cascade style
 

heres a few pics of my first waterblock, not finished yet.
tools used include dremel and drill press.
no cnc :)

maybe this makes things clearer:rolleyes:

and a first test to check flow pattern - mmm seems ok :cool:

heres the plenum chamber, had a bit of fun trying to mill that "freehand" hehe. its 8mm polycarb (thx dad!)

the needles are just that, disposable medical needles - 40x1.25mm OD, approx 0.85 mm ID. Hope the stainless steel and copper wont try to do anything with each other??

hope no-one will be offended by this rather crude imitation of a fantastic design: thx goes to cather, and all those who originally thought this up .
btw, dont worry, i dont intend to sell any:)

Looks very nice. How deep are the holes in the copper base? They look to be a little shallow to me.

Its always nice to see what a drill press and dremel can do with a little ingenuity. I bet that thing works great. The jets of water look like they are evenly spread across the entire surface, a good sign that you got that part right. I'd like to know what the OD of those tubes are compared to the width of the impingement holes on the copper and the depth of those holes and the thickness of the copper all-together. Looks nice!

hi Cathar and killernoodle!

ok, the holes in the pic are full of water so they dont look so deep ,but actually they are 3.5mm in a 4mm base:D

i managed to get the bottoms fairly flat (home made end mill, lol), but was careful not to get carried away when i saw what was happening to the underside of the base :eek:

holes are 1.8mm , tubes 1.25mm. i tried to optimise jet density, and kinetic energy, with a compromise as to machinability (60 holes is manageable, 300 is errrr, etc...)
no cfd unfortunately(!?)

anyways the end result is probably not far off from the original cascade, so hopefully results will follow likewise:)

Looks nice to me

Hope it works good :D

thx for the compliments Dieter:)

anyway, been busy again !

i wasnt too happy about some of the machined surfaces once i removed the protective film off the parts, so i had a shot at straightening the edges of the plenum chamber - this time using a g u i d e (hey, live & learn) and a lot of cooling water. got messy needless to say, but was well worth it :)

also cleaned up the outlet holes (again using water to lubricate the tool) and i think they look rather cool now.

was slightly apprehensive about the barbs as i cant find anything under 15mm in the shops, with at least a 1" thread - arrgh - , so decided to make those as well.
they dont look too stupid afterall, methinks.

just hope the araldite (slow setting stuff, should be stronger) holds up to its promises

:edit: note the scorelines on the top surface, dumbass hehe, lol

I really hope the block works, cuz I saw some posts of a member (jaydee16 or something) saying he had made several blocks which all failed...


edit: check this topic
http://forums.procooling.com/vbb/sho...5&pagenumber=1

But maybe they all failed becauce the "jets" (of whatever they are called) dont go into the holes...although I couldnt believe he would make the same error over and over and over again.

Looks nice, but how are you gonna bolt it all together? Did you tap the copper so you can put a bolt in it?

The problem is that the jet tube OD is slightly too large.

If your cup ID is 1.8mm, your jet tube's OD ought to be 0.9 or less.

But hey... try it and see!

I made several versions. All but a few had the jets IN the cups. It is not an easy design by any standard. Any jackass can drill a few holes, put in a few tubes, and drill some cups in a chunk of copper. That isn't the hard part. The hard part is finding a working combination. Cathar made 20 variations of the Cascade before he found the sweet spot. I made 10 and have not gotten what I was hoping for and all but given up due to better things to do.

You need the proper base plate thickness, proper cup diameter, proper jet diameter (inner and outer), proper length of the jets (how much time the water spends inside the jet tube), proper bore (depth and width) above the jet tubes for the water to spread to all the jets in predictable form, proper clearance for water to escape the cups and go to the outlet(s), etc.... It takes very little to make the block perform very well vrs. performing bad. Very touchy combination.

I would think that alignment in this design is critical. I don't know for sure but if the axis of the jet isn't totally aligned with the axis of the hole I can't imagine it impinging very well on the bottom of the cup

thx guys for bringing me back down to earth :)

bigben, im curious as to how you come up with the 1.8/0.9 ?

i admit i havnt been here long enough to have read about everyones experience of jet impingement blocks...maybe thankfully, considering the apparent lack of "success" these have had :shrug:
also i'm not too sure how to quantify success/failure objectively, but im fairly certain that this setup will give me more satisfaction than my actual full aircooling setup:rolleyes:

atm im just happy that the needles all line up in the cups (happy jackass hehe:) )

in answer to the questions:- threads in copper base, probably clear silicone sealant (i want to be able to take it apart + stay transparent)

Hey I rank up there with the king of jackass's. I got over $3,000 in making water blocks.....for myself. So don't take the jackass comment to harsh. :D I think I have spent more making water blocks in the last 3 years than I have spent on computer equipment. :eek: Thats all coming to an end now though.

thats ok Jaydee:) guess if i had 3000$ to spare i woudnt mind putting it into cnc, etc… either :)

<getalife mode>

anyways, i did a few flow tests with the block and a small pump (i still have some lapping to do before it gets the real test)

dont be misled by the number of points on the graph, i only measured one point, which is the pump/waterblock+line intersect. i'm assuming that head is proportionate to square of flow rate (pretty standard assumption)

pump is a hydor S30 (similar to L20) slightly modified at the inlet :). i checked the static head, 1.34m instead of 1.35m.
max flow approx 810l/h instead of 600 (overclocked hehe)

sorry about the SI units for you unfortunate yankees!!

conclusion:

the block is fairly restrictive, 0.7m at 5lpm (0.58psi @ 1gpm), however (hopefully) the pressure drop is put to good used by increasing the water jet velocity.

i was curious as to the effect of the base plate on the flow resistance (what proportion of the pressure drop is "used" by the jets, and how much is "wasted" in getting the water back up the cups and out of block.

Welp, i measured the block flow rate as 5.3 lpm with this pump and the 3m of piping (see graph)
when i removed the base plate, the flow increased to 5.9 lpm. (big deal, huh)
however… this translates to approx 0.45m @ 5lpm for the block without the base. (instead of 0.7m with the base) hope you're following.
Therefore we can say that the base adds a fairly considerable restriction - this is actually more than 50% extra head loss. (+- quite a lot:D )
However, interestingly, and perhaps suprisingly this makes very little difference to the flow rate (-10%).
Bear in mind that this is specific to one particular pump, but i guess you can generalise it if the system is operating in the "top region" of the pump PQ curve.

*This bit of info could be of use to those of you running 2 or more blocks in series…*

</getalife mode>

phew, sorry for going on a bit there, hope someone will find it useful:)

For those of you who are thinking of making you're own jet impingement blocks, I would guess that *one* of the critical parts is sizing the jets. This is perhaps the easiest bit to design "right", and will probably also make or break the whole project.

The graph below could be used as a rough starting point. If you know how many nozzles you're going to use, then it will give you an idea of what diameter nozzle you should be aiming at.
It would appear that 2-3 m/s velocity is not an unreasonable target at 5 lpm flow rate (the green lines). Less is going to be too sedate, more will be tricky without a high pressure pump.

Dotted lines correspond to my particular design :)

I'm waiting for the real-life tests :)
Or aren't you going to install it?

hey, I may have missed it, but how did you secure the needles? just had a tight fit or what? How much did the needles cost?

If you know a doctor, veterinarian,... you can get those for free ;)

The needles are a nice push fit, well I used the drill press for extra force:)

Cost breakdown for hardware is as follows: (excl. tools, time, beer, etc)

*8mm polycarb - offcut, free
*4mm acrylic - same
*4mm copper base - pulled from a 'borrowed' IGBT power module (big one, probably 600A or more :drool: )
*25 disposable medical needles 1.25x40mm (not used! ) - 1.5€ (+ a little embarrassment - make sure you're the only customer if you buy these at the chemists, lol)