Overclock.net › Forums › Cooling › Cooling Experiments › [UPDATED 4/7/2012] Cram's TIM + Thermal Adhesive Experiments
New Posts  All Forums:Forum Nav:

[UPDATED 4/7/2012] Cram's TIM + Thermal Adhesive Experiments

post #1 of 3
Thread Starter 

Hello members of OCN.

 

Today I am proud to present an experiment that might or might not benefit everyone. My experiment is all about mixing thermal adhesives (TA) and thermal interface materials (TIM) for various applications.

 

For my part, the reason why I wanted to mix these 2 compounds is because I want to have the flexibility of changing my heat-sinks as opposed to a permanent solution on my GPU. The fact that I have the option to sell my GPU with the default configuration or my custom configuration is there is a very desirable option for a buyer in my opinion.

 

So enough of the wall-o-text and let’s get started.

 

Materials:

Arctic Alumina Thermal Adhesive (AATA 2 part thermal adhesive epoxy)

Arctic Alumina Premium Ceramic Polysynthetic Thermal Compound (AA-TIM)

Arctic Cooling MX-4 (MX4-TIM)

Toothpicks

Sticky notes

Heat-sinks (1 forged copper / 7 aluminum)

Dead MP3 player

 

Let’s start off with the weird component on the list, the dead mp3 player. Well I found this mp3 player I remember it being a Sansa mp3 player I found on the curb, flattened by a vehicle. So I picked it up since the buttons were still fine and I could transplant the micro buttons to my other mp3 player missing a volume control. But that transplant failed…anyway this served as a guinea pig for this particular experiment as there are some chips in the mp3 player that resembled the VRM MOSFETs of my HD 6950. So if you’re going to do what I did please find some sort of PCB that has chips in them so you can test your mixing ratios out.

 

You’re going to need some test heat-sinks for this as well…the Enzotech MOSFET forged copper heat-sinks include 10 pieces and my 6950 only used 9 so I had an extra one to play with. If you don’t have the copper ones, the aluminum ones will suffice if you’re using aluminum heat-sinks.

 

Sticky notes…this is one of the most versatile tools I have, they can be used as mixing pads, toothpicks and much much more…trust me. For this experiment I’m using them as mixing pads.

 

Toothpicks…I would suggest the bamboo ones that you can buy at the dollar store as they are stronger than the cheap wooden ones. These are used to mix the AATA and the TIMs.

 

There is only one company that I know that sells this kind of compound, so for this experiment I will be using the Arctic Alumina Thermal Adhesive

 

 

tim_mix_exper_002.jpg

 

 

As you can see I have purchased the Arctic Alumina Ceramic TIM as I thought they’d mix well together. As for another TIM the MX2 has proven its worth for me before so I had to get the newer iteration of the TIM from Arctic Cooling. (BTW their support is horrible as I tried to contact them about the thermal adhesive that came with the AXP-II and no reply.)

 

Here comes the hard part…trying to get the ratios right.

 

tim_mix_exper_010.jpg

 

As you can see here my first batch was a 30/70 ratio for the AATA and I added a ratio of 50/50 for the ceramic compound.

 

tim_mix_exper_016.jpg

 

The second batch was as written on the note…40/60 ratio but no Ceramic TIM included in the mix.

 

tim_mix_exper_019.jpg

 

The third batch was also just epoxy with a mix ratio of 80/20 with no Ceramic TIM included.

 

tim_mix_exper_022.jpg

 

My rule-of-thumb (ROT) was to leave everything in a cool dry place and wait for 24 hours for the epoxy to dry.

 

tim_mix_exper_031.jpg

 

After waiting for 24 hours…I tried to take them off which was my original plan. All of them failed the test except for one mix ratio and that was a batch I failed to capture on camera. That batch was 50/50 epoxy and 50 Ceramic TIM.

 

The results for the first 3 batches weren’t stable enough except for the 50/50/50(TIM). I’d like to add the 40/60 ratio batch into the stable bracket; it did stick but upon taking a micro blow-torch to the rear of the PCB and heating up the chips and the heat-sink fell off. So far off to a very bad start.

 

I will be trying to mix another batch with the AATA and Ceramic TIM once I finish testing the MX-4.

 

So far the best results that I have are the following:

First Mix: (50% Part A + 50% Part B) = 100% AATA Mix

Second Mix: 50% AATA Mix + 50% Ceramic TIM

 

Now for the second TIM the MX4…let’s see how this does when mixed with AATA.

 

tim_mix_exper_039.jpg

 

After cleaning up all the parts with cotton-buds and lighter fluid (aka the most awesome degreaser) I proceeded to mix AATA and the MX4.

 

tim_mix_exper_043.jpg

 

As you can see this time I took what I learned from the first test which was a 50/50 mix of AATA gives the best results. So I had a batch of 50/50 AATA on the left and MX4 on the right. As you can see they’re listed on the mixing pads.

 

tim_mix_exper_046.jpg

 

So here they are on the clean but broken PCB.

 

I will need to wait for another 12 hours and I will report back with my findings.

 

 


Edited by CramComplex - 4/7/12 at 8:30am
Intel System 2011
(18 items)
 
AMD System 2009
(14 items)
 
 
CPUMotherboardGraphicsRAM
Athlon II X4 630 Propus @ 3.5Ghz Gigabyte GA-880GM-UD2H Rev.1 ASUS EAH5450 SILENT/DI/512MD2(LP) 4x2GB G.Skill NQ DDR3 PC 1600 
Hard DriveOptical DriveCoolingOS
Seagate Momentus XT LG GSA-H30N Scythe Samurai ZZ Rev.B Win7 x64 Ultimate 
MonitorKeyboardPowerCase
LG CS 460 - 32" LCD TV Microsoft Reclusa Gaming Keyboard Thermaltake TR2 750w Fractal Design Core 1000 
Mouse
Microsoft Comfort Mouse 6000 
  hide details  
Reply
Intel System 2011
(18 items)
 
AMD System 2009
(14 items)
 
 
CPUMotherboardGraphicsRAM
Athlon II X4 630 Propus @ 3.5Ghz Gigabyte GA-880GM-UD2H Rev.1 ASUS EAH5450 SILENT/DI/512MD2(LP) 4x2GB G.Skill NQ DDR3 PC 1600 
Hard DriveOptical DriveCoolingOS
Seagate Momentus XT LG GSA-H30N Scythe Samurai ZZ Rev.B Win7 x64 Ultimate 
MonitorKeyboardPowerCase
LG CS 460 - 32" LCD TV Microsoft Reclusa Gaming Keyboard Thermaltake TR2 750w Fractal Design Core 1000 
Mouse
Microsoft Comfort Mouse 6000 
  hide details  
Reply
post #2 of 3
subbed thumb.gif
[ WOMD ] Rainuke
(17 items)
 
  
CPUMotherboardGraphicsRAM
i7-2600k Asrock Z77E-ITX R9 290X Crucial Ballistix Sport 
Hard DriveHard DriveCoolingOS
Seagate 7200 RPM Samsung 830 128GB XSPC AX240 kit - With custom Components windows 8 professional 
MonitorKeyboardPowerCase
2x Asus VG248QE Coolermaster TK Coolermaster Hybrid Define R4 
MouseMouse PadAudioAudio
Razer Mamba Rainbow dash custom Schitt Modi Bose Companion 3 
Audio
Objective 2 
  hide details  
Reply
[ WOMD ] Rainuke
(17 items)
 
  
CPUMotherboardGraphicsRAM
i7-2600k Asrock Z77E-ITX R9 290X Crucial Ballistix Sport 
Hard DriveHard DriveCoolingOS
Seagate 7200 RPM Samsung 830 128GB XSPC AX240 kit - With custom Components windows 8 professional 
MonitorKeyboardPowerCase
2x Asus VG248QE Coolermaster TK Coolermaster Hybrid Define R4 
MouseMouse PadAudioAudio
Razer Mamba Rainbow dash custom Schitt Modi Bose Companion 3 
Audio
Objective 2 
  hide details  
Reply
post #3 of 3
Thread Starter 

[UPDATE: 4 / 7 / 2012]

 

So all with all the testing done and the pictures taken…I’m ready to present the results of the experiment. (I will be doing another experiment on store bought 2 part epoxy…stay tuned for that.)

 

tim_mix_exper_046.JPG

 

The time has come to remove the sinks from their bonds and as we can see on the above pic…the 40/60 ratio and the 30/70 ratio were the easiest to remove. The 40/60 ratio was a bit tougher than the 30/70 as the 30/70 just slid out of the bond.

 

I did the test twice on 2 separate PCBs but I couldn’t get pics of the other one as I have forgotten all about them…stupid me. The other test subject was blasted with a micro-torch to simulate heat coming from the VRMs during usage…I did a quick 30 second blow on each of the chips and all of them stayed intact to my surprise! The AATA + Ceramic didn't hold out as much as the MX4 mix.

 

tim_mix_exper_047.JPG

 

This is what you can expect from the 30/70 mix, a blob that isn’t really bonded well but through the testing this has proven to be the weakest due to the mix ratio obviously.

 

tim_mix_exper_048.JPG

 

This second one is what you can expect from a 40/50 mix ratio. As you can see most of the compound stayed on the test VRM and not on the heat-sink.

 

tim_mix_exper_049.JPG

 

And the star of the show the 50/50 mix ratio. I think the most effective mix I’ve done with the MX4 and the Ceramic TIM.

 

tim_mix_exper_051.JPG

 

I just forced the copper heat-sink left and right and it broke off pretty easily like the 40/60…unlike the 30/70 where the sink just slid off.

 

Now on to cleaning the sinks…

 

tim_mix_exper_053.JPG

 

As I mentioned before Lighter fluid is one of the best electronics cleaners around…they don’t dry up as fast as 99% alcohol and they are very effective in removing stickyness.

 

tim_mix_exper_054.JPG

 

After soaking them for about 30 seconds I proceeded to wipe them off with tissue. Then had a second pass with lighter fluid again and finally 99% alcohol to get rid of most of the dirt.

 

tim_mix_exper_055.JPG

 

And here are the results after cleaning for about 5 minutes!!! You can see how the copper clearly reflects the bottom sink…now that’s clean!!!

 

And for the results of the experiment that lasted for about 5 days (due to work and stuff) I am happy to present my OWN findings and recommendations…

 

BEST MIX RATIO FOR STABILITY – BOND – HEAT TREATMENT:

 

AATA 50% + Arctic Ceramic TIM 50%

AATA 50% + Arctic Cooling MX4 50%

 

BEST MIX RATIO FOR REMOVABILITY

 

AATA 40% + Arctic Ceramic TIM 60%

AATA 40% + Arctic Cooling MX4 60%

 

I personally use the 40/60 ratio as I clean my parts every 6 months.

 

So take what you can from this experiment of mine and I hope some of your questions as to mix ratios can be answered.

 

I’ll be updating this thread as soon as I get some 2-part epoxies! So stay tuned!


Edited by CramComplex - 4/7/12 at 8:33am
Intel System 2011
(18 items)
 
AMD System 2009
(14 items)
 
 
CPUMotherboardGraphicsRAM
Athlon II X4 630 Propus @ 3.5Ghz Gigabyte GA-880GM-UD2H Rev.1 ASUS EAH5450 SILENT/DI/512MD2(LP) 4x2GB G.Skill NQ DDR3 PC 1600 
Hard DriveOptical DriveCoolingOS
Seagate Momentus XT LG GSA-H30N Scythe Samurai ZZ Rev.B Win7 x64 Ultimate 
MonitorKeyboardPowerCase
LG CS 460 - 32" LCD TV Microsoft Reclusa Gaming Keyboard Thermaltake TR2 750w Fractal Design Core 1000 
Mouse
Microsoft Comfort Mouse 6000 
  hide details  
Reply
Intel System 2011
(18 items)
 
AMD System 2009
(14 items)
 
 
CPUMotherboardGraphicsRAM
Athlon II X4 630 Propus @ 3.5Ghz Gigabyte GA-880GM-UD2H Rev.1 ASUS EAH5450 SILENT/DI/512MD2(LP) 4x2GB G.Skill NQ DDR3 PC 1600 
Hard DriveOptical DriveCoolingOS
Seagate Momentus XT LG GSA-H30N Scythe Samurai ZZ Rev.B Win7 x64 Ultimate 
MonitorKeyboardPowerCase
LG CS 460 - 32" LCD TV Microsoft Reclusa Gaming Keyboard Thermaltake TR2 750w Fractal Design Core 1000 
Mouse
Microsoft Comfort Mouse 6000 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Cooling Experiments
Overclock.net › Forums › Cooling › Cooling Experiments › [UPDATED 4/7/2012] Cram's TIM + Thermal Adhesive Experiments