DT 5Noz waterblock

[DetroitAC]

Yeah, hide the nozzle was a decision I thought hard about. The nozzle is really easy to duplicate, you just see a picture and you have it, no need to get a sample and measure. I have no doubt competitors will catch up and some will use the same idea but implement in a different way. I am only delaying them by a month or two, but for me those few months equate to money. My guess is the nozzle will be public in under amonth from now.

[st0ned]

Btw it's all fixed with the webshop now. Thanks for fixing it in such a short notice

[avddreamr]

I guess patenting it is out of the question?

Looks nice, I might get one when I do my next watercooling upgrade. Good luck on the business venture

[stren]

I hadn't seen any photos of the thumb nuts before - so here's a photo from my build:

The actual block itself is pretty large - probably in order to fit the 5 nozzles, much larger than most other blocks I've seen. Ignore the barbs, they are just place holders.

[st0ned]

I've to look for sume thumb nuts as the ones that martin used, since I don't find those plastic ones that pretty.

[modster]

Stren, what were you using before? Can you post your performance increase once you get it running?

[Church]

modster: i doubt performance change will be THAT noticeable. Temps are on par with the best, but top blocks from all the companies are within 1-3C. Flow resistance is halved from next best .. but it's not the only block on loop and often importance of flow is overestimated. So imho it's not so much worth to upgrade existing if you already have some of top blocks (EK HF/CPU-360 or 370/Kryos/Apogee XT or HD/XSPC Raystorm), but simply consider it as the currently best block out there for your next purchase/build.

[modster]

modster: i doubt performance change will be THAT noticeable. Temps are on par with the best, but top blocks from all the companies are within 1-3C. Flow resistance is halved from next best .. but it's not the only block on loop and often importance of flow is overestimated. So imho it's not so much worth to upgrade existing if you already have some of top blocks (EK HF/CPU-360 or 370/Kryos/Apogee XT or HD/XSPC Raystorm), but simply consider it as the currently best block out there for your next purchase/build.

I agree and that's why I want to see how the block performs in a restrictive loop. I can see stren has quite a few blocks and I am assuming he is running all his blocks in series with one single loop. I am interpreting the test result a little differently than most people here and think low restriction may not be a good thing in an already restrictive loop.

[Boogerlad]

with this mounting system, is the only way to get consistent pressure across the four screws is to tighten to the maximum?

[stren]

I agree and that's why I want to see how the block performs in a restrictive loop. I can see stren has quite a few blocks and I am assuming he is running all his blocks in series with one single loop. I am interpreting the test result a little differently than most people here and think low restriction may not be a good thing in an already restrictive loop.

Modster - I'm actually running two loops on that board - it's a workstation build and cpu clock matters the most, so the cpu get's its own loop with a mora 140x9 rad. The motherboard/ram/gpu will get it's own loop. On the other half of the case I have a 990x and 3 gpus which will be on one loop, I have another 5Noz for that. I will be doing some temp testing - but given Martin's results I expect them to be within the noise floor of my unscientific setup

See the link in my sig for the build if you want to see more.

[DetroitAC]

Nozzle pictures have been posted elsewhere, here are some that Martin took

DT5noz-Internal2.jpg
DT5noz-Internal.jpg

[wez]

Wow, love it! Very interesting design

Will no doubt grab one if they ever are sold in EU (tax/duty is real messy for import here)

[zeropluszero]

lol. of course its high flow. there's nothing stopping the flow.

[DetroitAC]

"Artwork" of the flow paths
5Noz flow path1.jpg
5Noz flow path2.jpg

[Shteud]

DetroitAC
Very nice work!
This is very good so far are located enthusiasts looking for new ways and opportunities. Interesting idea. Waiting for repeating units in the series

[Church]

DetroitAC: But i wonder WHY you still get good thermal efficiency with THAT LESS flow concentration? Weren't others getting efficiency with concentrating flow via their jet plates right above core below IHS? Yes, they can also easily get flow resistance down by eg. cutting similar 5 wide cuts in jet plates of their blocks, but usually it will result in worse thermals then your block (at least from what i vaguely remember from skinneelab's EK HF least resistant plate test somewhat in between your block and best performing p#1/p#6 plates with most pressure drop). They can shift pressure drop/cooling efficiency ratio to different value if they wish for, but i don't get how you managed to get ratio itself that much better then others, and imho this nozzle system is not the answer, as it is simply parallelized/wider/lesser pressure drop waterflow path with less concentration. Or in other words imho with making your block more resistive (eg. closing two side nozzles) it's efficiency (flow resistance too though) has room to enhance cooling even more, while keeping flow resistance at least within EK HF's levels.

[Sparky]

Who cares, as long as it works!

[Kayin]

It looks to me like it's working on a simplified (and more open) Cuplex XT Di concept. He's getting the entire wetted area to give up heat with his version, though-it's really a quite ingenious idea. It's quite a concept-one I wouldn't have just stumbled across.

[DetroitAC]

There are a few things about this concept that are different than just cutting more slots in a jet plate.

The flow strikes a lot of fin leading edges, heat transfer is excellent at the very turbulent zone near a leading edge, this is why high performance air fins are louvered to give thousands of leading edges. Although if you just cut extra slots in a jet plate you get the same number leading egdes. But cutting extra slots in a jet plate just gives the flow a low resistance path at the outermost slot and a high resistance path at the innermost slot, it's not the same as this concept because each of the possible flow paths see pretty much the same resistance.

The flow from each nozzle is directed downward towards the bottom of the microchannels and then forced to turn and split into two more separate flows inside each fin channel. Then these flows collide into each other at the bottom of the fin channels, more turbulence .

Next comes the part where the nozzle does something unique, the flow comes vertically out of the fin channels and into the little tunnels of the nozzle where it is able to reach the block exit. So water flow only passes through about 1cm or so of fin before exiting, but there is a lot of turbulence in that 1cm even though they are smooth machined fins.

I agree with you Churchy, this concept could be "tuned" to get the heat transfer and resistance to many different operating points, the "next" nozzle is what I'm currently developing.

[CHiLLiN1369]

I Got one and was able to do a little shopping locally at the store...

[Church]

Hmm, so more turbulence from opposite direction flows instead of getting more turbulence from plain speed/concentration in one place as in other blocks? Hmm, sounds possible. I'd also check if your design somehow doesn't help for better thermal contact/pressure then common design.
As for your next block .. maybe not eliminate side inlets then, but make them all narrower by some 20(?)% and more closely located? Probably some study of cpu core size/form/placement under IHS plate of all the latest generation chips is needed to get more of cooling/turbulence right above him, less - on IHS sides without chip beneath. As i've stated previously, you have big margin of growing flow resistance till you get till next best block, so why not use it to get eg. another two degrees down?
P.S.
For bling side of future blocks you may consider All black. Including springs/mounting screws&bolts and so on.
P.P.S.
Another idea that might be tried - locating inlet/outlet fittings at non perpendicular angle to block surface? Imho this might allow more close/optimal inlet placement while still allowing the very biggest compression fittings side by side, and reduce flow resistance a bit. Probably harder to manufacture though.

[st0ned]

Wow, love it! Very interesting design

Will no doubt grab one if they ever are sold in EU (tax/duty is real messy for import here)

Well I got one in Portugal ( Ordered from the online store ) DT was kind enought to help filling the gaps so it didn't got any attention of costumes therefore I got it without paying extra taxes .

Some photos I took ( I'll show the injection plate since DT did it himself and I think theres no problem in showing it any more ).

[BeepBeep2]

Hmm, so more turbulence from opposite direction flows instead of getting more turbulence from plain speed/concentration in one place as in other blocks? Hmm, sounds possible. I'd also check if your design somehow doesn't help for better thermal contact/pressure then common design.

Mounting for this block is actually a bit less pressure than standard mounts if I read correctly.

[modster]

DT, how many blocks have you sold so far?

[eXa]

Very nice injection plate. I wonder how much blocking the to outer smaller holes will affect temperature and flow...?