What are the worst TIMs?

[:D:D]

By thickness, are you referring to the TIM?

Bond Line Thickness (BLT)

A quick google gives https://www.electronics-cooling.com...view-of-design-characteristics-and-materials/
old article though so maybe a bit out of date and some annoying pop-ups!

Nice PDF with no pop ups https://www.researchgate.net/profil...jYXRpb24iLCJwcmV2aW91c1BhZ2UiOiJfZGlyZWN0In19

 

[Noci]

Bond Line Thickness (BLT)

A quick google gives https://www.electronics-cooling.com...view-of-design-characteristics-and-materials/
old article though so maybe a bit out of date and some annoying pop-ups!

Nice PDF with no pop ups https://www.researchgate.net/profil...jYXRpb24iLCJwcmV2aW91c1BhZ2UiOiJfZGlyZWN0In19

Very interresting reading material, but if you read between the lines you'll notice the differences between the pastes we have now for PC's, they all operate within a small thermal window and the effect of the minimal differences in thermal conductivity, are neglectable. This is because the majority of the pastes are based on silicone oil with cereamic, aluminium and/or zinc-oxide fillers as heat conductors.
What differentiates the efficiency of them are the mixing ratio's and the viscosity, but still they are close to each other due to the physics of the components.

And indeed the BLT (as mentioned before: proper application of the TIM) is another major influencial factor if not even bigger, no matter if you have the best or the worst TIM.

Another showstopper can be the external factors as warping and bending of the mating surfaces. The only inluence the user has, is to properly install the coldplate with an even distribution of pressure.

Also from these studies I understand that the pump-out effect (which I personally never encoutered or recognised over the years, still using my 12 year old 20g syringe of Arctic MX-4) generally goes for all the TIM's in this product group, which is quite a concern. Of course per brand they make improvements by fine tunig the mixing ratios of the components to prevent from pump-out and that's where we see differences in our daily computing cooling adventures. Be noted that pump out is more significant on direct die than on an IHS due to die warpage during heating and cooling cycles.

In science Graphene is considered to be the rising star in the cooling techniques, amongst other technical disiplines where it looks very promising, due to it's physical properties. But because of manufactering challenges and implementation/application techniques still being developed, not yet being matured. In the (near) future we will see more of it, for sure.

For the ones who do not want to read research papers with technical mumbo jumbo, the following site has some nice illustrated answers on the most common questions regarding PC cooling. However they do get a little annoying with the recurring sales pitch (not that I can blame them for getting a piece of the pie ).

https://koolingmonster.com/insights

 

[R-T-B]

Clearly it's peanut butter.

No. It's insulation from your walls.

 

[Shrek]

https://koolingmonster.com/insights

Liquid Metal vs. Thermal Paste: Which One is Better? — Kooling Monster

A very enlightening article

" ... the real performance difference between liquid metal-based thermal pastes and silicone-based thermal pastes is not that big."

 

[LabRat 891]

BTW, after thinking on/off awhile. I have (experience with) the perfect "worst thermal paste".

A. It performs worse than most generic silicone-based pastes
B. It's been around for decades.
C. It's still manufactured and sold, today.
D. It's still 'useful', despite performing worse than even silicone-based thermal pads.

Arctic Silver - Céramique (2)

Arctic Silver Incorporated - Céramique 2

www.arcticsilver.com

Céramique 2 uses only ceramic fillers so it is neither electrically conductive nor capacitive. The tri-linear composite of aluminum oxide, zinc oxide and boron nitride

polysynthetic suspension fluid combines advanced synthetic oils

Basically, it's Diaper Rash Cream but, made w/ a Synthetic Petrol base and (purposefully) containing nano zinc oxides and alumina (IIRC, lab testing determined it's mostly zinc oxide)

It's useful for MOSFET TIM, and I've seen ASC (or similar*) used in finished-built products. It stays put, never 'dries out', and can be used well above 100c.

*ASC has a very unique non-Newtonian 'flow' and cleanup. Which, I've seen on/in non-PC related products, typ. on power ICs / FETs

 

[Shrek]

You should be able to hop onto Digi-key and/or Mouser and drill down some of the worst TIMs.

Only if you assume that thermal conductivity is the be all and end all.

• What if a low thermal conductivity paste ends up real thin... is this the best?
• What if a high thermal conductivity paste ends up thick... but resisists pump-out as a result... is this the best?

why is life so complicated?

*ASC has a very unique non-Newtonian 'flow' and cleanup.

Hadn't though of that... most interesting.

 

[R-T-B]

why is life so complicated?

The designers of this simulation clearly had to be german. It's way overengineered. (It's a joke, w1zzard )

 

[TPCEA]

https://koolingmonster.com/insights

Looks like a lot of useful articles! Thanks!

 

[Chrispy_]

BTW, after thinking on/off awhile. I have (experience with) the perfect "worst thermal paste".

A. It performs worse than most generic silicone-based pastes
B. It's been around for decades.
C. It's still manufactured and sold, today.
D. It's still 'useful', despite performing worse than even silicone-based thermal pads.

Arctic Silver - Céramique (2)

Arctic Silver Incorporated - Céramique 2

www.arcticsilver.com

Basically, it's Diaper Rash Cream but, made w/ a Synthetic Petrol base and (purposefully) containing nano zinc oxides and alumina (IIRC, lab testing determined it's mostly zinc oxide)

It's useful for MOSFET TIM, and I've seen ASC (or similar*) used in finished-built products. It stays put, never 'dries out', and can be used well above 100c.
View attachment 341385
*ASC has a very unique non-Newtonian 'flow' and cleanup. Which, I've seen on/in non-PC related products, typ. on power ICs / FETs

I never understood the point of Arctic Ceramic.

Ceramics are what are famously used as the best INSULATORS.

• Space shuttle orbital re-entry heat-shielding tiles.
• Insulation for smelting furnaces and other extreme applications north of 1500C
• Glass and ceramic high-temperature cookware for daily use in 220C ovens.
• Ceramic drinkware with handles so you can hold near-boiling hot water without burning your hands.

You don't even need a scientific education to know half of these things. Ceramics are f*cking awful at moving heat, and have been widely used as heat insulators from before humans even had the ability to record their learned knowledge. It's up to archaeologists to dig up 25000-year-old ceramic pottery and cookware and tell the stories that early mankind didn't have the record-keeping skills to archive.

 

[Shrek]

Ever grabbed a hot coffee mug?

 

[freeagent]

I thought Ceramique was alright on lower wattage CPU's.

 

[TPCEA]

I thought Ceramique was alright on lower wattage CPU's.

Sounds to me like it's garbage, then, if it can only handle cool CPUs... and coffee.

 

[freeagent]

Sounds to me like it's garbage, then, if it can only handle cool CPUs... and coffee.

Try it for yourself.. you might be impressed. Low wattage to me is 6 and 8 cores, but overclocked of course.

 

[LabRat 891]

I never understood the point of Arctic Ceramic.

Ceramics are what are famously used as the best INSULATORS.

• Space shuttle orbital re-entry heat-shielding tiles.
• Insulation for smelting furnaces and other extreme applications north of 1500C
• Glass and ceramic high-temperature cookware for daily use in 220C ovens.
• Ceramic drinkware with handles so you can hold near-boiling hot water without burning your hands.

You don't even need a scientific education to know half of these things. Ceramics are f*cking awful at moving heat, and have been widely used as heat insulators from before humans even had the ability to record their learned knowledge. It's up to archaeologists to dig up 25000-year-old ceramic pottery and cookware and tell the stories that early mankind didn't have the record-keeping skills to archive.

Correct.
However, the reason it's useful still, is somewhat buried in the marketing.

With a thin enough bond line, an insulating material is still considerably better than air (or vacuum).
Also, there are some oddball/old power components with metallic, (hot or gnd) casements. The extremely high breakdown voltage (and thermal decomposition threshold) of ASCeramique allows for use in HV and High-temp devices. It may* even 'better endure' ozone and nitric acid exposure from corona.

Try it for yourself.. you might be impressed. Low wattage to me is 6 and 8 cores, but overclocked of course.

It's cheap, "Non-Migrating", effectively inert*, and lasts nearly indefinitely. Agreed, seems useful still.
TBQH, I'd be interested in testing it on something known for 'pumpout' like Vega 10, and see how it does. With a thin bond line, and its propensity to 'stay put', I'm wondering if it has new practical use, today.
(not max. perf. but, max. endurance)
*Extreme Acids or Bases will still break down the carrier oil.

 

[Athlonite]

4) Old separated TIM in a tube like AS5 left from 200X in a garage.

I've never had that problem with any of the dozens of tubes of AS5 I've used over the years

 

[Shrek]

One reason I prefer tubs; if things do seperate I can stir before use.

 

[Chrispy_]

Also, there are some oddball/old power components with metallic, (hot or gnd) casements.

Well yeah, I could excuse it if that's the use case it was sold for, but nope:

 

[LabRat 891]

Well yeah, I could excuse it if that's the use case it was sold for, but nope:

View attachment 341515

If 100c+ sustained operation is in your design, it's 'not wrong'...
ASC'd be the TIM I'd choose if I was making a CPU-powered steamer/rice cooker

They used to advertise it as a Thermal Pad replacement, too.
Tried that, ONCE. I will say though, I was impressed how it stayed put even over many years. TBQH, Might be usable w/ appropriately-thick Cu/Ag thermal shims, for GDDR6-> today.
Put it on a China-Import Generic AGP 9600XT's RAM. It did the opposite of help its stability. This is also where-what-how I became so familiar w/ ASC's other unique properties.
Oh Well

One reason I prefer tubs; if things do seperate I can stir before use.

Protip: Bent Paperclip + Cordless Drill, insert into tube, and mix.

A. Remove the plunger; mixes easily, hard to remove air.
B. Carefully shape the paperclip so you can angle it into the tip of the applicator.

Either case, the paper clip acts as a mixing tine for the drill. Get the geometry+motion right, you can mix most of what's in the tube (kinda like peanut spread, really)

 

[TPCEA]

Try it for yourself.. you might be impressed. Low wattage to me is 6 and 8 cores, but overclocked of course.

Um, I'm not really sure, so I'm just going to ask - is there a direct causation between the number of cores and the number of watts? I would assume so...Also, do the number of cores and threads directly increase the watts and temps in relation to how many there are, or is it a log, syn or something increase? I won't be trying Ceramique unless I can get a really good price as I've overspent of late. Gotta walk the line a while because I had to take off due to work injuries that Amazon doesn't care about.

 

[Chrispy_]

If 100c+ sustained operation is in your design, it's 'not wrong'...

They used to advertise it as a Thermal Pad replacement, too.
Tried that, ONCE. I will say though, I was impressed how it stayed put even over many years.

That's why I'm still using my 25-year-old tub of Servisol, it's fairly viscous silica and zinc oxide grease with no added BS, so it stays put.

Crc-europe datasheets

Looking at the ingredients, I guess you could technically classify it as ceramic too.

Silica is obviously a ceramic, but I think it's hard to classify zinc oxide (which is the majority by weight) as anything else either. The only other significant ingredient is the polymer oil that lets it move and flow a bit.

So yeah, my rant about ceramics was poorly considered. I've been using this tub of Servisol for more than half my life and it's been indistinguishable from just about every other thermal compound that's ever been included with a CPU cooler!

 

[TPCEA]

Silica is obviously a ceramic

I think you mean silica ceramic, which contains silica, right?

 

[Shrek]

It almost doesn't matter which CPU thermal paste you buy (msn.com)

 

[Bobaganoosh]

Worst I've ever tested is Corsair XTM50. I bought it at a local Best Buy and I'm guessing it must have been sitting on the shelf for years if it performed that poorly. I don't know how anybody could buy that and consider it usable. I'd honestly rather use Arctic Silver 5 or whatever cheap stuff they put on a stock cooler. I needed something that day to do some testing with and it seemed almost as bad as not using any paste at all. At least 10°C higher temps than using PK-3 nano or KPx on a ~250W 12900k heatload (a non-AVX stress-test in EVGA's BIOS).

 

[Chrispy_]

I think you mean silica ceramic, which contains silica, right?

Nope, just silica, aka high-quality sand, and presumably powdered incredibly finely.

It's in the ingredients list I linked and 5% by weight...

 

[TPCEA]

Insights — Kooling Monster

koolingmonster.com

I read several articles and, aside from the revelation of .2ml on a 30x30mm CPU, unfortunately it was geared primarily towards low tech people. Naturally, every article suggested their cooling product, which is fine, although the article are mostly for marketing.

I found the article on fans to be disappointing. To make the article bigger (I guess), they included GPU and PSU fans, which most people don't need to know about, didn't provide enough guidance on choosing fans (e.g. intake vs exhaust; AF vs SP), and didn't even mention static pressure. I shall have to try the TIM calculator to see what the ideal volume for my CPU is, though.

I wonder if anyone has tried KOLD-01?