Yeah some pics would be nice... could you post some when you finally get the patent and everything worked out? I and others would love to see how things were constructed etc... I'll definiately buy one if they are cheap like you say... then maybe I can get some better C/W numbers....

BBen... Got any ideas on Pump---CPU Vs Rad---CPU temps/flowrate are concerned? My system seemed to be cooler w/pump---CPU which, in theory, would preheat outgoing water, but would have the most flow....

Actually, it's a myth (maybe pH should add it to the list) that flow decreases, the further away it is from the pump.

If the pump creates 100 gph flow, then that's the flow at any point in the system, but since the diameter of every point in a rig is not the same, then the flow speed changes, but never the flow rate.

What changes at every point also, is the pressure: as the water is restricted through all the points, there is a pressure drop, relative to the pump outlet. Each one of those points creates turbulence, and that would be where the pump heat comes from. Otherwise, pump heat comes from the impeller turning around but not moving water as fast as possible.

So although the pump itself produces heat, so does the blocks restrictions. Just out of curiosity, is there a 90 deg bend at the outlet of your rad? Look closely: is the rad barb a 90 deg bend from where the water comes out of the rad fins?

I'm unable to otherwise explain why you would get better results with pump -> CPU -> Rad, versus pump -> rad -> CPU.

I just found this thread - sounds like a good block you've got there Bruno, can't wait to see the finished version! I don't know whether you've made the decision or not yet, but I'd also say go with the screws for fitting the top, rather than soldering it. This allows people to open the block and clean it thoroughly, rather than just having to flush it out with some cleaning solution and hope. I've got a two year old DangerDen block which I reckon could do with cleaning out by now, but I can't get the top off :(

Plus, with a screw-on top, it would be easy to experiment with different materials - for instance perspex/lucite, I know there's plenty of people out there who like the clear-topped blocks. No reason to leave them out of your customer base :)

Well done so far, I admire the effort you've put into getting this far. Good luck with the rest of the project!

It's very important that you read the text instead of just looking at the pics, I know not much people do that but anyway...

Before I forget, it's definitive, we're going for screws + silicone for the sealing of the top/bottom contact. If you want to clean it just remove the screws, pull a little hard so the silicone will come off and then add a little silicone when you're closing it, you may not even need it but just in case..

I know, I know, it's wed I told you they would be cast on sun.. But I think it has worth the wait. Here they are, the "raw" pics as lots of people requested. These blocks were just taken out the molds and all I did was clean them a little so you can see the inside, they have no finnish at all and they're still pretty dirty but you asked for it :). Of course these still have to be machined and when they're finished they will look much better than the previous pic I showed you, also, that shiny hehe

http://66.227.5.144/images/block.jpg
"Rough" blocks (without any machining)


I casted 9 of them, some are exactly as these first pictures but some came with a very small "defect" if you can call it that way, there was a little copper accumulation in a corner, it wouldn't affect performance at all, just looks, so I would like to know if you think these "deffective" blocks are sellable or not. These little metal accumulations were caused by the melted copper being spilled in the mold too fast, the ones that were spilled slower came out to be "perfect" so I could cast again and get a buch of block without that little flaw or I can just make these available (after finnishing them of course), what you think?

http://66.227.5.144/images/defect.jpg
"Defective" block

@Rob C: Thanks :)

Let me know your toughts on these
Thanks
Bruno Facca

Great work !!!!!!

Casting blocks seem very interesting process, considering the great things one comes to at the end... Too bad the copper melts on so high temp, so no home blast furnaces!

Tell us more about the casting process, making molds, problems etc pls!

PS: do you know how much the shipping/postage fee would be for one block to Croatia?

I wouldn't even call it a defect: it's just a small handling error in the casting process. To tell you the truth, it's so small that I'm not sure what defect you're talking about (You'd have to take a couple more pictures from a lower angle).

It certainly wouldn't affect performance, so it's not a problem, IMO.

So now you are going solely with the 4 bolt design?

I personally like the tab holding blocks, but the block im using right now is much lighter.

It also takes something more innovative to hold it with the tabs.

But the design looks cool, and so does the casting method, can't wait to see one of them all final and everything.

Great job Bruno, this is the first time I saw someone go and try casting waterblock, I thought of it once, but the price for hss steel mold was too much, I was thinking preasure casting though. They turn out pretty good, just wonder how will the block look when milled, if there will be any bubles and how porous will be. Did you drop any additives in melted copper for better casting?

I'm glad most people liked the results so far.

@Puzzdre: 1058 celcius, that was not easy to achieve but once you have a proper blast furnace is not that hard, except that it spends a lot of gas. Shipping to Croatia is a little more expensive than to the US, USD$13. Just out of curiosity why can't forks and spoons be sent to your coutry by mail? :)

@Raccoon: As I said before, there will be 4 versions of the block, one for 4 holes, one for socket lugs and each of these will be available in 1/2 and 3/8 hose fittings.

@morphling1: I was also thinking about pressure casting at first, I got some air bubble problems in the beggining but I managed to make the molds in a way that the air could escape so I never got bubbles again. It will be that porous in the inside (good for performance) and the outside will be machined until it looks like a damn mirror :). I didn't use anny additive, just pure copper. I never heard of these additives, what are they?

Thanks
Bruno Facca

NICE! Hey, sorry I didn't get back with you in a fashionable time, I've been working on another custom "Brazed-HomeDePot-block!" It's going to work nice, I can feel it, but anyhow, I KNOW I'm going to be getting one of them blocks that you cast! I was one of the first to show interest, but good work!

About that "flaw" that flaw could actually help and add to more turbulence! So what most people consider a "mistake," usually turns out to be a hidden jew of genius; Posted-Notes, Teflon (GOD!), all kinds of stuff...

How much did you say these blocks were going to be? I'd like to get a 4-bolt style.... GOod work :)

:( :shrug: :shrug: :shrug: :shrug: :cry: :cry: :cry: :cry: !!!!!

LOL !!!!

This is the first time I heard of such a...a stupidity !!!!

Maybe people who make the regulations and the laws in my country still eat with their bare hand (I wouldn't be suprised...):D

Really, this is just another stupidity of the our national post office or such...

Dunno, maybe they think it's a new weapon against somebody or something... or maybe those spoons are smaller than the average mouth here...dunno...

stupid, stupid, stupid !:rolleyes: :rolleyes:

Bruno, if you're going with a screw top, what's your solution to sealing it?

Paul has a couple of informative posts in the waterblock design section.

linky

Hi bruno, People from Brazil are interested in your block. Please check www.hardmob.com.br
Congratulations. Nice work.

@V12|V12: I would also like to see some of your work, maybe you could e-mail me some pics when you have them.. The 4 bolts model will probably be available before the socket lugs one, theý will be USD$15.
About the "flaw" you may be right :) I'm glad you liked it.

@Puzzdre:
LOL :D yeah maybe that's it hehehe

@bibben2k: the contact surface between top and bottom will be very flat so it should seal itself without need for anything else but anyway I will be adding some silicone before screwing it, the excess will be thrown out when I aplly pressure to the screws so it will only be in any placeds that the water would pass (if there are any), that way both parts can still exchange some heat, so we don't waste the top surface area. Thanks for the link.

@Marvin: Check your e-mail please.

Thank you
Bruno Facca

I'm giving those 8 blocks finnish (they will be sold as a pre-relase), here's what I'm doing for further performance testing:

- 1 block with wall thickness reduced to +-2mm, not the internal divisions just the outside walls, bottom is "regular" thickness. Normal wall thickness for the sides is 3.8mm.

- 1 block with bottom wall thickness reduced to 2.5mm. Normal bottom thickness is 3.5mm

- 1 block with height reduced (as in some of the top shaved off) to <I'm not sure exactly how much for as I still didn't do this one>

- 1 block with the water inlet right above the CPU core, this one will have 2 water outlets, one on each side

The other 4 are "regulars", the last one I mentioned will have only it's top modded so it can also be "converted to a regular" just by switching tops.

As soon as I'm done making these blocks/mods I just described (don't forget I still have to cast the tops) I will have some extensive performance (Cº/W) data

Thanks
Bruno Facca

btw i found a block on ocau looks very nice, and performs pretty well too. gets about 10Mhz higher that a cyclone 5, which in turn gets about 30/40Mhz higher than a maze 3 on an xp at 2.0-2.2volts at just above 2 ghz

get an ocau account, its worth doing there are some people on there that really know their stuff, such as volenti's new evaporative (bong) cooler, then check the thread out:
http://forums.overclockers.com.au/sh...5&pagenumber=1


http://www.employees.org/~slf/concept/cb5.jpg


go to the thread for more info

edit: seems like cathar created his own thread here... check it out...

Sorry for taking so long to update you. 7 out of 9 blocks are done with the "heavy machining" part (yes it took me all this time). As I said the original model absorbed humidity from the molding material so the blocks got 1mm bigger (wider) on one side, that means if I had put together the pieces of those tools I'm making to "automate" the machining process they would be irregular and wouldn't work for the next batch of blocks. In short I'm having to machine all the 9 blocks by hand, again . That's why Í'm taking so much time. I expect to be done with the heavy machining by tomorrow, then it's the thin sandpaper and polishing them up.
I decided not to cast the tops, instead I'll just buy some flat copper bars and make them the exact size of the block, so the ACS logo will be painted instead of extruded, it will look very cool. That should make the production quicker and easier.
Some of you are probably wondering "how the hell is this guy planning on selling these blocks if it takes so long to make just a few of them?". The answer to that is, once I get things going (the tools are ready, I got the hole references and this kind of stuff) production should be fast, around 16 blocks a week in the beginning climbing to +- 40 blocks a week after a few months.
I forgot to mention the blocks will be cleaned using a *big ass* jacuzzi pump driving a very powerful and thin water jet. Sand blasting is not an option right now and I got this pump lying around so...
I also got myself 2 very thick (read strong) 120mm 220V fans for my radiator (wich is already very good) and a 316GPH pump wich should also contribute in a high performance testing rig for the blocks so I can get the best out of them.

@Pinkster: Thanks for the info, that is indeed a good forum.

Thank you
Bruno

First of all I'd like everyone that read is reading this to post their opinion on this:
Is it important for you that the block is polished (looks like a mirror) on the sides or will you be satisfied if it's not polished but it's really smooth as in you can't see a single scratch from the machining (a very very thin sandpaper would be used)?
Please give your opinion as this is very important.

Good news are I got myself a belt sander, that should help the machining process _a lot_.

Here are some work in progress pics:
Before you ask about the top thickness, it's 6mm, I know thats too thick (it will be 4mm in the final relase) but it was the only kind of copper bar I could get without having to order tons of it. Anyway it's not that heavy. Well here they are:

http://66.227.5.144/images/blocks.jpg

They look "brownish" and uneven because I work with water most of the time to cool them down while they're being machined so they are a little oxidized, of course that's not how they going to be (a thin sandpaper and they're new). They are already very flat and quite smooth, they will be even more as they're not finished yet. You can't see much detail in one of the pics so I also took a close up pic of one of those blocks, this is how they look so far.

Bruno Facca

Before posting any more pics I want to know if there are still people reading this thread, cause if no one is reading it anymore there's no reason for me to keep posting updates here...

It took me more time than I imagined to drill all the holes for screws and hose fittings because (besides copper being a bitch to drill) as I said before this batch of blocks are all 1 or 2mm wider on one side so it's hard to measure and determine the exact places for the holes in a irregular piece, also I got a cut in my hand from a top that "grabbed" the driller and started spinning like crazy, to avoid that and make the drilling process a lot faster I'm putting together something that will hold the tops (and the blocks) in the right spot in the table top driller.

I should have the C/W numbers in the next couple of days.

Bruno Facca

Well Bruno, if you look at the previous page, you'll find that this thread has been viewed 5'520 times. That's pretty good around here, you're in the top 1% of all threads.

Keep up the good work!

People don't always reply to stuff they read, I know I don't :D . Keep it up Bruno, it's easy to get disheartened at an early stage, you got this far, so keep it coming.

PS, the finish does'nt mean jack to me. it will soon tarnish anyway unless it's laquered, so keep it smooth but not mirrored ;) ...

@bigben2k and MadDogMe: Thanks :)

BENCHMARK RESULTS:

AMD Athlon XP 1700+ at 1467 MHz (stock speed), Vcore=1,76V, the following temperatures are taken when running at full load non stop for 4 hours:
All temps were taken from the mother board's thermal sensors so they may be reported highet than they actu ally are, for more info please read this: http://www.swiftnets.com/troubleshooting.htm
Heat dissipation is 64.7 Watts for the processor and (*IMPORTANT*) my 1200 Liters/h pump is rated at 12 Watts, I don't know how much of that is converted into heat, I will make it 6 Watts for the calculations, let me know what you think about this.
CPU temp: 37ºC
Case temp: 32ºC

We have a Cº/W ratio of 0,11!

Please let me know if you think any benchmarking methods I just described can be improved for more accuracy.

So what you think?

Thanks
Bruno Facca

Why do people always insist on quoting the case temperature as if it were the radiator air intake temperature?

@Cathar: No need to say it like that. And maybe some people do that because thei rads are inside their cases (not my case).

@everyone:
In fact, I should have measured the radiator air intake temp and the water temp and at the reservoir before turning the rig on instead of case temp, my mistake, tonight I'll post the new numbers and a test of another version of the block I have here.

You wanted detail, I'll give you even what you didn't want to know :)

PC specs:
AMD Athlon XP 1700+ at stock speed and Vcore
Soyo Dragon +
256mb DDR
2x Samsung 40GB 5200RPM in RAID mode 0
Geforce 2 MX 400 w/ TV out
Pixelview TV/video capture card
LG DVD ROM 16x
Samsung CD-R 8x4x32

Water cooling system specs:
Atman 1200 pump
Aluminium radiator with parallel flat tubing (not in the market)
2x 120mm 220V fans
1/2 tubing
pump - block - rad config


Here are some pics of the test rig, I didn't want to take those because the block I'm using has absolutely no finish so it looks very bad. Of course the ones for sale won't look anything like this. Also the soldering (for the hose fittings) is ugly because I was experimenting a new soldering method on this one.

http://66.227.5.144/bruno/benchmark1.jpg
http://66.227.5.144/bruno/benchmark2.jpg

Before you ask, the black box in the rad is the fan's speed control so I can make it silent when I'm going to sleep with the PC on or watch a DVD.
:D

Thanks
Bruno

Looks abit like my setup, sprawled all across the desktop! :drool: :D . Viva the day I can afford some silicone tubing and an incase rad!,(twin parralel The Deeps!, Hmmm!) damn heatercores bigger than my 'sweet~jar reservoir', barely fits on my table, let alone in my case!!...

I'm sorry about not posting the results last night as I told you I would. I decided to make a digital thermometer this afternoon, so I can get accurate measurements for water temp and room temp, the damn thing wouldn't work so it took me all night to troubleshoot it, now it's working and calibrated.
About the CPU I'll figure a way to use the processor internal thermistor, either by modding my mother board or attaching a circuit that can read from it, it shouldn't be too hard..

@MadDogMe: I kind of like it that way, I like to see the water running and also keep as little water inside my case as possible :D so the rad and reservoir stays out, of course I will have to organize it better..

Thanks
Bruno

I'd like your opinion on some temperature reading methods I will describe, I know none of them is very reliable but that's all I could think of, any new ideas are welcome.

I was looking into building a SMBUS interface using the maxim 6557 IC or similar but they're not available in the city where I live. I also couldn't find *any* digital thermometers other than those used to measure fever . So far I can only think of a few, not so good alternatives to the original mother board (socket) thermistor:

- If the thermal diode inside the palomino's core is similar to the one in the mobo I could just rewire the mother board so it reads from the internal diode, but I don't know the specs for the mother board diode and looks like that would be to easy to actually work, what you think?

- I didn't looked at it yet but most mother boards have an area in the inside of the socket that has no copper tracks, components or anything, if a reading from below the core (a temp sensor touching the bottom of the processor) is reliable I'll make a hole in the mother board for the wires and put a temp sensor in there there

- A thin (but wide) copper "bar" that goes below the water block, touching the core, having a temp sensor soldered in the other side (right below the water block) insulated from the water block with, well, something... My temp sensors don't fit between the block and the core.

- As suggested, just pointing a fan at the socket to improve heat convection (or is that irradiation? can't remember) from the CPU to the mother board thermal diode, after all they were designed to work with some air turbulence around them and they calculate the temp predicting air movement caused by a heatsink and not water cooling.

That thermometer I made for measuring water temp is fluctuating a bit (just like 1 degree), probably because of electrical noise so I'm re doing in, this time with better grounding in the PCB and all shielded cables, if anyone is interested in making one of these for yourself it's based on the IC 7107, just ask and I'll post a link to the schematics.

Thanks
Bruno Facca

UHOH, mobos are multilayerd (I read somewhere that they have 4 layers or more) so drilling a spot that looks like it don't have the tracks may result in killing the mobo...

I would scratch that option...:)

Get a DigiDoc5?, they have a flat thermistor with them that would fit between the block & socket to touch the core edge. getting a precise reading is damn near impossible anyways, totaly damn impossible with a socket thermistor :D ...

between the socket and a Digidoc you'd know you're somewhere in the ballpark though :) ...

Amazing read!
 

Bruno,
I'm new here and just learning, but I wanted you to know I read every post in this thread this morning and enjoyed seeing your project get so far. It was a great read!

Cheers,
Mark