microfin wb made from akasa ak360/370

[MadDogMe]

Wednesday morning. Sorry I did'nt PM a thank you straight off. I've had a wierd few days lately, out of sorts :shrug: . Nothing I won't bounce back out of though ..

[Blackeagle]

Ahhh, thanks MadDogMe. 4 working days to Less vs 6 to you, at half the cost.

And everbody has a bad day sometimes. Main thing is to have a good average!

[leejsmith]

Hi BE,

I can get you the ak350 and send it over if you like.

did you see my gfx cooler i made from the ak350 ? is that what you have in mind. The base of the ak350 is 5mm at it's thickest.



i havent mounted this yet as i need something on the back of the card to support it and it needs at least 2mm of the base.

On the thermal take base i took it down to 3mm and it made no differance on the temps. i will take it down to 2mm this week.

Lee

[MadDogMe]

How many of those bases could you get out of one ak350 HS Lee? I like the idea of the pins as well as it bieing a low restriction block. It'd be better than the job I'm doing for GPU and NB blocks at the mo'...

The size would be 25MM x 25MM pin area plus enough 'border' to seal the block (5~8MM usin M3bolts)...

[leejsmith]

the ak350 base is 60mm x 60mm on akasa.co.uk

if the gpu base could be kept to 30x30 then you could get 4 from one ak350 base but that wont leave much around the fin area.
if not then only one.

what if the base had a lip all the way around like a tong and groove. some thing else could make up the rest of the base.

i will do some 3d work later to show my idea.

Lee

[Blackeagle]

Hi Lee,

No, I hadn't seen that one. "Sigh", so much for making a variation of my own.

If you recall the post I put up on useing the Rotor type block, I had in mind to do much the same useing the AK350.

Would cut off some of the pins as you did for a seal area. Then use a small drill bit to drill dimples in to the base between the pins to increase the surface area. Then make up a Lexan top with angled inlet/outlets as I described for the Rotor type block.

So all in all not much differant at all from what you have there. Did you drill dimples between the pins on yours?

Question is would it match up well vs. the Rotor type?

How tall are the pins in the one you made? How tall were they at start?

Last how much do the original AK 350 coolers cost? I got the idea of doing it this way from a heatsink maker's site I checked out some time back. www.coolinnovations.com But I then found they won't sell to end users, minimum order of 500 pc, which was a couple more than I had in mind. LOL Then saw a review of the AK 350 which got me going again on the idea. But cost is a consideration indeed, as cutting up CPU coolers is bound to cost a good deal more than use of small heatsinks bought plain. Depending on cost the Rotor design might be a better way to go.

[leejsmith]

Originally posted by Blackeagle
Hi Lee,

No, I hadn't seen that one. "Sigh", so much for making a variation of my own.

If you recall the post I put up on useing the Rotor type block, I had in mind to do much the same useing the AK350.

Would cut off some of the pins as you did for a seal area. Then use a small drill bit to drill dimples in to the base between the pins to increase the surface area. Then make up a Lexan top with angled inlet/outlets as I described for the Rotor type block.

So all in all not much differant at all from what you have there. Did you drill dimples between the pins on yours?

no i didnt drill dimples on this one but i am going to convert it to 1/2" inlet / outlet so i could make some changes. I was thinking of moving the inlet over the pins.

Question is would it match up well vs. the Rotor type?

once the base is thinner and you have the dimples then is should work very well.

How tall are the pins in the one you made? How tall were they at start?
then pins are 8mm before i lowerd them down to 4mm.

Last how much do the original AK 350 coolers cost? I got the idea of doing it this way from a heatsink maker's site I checked out some time back. www.coolinnovations.com But I then found they won't sell to end users, minimum order of 500 pc, which was a couple more than I had in mind. LOL Then saw a review of the AK 350 which got me going again on the idea. But cost is a consideration indeed, as cutting up CPU coolers is bound to cost a good deal more than use of small heatsinks bought plain. Depending on cost the Rotor design might be a better way to go.

follow this link http://www.theoverclockingstore.co.u...?codeid=812438

they are £14.69. i will send it to you.
my dad thinks there is a scrap metal yard close to me. I get my heater cores from next to them i think so i will see if i can get some copper from them.

[leejsmith]

i now have the thermaltake base down to 2mm and temps are roughly the same.

It does take longer to get to it's highest temp than before but still gets to the same temp within the test period.

next 1.5mm

[Blackeagle]

Lee,

Thanks man, I look forward to it's arrival.

Has your item arrived yet? If not it shouldn't be much longer.

I'm surprised that the reduction in your bases thickness has so far made no differance at all. Still anything learned is good. Even if it isn't what one expects.

[leejsmith]

Originally posted by Blackeagle
Lee,

Thanks man, I look forward to it's arrival.

Has your item arrived yet? If not it shouldn't be much longer.

I'm surprised that the reduction in your bases thickness has so far made no differance at all. Still anything learned is good. Even if it isn't what one expects.

i dont have it yet but i guess early next week.

i will order the ak350 monday and send it as soon as.

i got fedup with the temp readings from the cheapo p4 mobo i use for testing. it only gives temp readings within .5C so i butcherd a temp probe i had removing the actual probe from it's protective covering. Using some very fine wire i etended the probe and superglued it to the underside of the p4.
As the wire is so small i was able to run the wires through the pins of the cpu. I also put a piece of wire on the oposite corner of the cpu to balance it.

i know it's not actual die temps but it's giving me a better indication of what temp the cpu is with the differant blocks.

i ran some tests again but as i dont have a 4mm thermaltake base to compare i used the maze 3 again.

the 2mm thermaltake base gives .5C better temps on the temp probe than the maze 3.

i will take it down to 1.5mm and try again but i cant see it improving that much maybee another .5C at the most.

i need to make a top the the akasa base i took down to 2.5mm with 1mm channels cut into the base and test this with the temp probe on the cpu.

Could it be the height of the fins or water not getting to the base of the fins. IMO the skivved fins on the thermaltake base maybe stopping water getting to the base.

[leejsmith]

so i took the base down to 1.5mm today.

i did the same test run i used on all other tests but this time i had the probe under the cpu.

first maze 3.

idle

water 24.7
cpu from system 31.5
probe 25.7

then at load

water 26.1
cpu from system 35.5
probe 32.6

now the thermaltake with 1.5mm base

idle
water 24.6
cpu from system 31.5
probe 25.4

at load

water 25.3
cpu from system 34.5
probe 31.3

should i take thebase down to 1mm what do you guys think ?

i do have to say testing these blocks is so frustrating. even more so as the wether is very warm in the uk at the moment and most of my previous tests the room temp was 20-21 at the moment it's 24-25.
trying to keep the room at a stable temp throuout the tests is a pain in the backside.
Lee

[MadDogMe]

I would try it if you think there's enought support on the baseplate to prevent it bowing under pressure...

God man don't knock the weather!!! , remember the rain!! , ...

[Blackeagle]

Glad to see you have some positive results Lee. 1C is a good start.

Don't think I'd go to much farther in thinning it down. Afraid it will flex on you if you do.

[leejsmith]

thanks BE,

i orderd the pin base today.

i dont have anything thin enough to cut between the fins.
unless i go thinner on the base what else can i do ?

i could look at the inlet to maximise flow rate while keeping the jet to spread water over the whole base.

[bigben2k]

To 1.0mm or not to 1.0mm, that is the question!

From your previous results, it seems like you've reached the far end of the efficiency range, with 1.5 mm. I believe that taking it down to 1.0 would be detrimental. If you do it, it would be experimental.

[MadDogMe]

Personaly I don't think you're getting any impingment yet, I think you'd see more than 1DegC with active impingment. Then the 1mm BP might yeild more :shrug: . I'd try making the 'slot' thinner, 1mm at least considering the width of a P4 core/HS'der...

[leejsmith]

I am going to make a couple of tops to go with the 1.5mm base.

first keeping the squashed inlet but one with 1mm and one with 3mm jet at the end.

then i will make a 2 piece top very simular to the white water. the quashed inlet will increase velocity but it's a smooth transition. with the jet plate it will create turbulance as it hits the smaller jet hole before it enters the base fins.

I am not too happy with the copper tubes as inlets. They work very well but the araldite instant i am using goes a bit soft where it's in contact with the water. I might cover the areas after it has set with silicon to protect it.

[leejsmith]

so i made the tops and found that 2mm and 3mm squasjed tubes gave the same results.
1mm squashed tube gave put .5C back i think thats due to this inlet taking 65 seconds to fill a 5l container compared to 30 on all of the others when i did a flow test.

and last a 3mm jet plate based on the WW jetplate and this took another .5 of my best results so far.

looking at the jests in action you can actualy see turbulance in the ww style jet where as with the squashed tubes you dont.

i am going to experiment with the jet plate and make another top.

After some tips from cathar i might reduce the width of the finned area so it's not covering the whole p4 heat spreade but still keeping the length so instead of 30x30 i will try 30x20. This should make the jet plate more effective.

anyway my mind doing overtime at the moment on cathars v2 tube cup block. taking 2C of WW is awsome drool:

[MadDogMe]

I'm surprised it did'nt improve it!, but then I'm only guessing without seeing the componants. I'd try them all on the pump and see which one 'shoots' water the farthest. That'll be the one with the most velocity. Whether it's the same submerged is beyond me though :shrug: ...

I agree!, Cathars latest block is very interesting!. Looks like one a competant homebrewer could make as well with basic tools , it's the one I'd be concentrating on ...

I don't know why I never thought of that before with the P4 h'spreader. It's something I've done myself with GPU chips, ignore the plastic 'package' and concentrate on the 10x10mm core in the middle. I've never put much thought into a Pentium core though. AMD all the way! (can't/would'nt afford a P4 )...

[leejsmith]

the ww style jet plate improved the temps buy .5C so thats 1.5C better than the maze 3. Maybee with a tweeked jetplate and or the smaller finned area it will improve.

WRT the tube cup block.

I have the copper tubes allready and i am working on a plan at the moment.
It wont be as small as cathars my drill press is pants. i have 2mm OD copper tube and was thinking of 3.5mm cups with 1mm between each cup but i will need a very good drill to be so close on the base plate.

From reading here i will need very high rpm with small drill bits on copper.

my other problem is getting copper for the base i do have a second thermaltake base that i could use but it's getting expensive to keep using a £20 cooler to experiment with.

i dont expect to get to the same performance but to be even close to WW would be wicked.

[leejsmith]

jet test 1

with no template this is not exact but lets me know it works


i drilled 2mm holes into the acrylic and superglued 2mm brass tubes into each hole.

then cut all the tubes to the same length and made sure there was nothing blocking them.

after cutting a chamber into the top and also making a little funnel shape at the top of each hole you can see it works very well



next to make a template and make sure it's spaced correctly.

i need to do some practice runs drilling the holes close together into a copper base.

[Pritorian]

Wow!

Looking really good!

-P-

[bigben2k]

That's the right stuff all right!

[MadDogMe]

It is going to be very difficult drilling the holes in both the base and jet plate, especialy with a wobbly old drillpress!. Can you get/make something for your Dremil that would work. The router attatchment maybe?. I'm thinking the tolerances may be tighter than a drillpress?...

That's goodly looking jets coming out of there. Maybe make up a 'jig/template' to hold them in place once glued for the final version?...

Were the 2mm brass tubes brought from a hobby/model shop?. I thought they might be too big (I/D) but they look OK ...

[bigben2k]

There's a drill press attachment for Dremels, but I'm sure it's even more wobbly.

I'd consider aligning the tubes on say, a board with nails, then glue them together, with resin or something.