LOL! :D . It's water cooled not water pooled!, at least I hope so! ;) . I've had a hectic week, I had to go to hospital to have an X~ray taken of my jaw so I can have all my teeth pulled out!! :eek: :cry: . I blame it on my American decendancy conflicting with my German surname while being English born. That or sleeping with my head under the pillow!, damn tooth fairy mugged me!! :mad: :p ...

The block?: It's still working fine, I hav'nt had time to do anything to the block though, to try out improvement ideas. I'm interested on how well the base being thinner works out with your latest though...

___________________________________

Of interest?:
I found a small circular saw blade magickly appeared in my toolbox the other day!(along with a brass barb with a 'jetted' bottom!), it's about 2' in diametre an produces an approx' 1mm wide cut!!. the only problem is it has a big hole in the middle for mounting(approx 1CM). So a custom arbour is needed. Seeing that Dremel router gave me ideas :D , now if we could get hold of some 5mm thick copper!...,,,

Laters Dudes :cool: ...

dude that was the most painful post ever. best wishes on the teeth pull :eek:

so very close but so sweet

http://homepage.ntlworld.com/lee_smith/v3-6.jpg

i got my amd mobo today and it has no mount holes do you think it would be ok to drill holes into the solder pads.

this only cost £25 so if it breaks i dont realy care.

http://homepage.ntlworld.com/lee_smith/amdholes.jpg

off topic the silver trim , uv dye arived today along with the danger den hold down set and 1/2 y

http://homepage.ntlworld.com/lee_smith/trim.jpg

have you checked out hoots thread at overclockers he made his own block from scratch with a table saw and dremel and it looks cool.

http://homepage.ntlworld.com/lee_smith/catclone1.jpg http://homepage.ntlworld.com/lee_smith/catclone2.jpg
http://homepage.ntlworld.com/lee_smith/catclone3.jpg http://homepage.ntlworld.com/lee_smith/catclone4.jpg

That does look sweet, pretty much what I had in mind for with regards to the circular saw blade (you can buy better sized ones that are even thinner than 1mm for under £30, but this is in my hand! :D ). I would've used chrome allen bolts that look better than crosshead/phillips screws though...

WRT holes, I've heard of people drilling them out on certain Asus mobo's with no problems. If the solder goes right the way through/is on both sides, it should be fine. Make sure you use the right sized drill bit and centre the holes precisely though, so you don't cut into the PCBoard or/and traces...

_____________________________________

Would you be interested in making a block simular to Hoot's(or 'cross cut' into pins) with your router if I can find a mount or another blade to mount in the Dremmel?. I can't find an outlet for flat copper bar stock near me though. How about you?...

PS, what size is the inner hole on those 1/2' Y's?. I have some oversized ones (named 17mm, but closer to 13mm outer. The barb/flanges are 17mm), but the centre narrows down to 10mm as it gets near the triple join. They could be safely drilled out to 11.5mm though. 12mm at a push. Would one of these be better than the 1/2' one/s you have?, if so I'll post you one :) ...

PPS, I would of used a 13mm inner Cu T rather than a plastic one, unless you drilled it out?. They make them so restrictive when the inner size 'funnels' down as it reaches the join :mad: ...

i think i will go and get some lotto tickets tonight as i am feeling lucky.

I checked the mobo one last time and went for it. Drilling 4 holes in the center of the solder pads. After checking the back side i noticed that when the drill had pushed through the pcb it lifted a little of the board and a track along with it.
thanks to my soldering skills from a previous job a quick jumper fix and the mobo works.
http://homepage.ntlworld.com/lee_smith/track.jpg

_____________________________________



neither do i but i do have some links. 5mm thick was it.

if we can get some copper i would give it a go.

i am going to make one of the blocks in the new design thread with the dimples cut into the base and jets over the dimples should be easy enough and i have a akasa base that i can use.

the brass one is 10mm id it could be drilled out to 11 but not 12.

if will wait to see what the poly ones are like and if they are no good then one of yours would be good thanks.


on my main rig the tube is 3/8 and the t is 1/2 so they are ok for that but not for the test rig that has 1/2 tube i will get one.

i have the maze 3 running on my amd board on my test rig (a bit of board with the stuff on top).

i have a temp probe glued to the side of the amd core , a water temp probe and ambiant temp. I will leave this running at idle for a while and do some tests later and Then my new block.

Lee

Yep, those 'jet & dimple' impingment designs look good, and very easy to make with the most basic of tools (woohoo!! :D ). The 3mm bases are a perfect thickness as well IMO. Thin enough for good performance bringing the dimples to near 1mm thickness yet sturdy enough to stop flexing/distortion...

I don't know if they are a perfect design for a large P4 heatspreader sized chip though, ????. Have you thought (or read) much about removing the heat spreader?...

so good news first.
after several tests over 2 days i have come to the result that my block is on an average 1c lower than the maze 3.

my rig is a asus a7v-ml and does not show cpu temps in the bios so i had to use a temp probe on the side of the core. I super gluded it in place so i never moved when i changed blocks.

i tried to maintain a room temp of 20C while i ran many tests.

but i found that it reacts very quickley to ambiant temp / water temp. whe the maze 3 will hold temps for a little while and change gradualy. I pot this down the thicker base on the maze 3 tacking longer to warm/cool.

bads news i cracked the lexan top near the screws that hold the top to the base and it started to leak it bit. I have bodged it back to health but will test it for a while before i run it again.

How did you crack it?, overtightening?. They really don't need much pressure if you use silicone sealant...

it must be when i put the fancy nutcaps on.It is still leaking so i will have to make a new top.:(

when i opened the block i found loads of crap in the channels that would stop the water flowing into them. i guess i didnt clean the rad is well as i thought.
this might make temps a little higher too.

Been awhile,

me makes note to self, no fancy hex nuts with plexi tops.:D

That's a bad do man, sorry to hear it. Might try a nylon washer between the nuts and plexi next time, might help.

I've a piece 3 7/16" X 2 7/8" X 1/2" I could send ya. I don't recall your size requirements at the moment.

Another very good place for Lexan is ebay. I just bought some on there, 2pc. 6 X 38 X .5" both shipped to me for a $20. So I don't have to be concerned about just getting it as small scraps now for a bit. Don't see the thick & clear stuff real often, I was lucky & bought if for near nothing.

That base of Hoot's is another good looker for sure.

BE

edit:

That's the one tough thing with a used core, crap that may be in it.:shrug:

Hi BE,
At least i got to compare againts the maze 3 for a little while.
getting 1C better is going in the right direction.

i have a link here for lexan http://www.engineeringfindings.com/cat22/cat22pg78.pdf

they have no lexan / polycarb on ebay.co.uk.

the last pice was 3 inch x 2.5 inch but 3 x 2 would be fine as i am using amd holes on all blocks and the bigger piece was for the p4 holddown.

i am going to make the rest of my prototypes from the acylic i have but when i have it at the best i can get it some lexan would be fantastic for the one i will actually use.
:)

what do you think of cathars new creation ?.

Lee

If you're interested in buying the Maze3 you can always make me an offer ;) . It'd be more use as a benchmark to you than a block to me :D . Whatever, I'm not in any hurry for it as long as you don't forget about it (I've got a memory like a sieve) :) ...

_____________________

It's nice to know that there's another performance option. Makes me wonder how much of WW was impingment and how much was micro~channels...

I like Cathar's new block he's working on pretty well. I think with a few tweaks he may have a new champion. And if it does beat out the WW I'll bet D-tek will freak!:eek:

I'll be buying a CPU block soon, his new design may become the third block on my short list. His WW & Morphling1's block being the other two.

I've a idea to throw out to you guys. I'm thinking of making my NB & Gfx blocks. Here is what I've got in mind so far. Base to be of the Rotor's design. Tops to be .8 Security grade Lexan. With the strength that the security plexi offers I'm thinking of drilling into the top on a angle, then insert copper tube for the inlet/outlets. The angle of the tubes in the top to improve the routing of my lines from CPU to NB to Gfx. I'm thinking it will allow them to be a little shorter and run tighter to the MB and Gfx cards.

The Gfx would also have a top like this with it's inlet angled upward for the incoming line, and outlet at a right angle aiming towards a DIY res I'm going to make. The res will be made of plexi tube and be as tall as needed to put the inlet tube at the same level as the outlet tube from the Gfx card. Combined with my single pass rad this should give me a system set up with Xtremly short tight lines. And a good deal of my lines (all curved sections) will be copper and all lines to be 1/2 or 5/8".

Have I gone over the top??? Thoughts??

Lee,

(need blushing smilie here) I'm afraid I don't have your addy anymore. Drop me a PM again please.


BE

that all sounds good to me BE but i am suprised your buying a cpu block i though you would have something made.

here are some pics of an impingment block i made today it's very rough just so i can see what i could do. I made this in about 3 hours again from an akasa base.

http://homepage.ntlworld.com/lee_smith/impinge2.jpg.jpg

i even managed to cut the area around the jets like cathar but it's not even close by 1000000000 miles to cathars block what fantastic work he makes.The dimples are 1mm above the cpu core.

http://homepage.ntlworld.com/lee_smith/impinge1.jpg.jpg

it's made for p4 again. i have it sitting waiting for the silicon to set then i will test it againts the maze 3 soon.

dont worry i wont forget. once i get the top back up and running on my block and i confirm the temp differance i will send it back.

The answer is that it's quite naturally, both.

The degree to which one "dominates" is dependent on the flow rate, as is theorised by Les and his base-plate thicknesses. At higher flow rates (around 4lpm and up) the jet impingement aspect starts to become the primary effect at play, below that the mini-channels are the primary effect as heat is spread out and up the walls more. Both effects are always working in unison.

This is something that I've re-established with the Cascade block and different sized holes in the base plate and the amount of copper between the cups, which effectively are the cooling equivalent of the fin walls in the White Water.

Too much copper and the jet impingement doesn't have enough total surface area to compensate for the spreading resistance of the copper cup walls. Too little copper for the walls and the heat that the jet impingement isn't dealing with (adequately) can't spread easily enough to the cup walls to be cooled.

The jet impingement action has it's limitations. Given enough flow and jet velocity the need for fin walls can be effectively eliminated, but not with the pumping pressures we're used to dealing with.

Of course, extra pumping pressure goes hand in hand with extra pump heat, and so we have to make a trade-off here as well. It's never simply a matter of maxing any one aspect because this always comes with its own separate issues. It's all a complex juggling act that even now I'm still learning the many facets and balances thereof.

Lee,

As to the CPU block, I know my limits vs. what I want in performance. I'd spend more buying tooling to TRY to match the WW than one will cost me.:shrug: A lot more, as I have no mill.

But nothing on the market for NB & Gfx impresses me anywhere near so much. None of the blocks on the market I've seen shows the degree of refinement of those CPU blocks. Thus I'll take my own shot. And Gfx heat loads are increasing fast.

One thing I'm wondering about is if I should just use a dimpled base for the NB. Would give lower resistence, and the NB doesn't need it as much as the Gfx or CPU.

I'll state my intention to extend the Cascade design to a GPU/peltier suitable block. I've got the plans drawn up, but developing two block in parallel is more than my present resources can cover. Better to get the CPU block right, and carry the concepts over to the GPU block, which is what I intend to do. The resultant GPU block, should be a very near match for the White Water by itself.

Well that's sure a interesting thought Cathar. And would offer a very strong matched set. If you've got it out by the time I finish my build I'm just starting to buy for I'll go for that. As I'm in no hurry that won't be for a couple months.

And you'd have, by large margin, the best performing Gfx block in the world. Given the total lack of uber performing Gfx blocks it could perhaps be a better seller than the CPU block for you. A more limited market, for sure, but one you'd dominate, while the WW would offer your new CPU design strong competition. But then I'm sure you see that, just as the same applies for the area of pelt blocks for the CPU. The Swifty blocks will offer competition in the pelt area, but the aluminum tops are a strike aginst them, & the Cascade will have a lower C/W as well.

Congrats.

edit: Clarity

I was just wondering the other day why Swiftech never bothered with polycarb :shrug: ...

hi cathar welcome to bodgers corner.

Do you have any formulars for working out the best base plate thicknes and fin height with the fins i have on these bases ?
if i remember correctly the fins are .5mm thick with 1.5mm gap. The base plate is 3.5mm before i made any changes.

Hope you dont think i am trying to copy ww. I dont have any access to mills and have to use materials i can find and the akasa base is working very well for me at the moment.

Do you have access to an heat sensitive camera you could use that to gauge how the heat is spearding from one dimple to another. My guess would be you want the cooling effect of each dimple slightly overlapping all others around it to maximise cooling with the optimal number of jets / dimples.
Just a thought

i am not worthy

Lee

Given those dimensions I'd use a 1.5mm thick base-plate if you had the central inlet jet like the White Water, and 2.0mm without the jet.

Hey, it's fine if you want to copy the White Water for yourself. I've never had any problem with anyone doing that. All the better for people to enjoy doing something themselves if they're into that sort of thing, and have a decent performing block at the end of the day.

Dimples? Sorry, can't say I've read the whole thread through yet. I'll take a look and get back to you.

i have the base plate at 2.5 but cut 1mm channels between the fins. I do have sort of a jet on the center inlet.



thanks cathar :)

i ment the dimples / holes on your latest block the casecade mini cup.

how many Degree C is the WW beter than a maze 3.

Lee

finaly i have the thermaltake coolers.

http://homepage.ntlworld.com/lee_smith/tt2.jpg

the base is a fraction larger than the akasa and also 4mm at the thickest part of the base.

The fins have a 1mm gap and are about 0.25mm thick.
the akasa have 1.5mm gap and .5mm thick

http://homepage.ntlworld.com/lee_smith/tt1.jpg

i will have lots of testing info soon compareing

maze 3
fin base with 3.5mm base 1mm channels
fin base with 2.5mm base 1mm channels
dimple base with 3.5mm base 2mm dimples.

so far the 2.5mm base is better than the others by 1C but they are all performing very close to each other including the maze 3.

20 fins per inch, nice!

I find it odd that they would use a 4:1 ratio, of channel-to-fin. I understood that the ideal ratio was at least 8:1, but maybe that was for aluminium:shrug:

In water, that ratio is actually high, where Cathar calculated it (under set circumstances) to be ideal somewhere between 1:0.75 and 1:1.5 , but it is known that a higher fin density will provide better performance, albeit with a higher water pressure.

This should be interesting!

my only problem with this base over the akasa is the way the fins are slightly curved at the base. This might stop me from cutting channles between the fins. I might have to take the base all the way down to 1.5mm as cathar suggested instead.

i just checked and there are 27 fins per inch.