[Review] Round 2: Heatkiller 3.0 LT and LC Tested + How To Improve the Heatkillers!

Preface
This is a quick version of my Heatkiller 3.0 LC and LT Testing....it's part two of my Roundup #2 of waterblock testing. I have some cool things in store for testing (including ways to significantly improve the performance of some blocks) and will be testing a lot of blocks. Anyway, I'll cut the boring parts out of the test logs and post up the Heatkiller results in this thread :)

If you have any questions, check out the full review (link above)...if they're not answered, ask away here :)

Test Results

First up, finding the best orientation for each block:
http://vapor.skinneelabs.com/i7/Roun...rientation.png
Here we can both the LC and LT prefering the Vertical orientation (barbs parallel to the socket latch) and the LT benefiting from having the plastic divider at the inlet removd (not the metal plate! Do not remove that!). Overall, the LT being roughly 1C better than the LC at this pumping power.


Improving the Heatkillers!

In my observation of the blocks, I actually noticed a design flaw. The base, while it sits flush with the internals when you just place one on the other, does not sit flush with the internals once you tighten it down and form the bow. Basically, when you force the base to bend outward at the center, you're causing a gap between the injectors and the base. That means flow 'escapes' over the microchannels. While that lowers restriciton, it's also less efficient thermally. I decided to close up that gap with inexpensive, completely waterproof, removable, and easy-to-use 100% silicone caulk.

http://vapor.skinneelabs.com/i7/Roun...r/Silicone.jpg

It doesn't have to be perfect, but using Q-Tips and silicone caulk, I formed a compresible and resilient 1mm layer on top of the impingement plate. That's what it looked like before I installed it. What's really interesting is what the performance looks like once it's installed! From here on out, I'll call it the "LT+" :D

http://vapor.skinneelabs.com/i7/Roun...timization.png

I've included performance from three different baselines: 1) the best-as-tested LC config, 2) the best-as-tested LT config, 3) the LT's base with the LC's top (Frankenblock!). The showstopper is just how much performance improved from adding that little bit of silicone, that's the "LT+" on the chart. We took what is largely considered the best block on the market (and the best I've tested so far), and made it better, noticeably better. Flowrate went down, as predicted, but not by much--from 1.77GPM to 1.55GPM (more flow than the Supreme LT at the same pumping power).

Now that we know which orientations are best for each block, let's run the full flowrate spectrum and see how they respond to flowrate.

• Very High Pumping Power: All three MCP355 pumps and the D5 are on at full speed--this has a very similar PQ curve to a pair of RD-30s at 20V.
• High Pumping Power: Two MCP355s with EK V2 tops are on at full speed. The other two pumps are off.
• Medium High Pumping Power: A single MCP355 with XSPC V3 top is on at full speed. The other three pumps are off.
• Medium Pumping Power: The stock D5 is on at full speed and setting 5. The other three pumps are off.
• Low Pumping Power: A single MCP355 with XSPC V3 top is on at minimum speed (~7.7V, ~2450RPM). The other three pumps are off.
• Very Low Pumping Power: The stock D5 is on at minimum speed--setting 1. The other three pumps are off.

http://vapor.skinneelabs.com/i7/Roun...tkiller25W.png

http://vapor.skinneelabs.com/i7/Roun.../HKLCTable.png

http://vapor.skinneelabs.com/i7/Roun.../HKLTTable.png

http://vapor.skinneelabs.com/i7/Roun...HKLT+Table.png
Note: I do 5 mounts at "Medium High" then take the best config of a block and test the whole flow spectrum (after a TIM curing session) then realign that curve with average of the 3 median mounts to give you the "Adjusted" data.


Many More Graphs
I've included the core graphs and data here...but there's more in the full review, which is here. Conclusion + thoughts are also in there...I suggest checking it out if you want more info, thanks for reading! :yepp:

Working on the overall comparison review/charts...it'll be updated with every installment I do :D

(also almost done with Apogee GTZ and GTZ SE testing)

Quote:

---

Originally Posted by Vapor

In my observation of the blocks, I actually noticed a design flaw.

---

:O Lies..

German engineering has no word for design flaw..

:rofl:

Great review eric.

Quote:

---

Originally Posted by NaeKuh

:O Lies..

German engineering has no word for design flaw..

:rofl:

Great review eric.

---

:ROTF:

That is pretty interesting that you got over a degree of improvement @ around 1 GPM by filling that void.

Great review. :up:

Great review! I have a question though, wouldn't having enough pressure from the mounting cause the CPU IHS to push back against the Heatkiller bow do a similer (if not the same) job at closing the gap with the silicone caulk? I believe with a backplate it is quite safe to apply over the 200N pressure suggested in its mounting instructions. (which I believe was assuming a backplate-less installation)

At what mounting pressure did you mount the HK at? Did a quick word search for "pressure" but turned out at the beginning which you mentioned will "present results from increased mounting pressure". Am I slowly getting blind reading all the nVidia Vs. ATi rubbish as I couldn't find the results anywhere in the full review... :(

I tested a few mounts at full tightening, it hurt results a little (~.1C, nothing significant). 200N is what I had all published mounts at (according to the Heatkiller's spec/install sheet).

I really only plan on fully investigating increased mounting pressure in scenarios when there's a clear lack of it (D-Tek 1366's Pro-Mount, for instance).

While I hate the Heatkiller's mounting mechanism, it is fully functional and they have noted what the ideal pressure is.

Ok, first, great review as usual. Now some questions:

1 - Does the parallel configuration for the mounting apply to the regular Heatkiller?

2 - Do any of these improvements apply to the all copper Heatkiller?

Thanks

Quote:

---

Originally Posted by Utnorris

Ok, first, great review as usual. Now some questions:

1 - Does the parallel configuration for the mounting apply to the regular Heatkiller?

2 - Do any of these improvements apply to the all copper Heatkiller?

Thanks

---

1) Eh? It seems to evenly apply to both that I tested...I'm sure it does to the Copper as well (it's due to the orientation of the bow/microchannels).

2) Yes, I'm sure these improvements apply to the Copper version as well.

Thanks, time to fiddle.

Sealing it up... reminds me of the Fuzion V1 washer mod, only a bit more involved with the silicone. :D

Excellent work and presentation once again! :toast:

So I realize there may be some concern with the silicone mod...but as long as you use 100% silicone, you should be okay. It's insanely tough and totally impervious to water.

Avoid products that say "Easy water clean up" or "Acrylic caulk" or anything like that. Look for "100% silicone" on the packaging of the caulk.

Nice review! Maybe they will release a second edition with a rubber sealed plate, and while there at it make a better mounting system, or at least use real "thumbscrews". Also looks like I could have gotten another ~1C by using my second pump since I have one big loop with dual gpu full cover block, even subtracting the ~.3C or so from additional 18W.

Excellent review there mate, much appreciated!:clap:

You would be better off with an o-ring around the inlets me thinks... I'm glad other people are starting to think again. I just haven't had the time. And vapor, not only have you sealed the water flow but most definitely you have increase the bow as well, same concept as the Fuzion.

Yeah, it does visually change the bow...it broadens it and probably strengthens the bow as well. The temperatures the LT+ was getting were so sweet to see, that's for sure.

For anybody wanting to do the mod, Q-Tips are your best friend, don't be afraid to use a lot of them as well (they're cheap). They're great for moving the silicone and also have a great straight edge for making a smooth plane (though I didn't do too well on that :lol: ).

Vapor,

Firstly, great job! Clearly a lot of work goes into these tests - very impressive work :clap:

Looking at the graphs, I am trying to guage what kind of result I should be getting vs. what I am actually getting. So, are the "Core Temperatures Corrected for 22 Deg C Air Intake and Typical Radiator Ability (Pump Heatdump Included)" curves the ones intended to be representative of what my own rig might achieve? Sorry if this is a dumb question :shrug:

Also, if I said that with my rig (as per my sig: Core i7 D0 at 4.2 GHz, 1.30625v Vcore - shows as 1.280v in Cpu-Z - at 22 Deg C ambient) I get a max core temp of 62 Deg C (measured with Real Temp v3.0 running LinX 0.6.3 and All Mem of ~4.6GB) with the other cores 1 to 2 Deg C lower, would that be considered to be good or not? I realise that your testing is with a C0 at much higher Vcore.

Basically, all I really want to know is if I have a "good mount" as I see from your tests that over 5 mounts it can vary by almost 1 Deg C. I have only done one!

P.S. I am using Shin-Etsu X23-7783D TIM after reading your review :D

That`s a great job :up: Thank you Vapor

Quote:

---

Originally Posted by M.Fixit

Vapor,

Firstly, great job! Clearly a lot of work goes into these tests - very impressive work :clap:

Looking at the graphs, I am trying to guage what kind of result I should be getting vs. what I am actually getting. So, are the "Core Temperatures Corrected for 22 Deg C Air Intake and Typical Radiator Ability (Pump Heatdump Included)" curves the ones intended to be representative of what my own rig might achieve? Sorry if this is a dumb question :shrug:

Also, if I said that with my rig (as per my sig: Core i7 D0 at 4.2 GHz, 1.30625v Vcore - shows as 1.280v in Cpu-Z - at 22 Deg C ambient) I get a max core temp of 62 Deg C (measured with Real Temp v3.0 running LinX 0.6.3 and All Mem of ~4.6GB) with the other cores 1 to 2 Deg C lower, would that be considered to be good or not? I realise that your testing is with a C0 at much higher Vcore.

Basically, all I really want to know is if I have a "good mount" as I see from your tests that over 5 mounts it can vary by almost 1 Deg C. I have only done one!

P.S. I am using Shin-Etsu X23-7783D TIM after reading your review :D

---

Yup, that's the graph for you to use to find optimal pumping power (which you're at, according my testing).

As for temps, every CPU is different, some are VERY different, so I wouldn't try to compare your CPU to mine. Your temps do sound good though, if you want to remount a couple times and see if you're getting roughly the same temps, it might be worth a shot. Overall though, I think your temps are good, if you used sound mounting technique, I wouldn't worry. :)

very interesting "silicone caulk"

Wow. Well done on improving the best waterblock out there!

Those HK blocks are awesome! me want me want !! :slobber::slobber:

I wonder if it is possible to use Bitspower 1/2in. Compression Fittings for 3/4in. OD Tubing with the HKs ! :confused:

Quote:

---

Originally Posted by Vapor

As for temps, every CPU is different, some are VERY different, so I wouldn't try to compare your CPU to mine.

---

Thanks - I appreciate that CPUs do differ, and sometimes widely, but as I only have experience with one CPU and one W/c block your observations / opinions provide valuable insight for me :D

I expect to change my fans to push/pull at some stage, so I will have a go at doing a few mounts to see what I can achieve. Alternatively I may go straight to the Indigo Xtreme TIM and be done with it :up:

Quote:

---

Originally Posted by Salamndar

I wonder if it is possible to use Bitspower 1/2in. Compression Fittings for 3/4in. OD Tubing with the HKs ! :confused:

---

I am using 1/2" id / 3/4" od BP compression fittings with 45s on inlet and exhaust on my HK. On their own the compression fittings don't fit :down:.

Could you show us a pic. of your block?

" OK that sounded nasty !!" :p:

Quote:

---

Originally Posted by Salamndar

Could you show us a pic. of your block?

" OK that sounded nasty !!" :p:

---

Sorry, I don't have any pics of my rig just yet ;). However, I am sure a bit of googling will find you some pics of similar installs. I will try to dig out a link of one similar -but in the mean time try this: http://media.photobucket.com/image/h...0/DSC01737.jpg

I have a question and it was probably answered but I need it in lamens terms. I have a HK 3.0 LT and I have a reverse atx case( A+ Black Perl) and essentially mounted the block upside down the on motherboard. I hope this makes sense how I explained it. I am wondering if this would have any effect on temps. I can reall see how because it is still mounted vertically on the chip.

Anyone tested this improvement yet? Results? (Besides from Vapor)

Quote:

---

Originally Posted by Jonni Nitro

I have a question and it was probably answered but I need it in lamens terms. I have a HK 3.0 LT and I have a reverse atx case( A+ Black Perl) and essentially mounted the block upside down the on motherboard. I hope this makes sense how I explained it. I am wondering if this would have any effect on temps. I can reall see how because it is still mounted vertically on the chip.

---

Nah, your performace will be fine.. theres isnt some "huge" difference between how is it mounted or where..

A little less messy way of Vapor modding Heatkillers;)

If you have a used fan vibration gasket (as on picture for a 92mm fan) cut the side off
and tape it on a smooth surgace (I used a 2mm rubber pad). This ensures the entire
surface of the silicon pad is of the same hight ...about 1mm

Then put the silicon gel (I use the same as for sealing kitchen sinks, bathtubes
...100% neutral silicone) on the rubber pad and spread it with a triangle
ruller to get a
smooth surface. Put it away for a day or two for silicone to dry.
http://www.shrani.si/t/35/CV/3aGjw9C...rafija0316.jpg

The next step is to draw a template. Put it on a plain paper, and then cut it
out. The paper will stick to the silicone, so no need for any additional taping.
http://www.shrani.si/t/1N/in/2Y5vhTM...rafija0317.jpg

You can cut this out with a carpenter knife or other sharp tools:D

This is how it looks on a LC top. Also this can be easily removed when
you take the block apart.
http://www.shrani.si/t/1W/r7/1A8nEYm...rafija0318.jpg

Hope it helps you.

Awesome way of doing the mod there! :toast:

Just brainstorming with your idea:D:up:

Minor update on the HK silicone mod:

After letting it sit for a few days I took the block apart and found the
following:

- the silicon gasket is compressed to about 0.2mm (screws fully tightened)
- the 1st and the last channel is blocked by silicone ...this would result in no
or very little flow in these two chanells ...this might not occour in LT/Cu
block as it has very narrow channels (51 for LT/Cu vs. 34 for LC)
- at the ends of the top the silicone is very very compressed and might tear off

I then removed the silicone gasket and assembled the block as per default. With
screws fully tightened the top sits on the chanells ...I tried to insert a piece of
paper (~0.1mm thick) between the top and base and could not do it (not at the
center or at the sides ...the center of the top has a small bulge) .

NOTE: the above applies for the LC block, the silicone mod might affect the
LT/Cu block differently.

Great review!!! :clap: even got me to buy one :D

Quote:

---

Originally Posted by Bladesinger7x

Minor update on the HK silicone mod:

After letting it sit for a few days I took the block apart and found the
following:

- the silicon gasket is compressed to about 0.2mm (screws fully tightened)
- the 1st and the last channel is blocked by silicone ...this would result in no
or very little flow in these two chanells ...this might not occour in LT/Cu
block as it has very narrow channels (51 for LT/Cu vs. 34 for LC)
- at the ends of the top the silicone is very very compressed and might tear off

I then removed the silicone gasket and assembled the block as per default. With
screws fully tightened the top sits on the chanells ...I tried to insert a piece of
paper (~0.1mm thick) between the top and base and could not do it (not at the
center or at the sides ...the center of the top has a small bulge) .

NOTE: the above applies for the LC block, the silicone mod might affect the
LT/Cu block differently.

---

I was thinking the same thing, I'm going to try the opposite side with silicone. It makes sense since the problem to begin with is that o-ring being to small between the top block and diverter plate. I just siliconed the top block around and on top of the o-ring. Not sure how I can test this, but if it stops the dingling sound the diverter plate makes when I shake it I'll be happy.

Anyone test with a CU block? I remember when I got mine new, I could shake it and hear the plate chinging around. I dont feel very good about the silicone caulk, but a nice o ring or even a cured silicone gasket cut to shape would be nice.

i have the CU nickel plated. i thought about putting a bigger o-ring in there, but was afraid that it would be hard to keep the shape trying to fit it in a smaller groove. good idea a flat gasket might work. if it were possible a thicker diverter plate with original o-ring might be the best option.

Finally got around to doing the silicone mod to mine. What with my mobo being out on RMA, there's plenty of time to let the silicone cure before reinstalling. I wound up trying a slightly different approach, just applying the silicone to the top only (o-ring is underneath still). I made sure to uncover the impingement plate location pins and clean up the inlet with q-tips. The silicone is about as thick as the location pins are above the surface. Got it pretty smooth using a pocket screwdriver. I'll give a couple of days before trying to reassemble.

http://i48.photobucket.com/albums/f2...9/100_2505.jpg

holy batman necro!

oh, c'mon now. It's only 11 months old. :)

Quote:

---

Originally Posted by Rock&Roll

oh, c'mon now. It's only 11 months old. :)

---

Which means you dragged it up from page 114 = dead topic. ;)

The top 3-4 (maybe 5) pages I can understand, but not page 114.

To all would be Necros, If the topic is more than a month old, please let them rest in peace.

Hmm, most probably thread got found via search .. and is it really better to make new one instead of posting in prehistoric one?