Water cooling kit for RPGA989?

got another reply from Jim i guess it doesn't have to sit that long or something...still figuring out the details. thermal glue is like what 4Wmk this is 87Wmk huge difference

OK so the indium TIM/cold solder doesn't actually bond it I guess. Misunderstood that. Talking to him now about some options on how to apply this and some different methods. Also I am waiting to hear back from him because I may also try using the indium as a TIM for the CPU. I think from what I understand so far that it is a huge improvement of regular thermal compound. Waiting to hear back and then I'll get back to you guys.

 

it might actually work.
TIM's are used as a compromise because you can't solder heatsinks to chips. you would fry the chip. imagine a cpu @500ºC.
But how much of improvement over TIM? is it really worth it?
I hope your experiments can give us an anwser.

 

What type of temps ate you expecting with this setup? If the water cooling is disappating like it should, I don't think it's going to matter what you use.. Hate to see you go through the extra time and money and see little gain.

It is however quite interesting what you're proposing though. I've seen people discussing attempting to solder a heatsink to their CPU, but they ran into trouble putting the two together quickly enough, and making sure to not damage the chip. Once they're together, it wouldve been a pain to make any adjustments..

I look forward to more updates!

 

don't tell him that :X
we need someone to do the experiment for us. lol
maybe it will work. we will never known until someone tries.

 

man today i turned off my laptop cooler to make a phone call and forgot to turn it back on and for 1 whole hour i was at 105gpu and 95 cpu...i was about to jump out the window in pain. So pissed off at myself that i forgot....i am amazed it didn't break. I bet i lost liek 6 months of life from it sigh.


Over the weekend i will give a massive update to everything i have learned about indium and what i will try to do. It appears to have some great potential. I will elaborate over the weekend so stay tuned.

 

I find this thread really interesting... Keep going

That's not true. Check out Koolance's products - as far as I remember they produced high quality QR, but they were quite expansive.

" I find your lack of faith disturbing "

 

you are tottaly right.
Koolance "No Leak" Quick Disconnect Demonstration - YouTube
really impressive.
they seam a bit bulky. hard to fit them on the bottom. but really great stuff.

 

I was just looking at the koolance stuff ^^^

What kind of gph are you needing from a the water pump you'll be using? You've got me intrigued.. I've been dismantling my laptop to see if I could contain a cooling system in the laptop itself. Of course id only have to cool a cpu. I'm going to try to secure a second cooling setup and I found another bottom half of my shell that I may end up mutilating. I can envision everything but the pump placement. Idk what gallon per hour rating is required for such a small system..

If I yank the disc drive, I could stick a reservoir in there I think, but I'd have to figure out how to cool it...

 

honestly to do an internal watercooling system would have to be some very nice custom made parts...that would be ridiculously hard. I would harass dell to buy out that water cooling system that those guys made for alienware laptops. it would give dell some serious edge with cooling.

I have a phone call with that guy tomorrow so i got to finish reading up and pulling together questions on indium. If indium TIM works as good as it looks would you guys buy some off of me if i bought a small kit? It would cost about 5 bucks a TIM(per use) if i sold it individually. From what I have read and been told its pretty awesome but i am waiting to try it. Jim was saying that is what is used many times in high performance servers. I am hopeing if the indium TIM works really well its one of those things you will never have to replace because it works so well. hence why 5 bucks a pop would be worth it. it is like 10 times more thermally conductive then TX-4

I am waiting for gentech to get back to me if ASUS has any extra heatsinks for my laptop. I rather buy from gentech then some random site.

 

the asetek watercooling notebook system is a prototype and as other prototypes they made they will never ever sell it and it's just vaporware. believe me I've tried hard to obtain one of those.

to make a full system fit inside a notebook it would only work in very large ones. it would make it even more heavier. it would become a transportable and not a desktop replacement.. battery life would be measured in seconds, not minutes or hours.
you do need some internal space. but I'm almost 100% sure it can be done rather easily if the notebook is big enough. like a 18" one.
you might have to remove the ODD and maybe some HDD's too.
fitting a reservoir and a pump is the easy part. there are small ones that would fit.

the BIGGEST problem is how to cool the system. because you do need a radiator. and radiators must be big in order to cool the water. that's the hard part. you need a large surface area for a radiator + fan. the only way to fit those would be on the bottom.
it had to be a custom made radiator. it could eventually fit the HDD bay. where 2 2.5" HDD stack on each other.
you could also remove the battery and use that space.
I don't see other way to do it.

 

i think the best way would to make not heatpipes but water pipes. Take the current laptop heating system and turn them into water pipes with fans. That would require little modification and give you great cooling because you wouldn't loose the fans. Going fan less is too hard.

 

the mod I described didn't required to remove the fans. you just add a waterblock to the chips keeping the heatpipes and fans.
then you would route the tubing to the pump and reservoir.
going fanless isn't hard IF the system has a good waterflow and the radiator is enough to cool the system, which I doubt.

actually I have many doubts about the asetek system. sure it overclocks well BUT the notebook already overclocks greatly on stock heatsinks without any mod.
It puzzles me how can they cool the water only using very small heatsinks.
they use waterblocks with copper fins as radiators. that's not very effecient.
actually they are just using the same stock cooling system. they just replace the heatpipes with water and added a pump.

 

wasn't it fanless?

 

yes it appeared to be fanless. that's why I doubt that the stock copper fins from the cooler are enough to cool the whole system. it doesn't make much sense.

 

thats why i say you should keep tjhje current design of the cooling assembly but make the heat pipes water pipes and add a pump and reservoir..that would be epic. Would require little modification.

 

and how are you planing to cool the water? magic? you need a radiator.
those very small heatsinks on the sides of the notebook won't do it.
that's why I think that asetek product is bs.
actually they are claming only -15ºC cooler then without any modification.
well, I get less 15ºC just with a notebook cooler.

it's like lenses and cameras. you can have a 50mp camera phone that it won't take better pictures then a 4mp dslr. good pictures need good lenses, good lenses need space.

you can get the best pump, waterblocks, tubing, reservoir etc etc. it doesn't matter. a good watercooling rig needs a good radiator. a good radiator needs space.

if you add the radiator externally then your problem is solved since a lot of space is avalable. But if you are going to do it, why do you need a reservoir and pump inside the notebook? because you want to use it while on battery? but if you do that then how do you cool the water?
got it?

 

it would work just fine. The point of water cooling is it pulls heat off better than heat pipes. so even those heatsinks with fans would work better with water pipes than heat pipes.

 

I've my fingers crossed for that. I really hope it works.
I've my doubts but I think you are right. Because most cheap watercooling kits already beat the high end massive cpu coolers full of heatpipies and loud fans.
even a small 80mm radiator would do. but it isn't much of an improvement. 5-10C which is interestingly consistent with the asetek results.
Corsair Hydro H60 CPU Water Cooler Review - The Temperature Testing Results - Legit Reviews
if you increase the waterflow then things can be a lot different. also if you add copper pipes instead of plastic tubing then it will cool a lot more too.
good luck for your soldering skills! you'll need it.

 

It was proven many times on LC related boards, that increasing waterflow does not make a big improvment. Or any at all, in most of tests. The only thing that actually matter is the size of radiator.

I *really* doubt that water is better heat conductor than copper/heat pipe or even aluminium...

As it was explained here:

you can't just beat physics - water cooling, to be effective, needs radiator. The bigger radiator, the lower the temps will be. If you don't use radiator, the water will not get cooled down, and the whole point of LC will die.

 

look...water carries tons of heat...the point of what i am talking agbout that custom made water cooling system for the alienware. You normal have heatpipes and heatsink and fan. you turn it into. COPPER water pipes and heatsinks and fan. that will work much better than a regular heatpipe. I am not talking about in that alienware system using plastic pipes. That would be pointless...also dangerous. Also increasing water floor helps some but not by a lot. it also depends on how how it is. if your running at 80C increasing water flow will help but what really help is increased radiator or fan speeds. The water cooling for an alienware computer with copper water pipes would work very well. if you also increase tha fan speed it will work even better. Also that asetek or whatever system connected both heatsinks. With water cooling and connecting both heatsinks that would make a world of a difference


EDIT: shame search this forum and found this but the video is removed. Wonder what he was trying to do :/

I got my cpu heatsink and heatsink in the mail today. Anyone know of a way to see if it has gas in it or a leak? Also I plan on buying some parts over the weekend to start building my laptop cooler. I'll update more when i find all the parts i think i'll need.

 

don't expect that much of a difference. even asetek itself says they only got less 15ºC. so if they say 15ºC it will be less then that.

agreed that waterflow is not the biggest issue. not even the waterblocks. but sure it must have a certain flow otherwise performance will be worse then air.

well me too.
but on the other hand a heatpipe as a very slow liquid flow unlike a watercooling system.

but anyway I think this review proves it.
Thermaltake Tide Water Review | techPowerUp
the radiator is very small. imo a good air cooler is better then that.


I guess the best way is to have an external pump, reservoir and radiator. After all those components inside the notebook would have some disavantages.

 

15 c though is a very good difference for that set up
i wouldn't have expected much more then that

 

It doesn't work that way. Water hold more meat in it than a heatpipe, but a heatpipe conducts heat a million times better than just water. If it worked that way, why on earth would we have have heatpipes?

 

you didn't read what i said.

 

I did. Water is not better than a heapipe for transferring heat like that. You think that having water transferring heat to the tiny heatsink only inches away would be better, but it would be worse. In this application, heatpipes are the best thing.

Just attach a real water block to your cpu and have a reservoir, pump, tubing, fans, and radiators outside the laptop.

 

i was talking about the alienware asetek design is better then a heatsink design...you really are not listening

 

that's what they claim and what I've many doubts about. Ever wonder why they never released the product? I bet it would be many buyers for overpriced watercooled alienware notebooks.

those small radiators really don't convience me. if there's really an improvement, it's just a small one.

if you had an external radiator you'll have a massive cooling improve.

But I guess we'll never know for sure until someone tries it.

 

Hows the progress of this project. I've yet to see pictures.
I dont see what the problem is unless he's being serious about it being integrated into the system, in which case with the space given watercooling will be far less inefficient and given the TDP loads would be a waste imho.

Heatpipes are far more efficient than watercooling by far when confined to a small space, where TDP is limited.

In larger applications watercooling wins by a landslide not in efficiency but capacity.

Without getting ridiculous, heat pipes are difficult to work with in large applications (or high performance systems for example).

#1 Heatpipes have far lower limits in how much heat can be stored when the heatpipes cannot handle the dissipation of the heat (and trust me some high performance desktops can surpass the heat dissipation limits quite easily).
This means after a long period of times they become saturated and you get massive heat spikes and eventually overheating. Since heatpipes must make direct contact with the source of the heat, when saturation occurs heat transfer stops or slow downs to a crawl and the heat source gets hotter.In watercooling, the resorvoir can be placed farther from the heat source and thus operate at higher efficiency than heatpipes (afterall at this point heat pipes are not working at all).
#2 Heatpipes are highly dependent on the proximity of heatfins, meaning they must be close and dense. The problem rises in that the heat exchange between heatpipes and heatfins gets limited by the amount of heat that can be carried at any one point in a heatpipe.

Water cooling on the other hand can hold vastly more energy in the water compared to heatpipes. Meaning it takes far longer to saturate, and the capacity of heat stored is only limited by the reservoir of water which can be added at a good amount higher of "heat storage" per cubic inch of space compared to a heat pipe (1 cubic inch of water holds more heat than 1 cubic inch of heatpipe, however water is less efficient at thermal transfer).

 

well i am waiting to see if the "heatsink/cooling assembly" I order is the right one. Second I have a set plan on how to do this to properly document it.

This is how it will go
1st month is ICD
2nd month is "Heat Spring" TIM
3rd month is adding water cooling using thermal glue
4th month is adding water cooling with solder
I think there might have been one or two different test I wanted to do. I am testing each one for a month to get a long term test to compare results.


I will be logging ambient temps, humidity, load, gpu/cpu/case temps, and anything else I think is worth while. I'll export all the numbers into charts for comparison so you can see exactly how each worked.

Please note I will log idle temps probably once a week to see how those work and compare. Also I will start off with ICD and max safe OC. Once I use Heat Spring I will log 1 week of those temps and move onto higher OC/overvolt to see that upper limit. (assuming it works better)

This will show how much better it works with same power usage and how much higher clocks you can reach with it. Also I will note ~ power usage at various times. (note not a 100% accurate reading since it is from my UPS and has monitor, fans, and laptop running off of it but I will try to keep it as close as possible to that so you can get a general idea.)

Also I have a couple ideas on how to make my laptop cooler/watercooler case. I need to make that first so tests are consistent. I hope to start this by August 1st. I am currently trying to sell my R4 for testing funds.

I would life to test this with indium solder but I have to use ENIG processes on the copper pieces but I don't know of a place that does that and if I can get in on a run. I obviously can't afford a full run that is 100-1000 dollars.

Also I will try a separate test while doing these in between. A large amount of my heat is removed by the laptop cooler that is on crack. I will try a test without the laptop cooler on and see how each TIM works. This is to give an idea on how these TIMs work in a "pure" laptop environment.

EDIT: This will take anywhere from 6-8 months because I want to be thorough and not miss anything and have any accusations of being inaccurate. Also I wwill only test the CPU because my GPU is acting finicky and I dont want to chance it. It started having artifacts while i was using it and have to reboot computer so I no longer fold/stress test cooling with it.

I will post progress as I have info on any updates in the main post. I will finish all of this and add it to my ultimate cooling mod for the G51j when i am done. I will start to write this up in a word file with all the charter indexes and stuff for an easy to transfer and navigate source.

 

i got my u3 or whatever in the mail a week or two ago..haven't gotten time yet to build it into a box that'll fit a water cooler in it yet....still need to sell m17x R4 before i can buy water cooling and heat spring TIMs but it'll be a whole month just testing on air and ICD I think not this weekend but next I'll build it. I also still need to buy external temperature monitoring device. I got to look at an old thread to go buy that but I'll look this weekend for that

 

Look into a custom block. Something micro pin low profile VGA and get it made in silver. Plumb it with 1/4" copper pipes back to where your heatsink was then hook some QD connections to it and use a DDC+ 18W with a res top and a 240mm radiator. DDC+ will have the head you need for a restrictive flow block to keep your flow rate good.

Have you thought about useing a Peltier? Then use the heatsink to cool the hot side and the cold side to cool the cpu? Hook up a thermo circut to it to control the power to the Peltier so you dont get the cpu all frosty.

 

peltier cooler would be tough to keep good contact. I did some measuring and the holes for RPGA989 are close to some standard water blocks so whenever i buy one i will see if i can jerry rig it to the case. Not sure i was doing off some dimensions in the details and it seems fairly close but i'll have to wait and see.

I am more interested in trying Heat-Spring TIMs first with regular heatpipe cooling.