nozzle design help
 

Hi,
first i made this block:

http://homepages.paradise.net.nz/max...post/block.jpg

with this base:

http://homepages.paradise.net.nz/max.../post/base.jpg

its inlet barb narrows down from 10mm to 7mm ID to speed up flow for "impingment".

http://homepages.paradise.net.nz/max...ost/nozzle.jpg

Butt didn't perform any better than with a plain old 10mm ID fitting though.

Any ideas for improvement? I made this base yesterday with thinner copper and smaller drill holes:

http://homepages.paradise.net.nz/max...t/newbase2.jpg

and thought maybe a grill after the inlet barb with 9 holes drilled like this

o o o
o o o
o o o

Have you had success with anything like this or anything else or am i barking up the wrong tree?

I've done exactly that, (the grid of jet holes), now on 2 different blocks, and with somewhat predictable results.

try and position the jet's so they fire at the center of the divots and you'll be set, use 2mm dia jet holes if your pump is around 1.5-2m head, use 1.5mm holes if it is above that.

Oh and they perform well enough that I'm currently using one over my micro pin block.

Makes you wonder what designs we'll be seeing next...

yea, I like seeing these designs that perform so well yet are within the capability of most competent DIY'ers to construct.

max;

does that block have much of a gap betwen the copper and the perspex? since it doesn't look like much from the photo's, may be causing too much back pressure and hindering the impingement.

There is a 3mm gap between them. I only have an ehiem 1048, do you think i should even bother with this impingment thing with that pump?

3mm is ok, even with a pump like that it's worth a try, just go with the 9 x 2mm jet holes and see how it performs.

YEEEEEHAAAAA!!!!!!:D:D:D

Finaly i got a better design after getting nowhere for ages.
I used this top:
http://homepages.paradise.net.nz/max...ost/nozzle.jpg

and this base:

http://homepages.paradise.net.nz/max...t/newbase2.jpg

and got 4 C better than my other blocks and i haven't even tried the 9 hole thing:D

I am very happy:)

Just goes to show...thinner is better :D

I think thinner is better for inlet over centre designs but for maze designs it definetly isn't (IMO)

That's right.

The simple Maze design has no additional surface area over the core, which is bad, and only allows it to work well with a thick bp.

In this design, and similarly Cathar's WW, the fins/dimples give us the additional surface area that we need, and with the jets, the turbulence (both for WW and this design) required to improve cooling, most efficiently.

Remember, WW's bp is less than 1 mm! I don't know what you are using, but this design is the cheapest, closest, easiest thing to WW that just about anyone can build. I would *roughly* predict that 1.5 mm should perform nicely.

Some more info here .

Volenti! :) : How does your micro pin block(or current best?) perform compared to your 'direct~die~jet block' you had before?...

what i was wondering about:

max; how did you manage to narrow down the inlet barb?

Oh the temp readings are impossible to compare between the intel system thet the direct die setup was on and the AMD system I'm using currently.

however I made a base up for the direct die block and tried it up against my micro pin block;

http://users.bigpond.net.au/volenti/imp_block.jpg

(note those are 9 holes in the base, not divots)

it (as best I can measure) matches my micropin block, I have made another block with just one intake/exit and 9x2mm jet holes (verses 9x1.5mm) that better suits the pumps most commonly used, it is predictibly slightly lower performing (by ~1.5 degrees) but "adequate" for the job, I'm currently still using it since it makes the plumbing cleaner in my case.

This is where BB's large outlet steps in methinks?, it's very important to have as little backpressure as possible with impingment cooling methinks?, plus it'd give a uniform 'spread' to the jet if the waters not all travelling the same direction towards a single outlet?. Does that make sense?...

PS, does it mean that water is contacting more surface area at the maintaned high velocity when holes are used instead of 'divets'?...

I think the main reason the 2 barb block is performing slightly less is because of lower jet volecity, not necessairly the restriction on the outlet (though that is an important aspect to get right)

When I'm testing the blocks for flow patterns I introduce a small "leak" on the suction side of the pump so I can "see" what the flow is doing inside the block(the iwaki pump churns up the bubbles very fine), the single outlet has no noticable effect on the jet's themselves, only after the water comes out of the jet holes does it make a beeline for the exit barb.

I use the holes like circular "fins", and the water flowing up the side of the hole should be almost as fast as the jet that's entering it, this also allows me to use a relitively thick base to mantain strength, I believe a thin >2mm divoted plate would work as good if not slightly better but copper strength starts to become an issue then without added support which would get in the way of the jet's most likley.

Well it comes with a 1/2 inch hose tail and narrows down to 7mm. I usually drill them out to 10mm ID but got one that i hadn't drilled out yet.


I made one that had a grid of holes but it performed about 3 C worse than than the one above with just one 7mm jet. Probably because of my 1048 pump. When i had the pump in the open the 7mm would travel much further and faster than my multihole jet one.

This is really quite freaky. Ive only just stumbled across this thread...........honest I have, and

Ive just come back from the workshop where I have been drilling a block in a very similar fashion to the ones from Max and Volenti.

Mine has a copper middle section with a 20mm circle of 2mm holes (37 in total) acting as jets directing flow directly into dimples in a base section. The water then flows away to the sides and out.

God knows if it will work, since the appliance of science has been distinctly lacking in the design, but hey, it costs nothing to try ;)