly this stuff reacts to pressure? The process of screwing the mount down is rather quick compared to the time I let it sit clamped together. BTW, I left sample #1 clamped for 11 minutes. Sample #2 got 17 minutes. I thought this might prove interesting. Everything is fully noted on the envelope.

Anyone else going to test a lapped setup?

Regards,
Bob

[tastymannatees]

That is a good thought. I may try it. I'm wondering how much of the edge print is due to mounting the sink though. Do you know how quickly this stuff reacts to pressure? The process of screwing the mount down is rather quick compared to the time I let it sit clamped together. BTW, I left sample #1 clamped for 11 minutes. Sample #2 got 17 minutes. I thought this might prove interesting. Everything is fully noted on the envelope.

Anyone else going to test a lapped setup?

Regards,
Bob

Bail_w does not have that edge print (Bail_w is winning the best contact contest so far)

basically the print is done when you first clamp and you reach max load. the 10 min is to allow the mechanicals to settle in. Probably does not make much difference but am trying to keep things uniform.

[Martijn]

Since everyone is posting their pics: this is the first one on the stock Intel sink. As you can all see, contact isn't perfect. I will be testing again with TIM to see if the pressure is then spread more properly.

[tastymannatees]

Shooting from the hip I would say the IHS is concave and the edges have no flex so there is a void in the middle making the paste overly thick with most of the heat conduction through the paste rather than split uniformly between direct contact and paste with a smaller BLT as in Bail_w's pic (Bail_w still in the lead)

Slightly surprised at the range covered so far

[Martijn]

Oh wow! I just took off the cooler again and I have never seen contact like this before. This stuff amazes me every single time

Just look at the thin film that's left on the CPU

[tastymannatees]

The paste left center is pretty thick and would tend to confirm the contact result. The optimum bondline should resolve to around the compounds particle size < 40&#181;m. So good contact should look a little like the faint circular halo on the paste impression matching the pressure film result

[Martijn]

Sorry for the double posting. Anyone have one or two extra pressure tests left? I blew a few, but now I don't have enough samples to test on my other heatsink

PM me please.

EDIT: No more double post lol.