Bushman-Tech WB - initial construction info

[DX2]

Initial construction infomation for below design

http://forums.procooling.com/vbb/showthread.php?t=12903




Ok finding the dimensions for the actual block is not as easy as id thought.

Having dug around procooling.com's forum and looking at various AMD socket specifications http://www.amd.com/us-en/Processors/...9_7203,00.html
ive come up with the following

whole size of cooler. according to amd pdf socket 940 specs
thermal specs cert. cooler in monunt is 77mm X 68mm @ 5 ribs = 15mm/rib ie 1 design unit = ~3mm
cam outline is 55mmx50mm @ 5 ribs = 11mm/rib ie. 1 esign unit = 2,2mm
cpu package 40mmx40mm @ 5 ribs = 8mm/rib ie. 1 design unit = 1,6mm

i was aiming for 1mm - 1,2mm design units, which means i will have enough space to increase the size depending how frustrating hand making this will be.

http://forums.procooling.com/vbb/showthread.php?t=12840
pin/fin height, procooling rec = 4mm , mine = ~8
base thickness, procooling rec = 1mm - 3mm


Practical Examples:

swiftechs apogee block is 50mm x 50mm
Base plate dimensions: 2" x 2" (50.5 x 50.5 mm)
http://www.swiftnets.com/products/Apogee.asp

http://www.cooltechnica.com/Merchant...egory_Code=VCW

pin/fin height = 4,5mm

Materials Aquisition:
Base:
Aluminium slabs from Ebay. 135mm x 50 mm x 20 mm
http://cgi.ebay.de/ws/eBayISAPI.dll?...e=STRK:MEWA:IT

Copper eventualy, but the ease of machining Alu Vs Cu, just drives me towards
Alu for initial tests and prototypes.

Lid:
for initial tests and prototypes, aluminium.
later on some plexiglass or acrylic polymer thingy. again probably from ebay. as i dont know where to get the materials locally. but if all goes well with initial prototypes ill start asking around.

Construction plan:

print a scale side view, glue to block, start dremeling / filing / drilling.
very DIY, low budget, fun.
as i imagine it my primry tool will be

• dremel cutting disc
• small files
• drill

In the design thread i have a little video attached where i kinda put the tools to the block, how i imagine i would do it. ill attach it to this thread too.
*.mp4 in a *.rar archive


convert to g-code for cnc, but thats well into phase 2, as i need to ask my dad about talking to the machinists at hiw work place. and what kind of bribe is needed for them todo some stuff for me. i hope a case of beer is adequite. lol


Questions:

Aluminium corrosion, is there any reason not to continue using the prototype blocks on less demanding system components for the semi long run?

ive been asked about budget. Well low. poor soon tobe university student low. which also translates into. ill cross the various bridges when i get to them. the initial prototypes are far more a fun hobby relaxed endevour. Thats why im willing to spend several days with a dremel working my way through a solid block of aluminium.

ps. nice little one page basic of water cooling. Something similar would be cool for the wikki http://www.directron.com/waterinfo.html

Have i missed anything? anything i should be awre of before i start my little DIY project?

I guess all i need now, is a bit of luck and some patience.

[bigben2k]

Keep in mind that Alu will react differently than copper; you should know that, and why.

I wouldn't worry about corrosion for now, but it is going to happen. Avoid mixing metals.

[BGP Spook]

Is that all going to be made from the same Alu block as one piece?

That looks pretty difficult to make.

Unless I am mistaken the thermal resistance is worse accross two mated surfaces compared to a single piece of the same size and shape.

[bobo5195]

Originally Posted by BGP Spook

Is that all going to be made from the same Alu block as one piece?

That looks pretty difficult to make.

Unless I am mistaken the thermal resistance is worse accross two mated surfaces compared to a single piece of the same size and shape.

v much worse.

Think cpu to heatsink.

[DX2]

Originally Posted by BGP Spook

Is that all going to be made from the same Alu block as one piece?

Yes, all the same block of Alu.

i havnt gotten as far as radiators yet. but chances are ill want to get a good copper one that i can keep for the full setup. and if it work out well with the alu prototypes. ill make it in copper. hence, eliminating the mixing of metals.

That looks pretty difficult to make.

The one picture i chose for this thread isnt exactly how i plan todo it. but it is an effective representation. In the design thread youll see its pretty similar to the apogee , as its a flat base with fins sticking up, and the lid is cupped over them. Rather than the lid being flat and the waterblock being cupped. Makes building it much easier for me.

[DX2]

My imagined construction plan

(a frame from the attached video)

From a basic Case Modding guide. This is pretty much my tool set. atleast thats how i imagine it.
http://www.creativemods.com/content/view/99/39/

[DX2]

A really great reply to my thread from the futuremark forums. I ported my thread to their forums. sorry to cross refference and reply. im trying to maximise the info in each thread without sending people back and forth between threads

Originally Posted by RealTime@futuremark.com

Having been a machinist I wish to add a few things:

1) Choose your aluminum carefully. I cannot read German, but I couldn't find anywhere that stated the alloy type. If its 6061 or above, you may have some very long nights ahead of you. Softer alloys cut easier but tend to be more difficult to grind (eg cutting wheel)

2)Using a Dremel for this work will be EXTREMELY tedious by hand. I suggest springing for a Dremel router attachment and a good bench vise as it will help in keeping your cross-cuts more consistent.

3) Buy GOOD cutters for your dremel. I would turn to an end mill supplier and look for high-speed 1/8" (3.175mm) shank steel end mills to fit into your dremel. I would suggest staying away from the Dremel brand cutters as these tend to be cheaply made and dont last long. The smaller end mills are reasonably priced and last MUCH longer. Just buy the shape cutter you want. Here is an example of what a quality end mill would run you. First one in the list. I would also recommend a 45 degree corner chamfer as this will have less bite and allow you to hog out sections with less "skipping" (where the tool bites in and jumps haphazardly across your workpiece).

4) Use lubricant, like a cutting agent or machining oil for cutting/drilling. This will keep your tools sharper, longer.

5) Wear a respiration mask to filter out aluminum dust when grinding (and cutting if you choose not to use a lubricant)


The design looks sound, but I am not sure you'll see much benefit from the vertical cuts through the "fins" in the water channels. I played with a similar design on a larger scale for blowmolding when I worked as a machinist and the greatest cooling differential was the size of the fins. Any cross-cuts or misshaping caused drops in flow which negated the added cooling performance the extra fluid turbulance and surface area provided.

I think you should by all means try your design, I just wanted to share my experience.

Good luck! I will be watching this thread to see how you're coming along.

To which i replied greatfully,

Thanks so much for the tips RealTime. i really appreciate it.

Im not sure which alloy im getting there either. ive seen Material: AlMg 4.5 Mn and Material: AlCuMgPb (Dreh- und Bohrqualität <-- turn and drill quality) The auction that id shown didnt say. Others ive looked at mentioned those 2 as far as i saw.

There have been several questions about the holes and such in my design and ll be prototyping my way to that final design nothing the performance at each stage.

just the fins
fins + horizontal ribs
fins + horizontal ribs + cross cuts
fins + horizontal ribs + cross cuts + vertical holes
fins + horizontal ribs + cross cuts + vertical holes + horizontal holes

fun experimentation along the way. See design thread for details.
yeah i can imagine its going to get tedious, and probably frustrating when i break a fin, or rib, or tool.

Again a really appreciate the info. Ill be sure to update along the way.

edit: testing will be interesting. Since i dont own a single piece of watercooling equipment. i either need to find someone around here who can quickly and easily swap in my block for a qucik test, or a more elaborate, bushman tech method. id imagine somthing like.

tap + thermometer -> hose -> waterblock + stove -> hose -> sink + thermometer.

other equipment includes, something to push the block down on the stove (aka die simulator) with uniform pressure. which will probably be a brick.This ought to get fun!

yes i know im putting the WBTA to shame (Though im not a member) and procoolings high standrds in block testing too. but i hope to redeem myself through innovation, creativity, and humour. And perhaps a sprinkle of enthusiasm.

[bigben2k]

That guy is 100% correct. I would have mentionned something, but it seemed too obvious.

As for testing, you're not likely to be able to measure any temperature difference with that setup: you're looking at a fraction of a degree C, under normal circumstances. What do you mean by "stove" exactly?

I would focus on finding someone to properly test it.

[DX2]

i must admit, i didnt really think about the aloy.

the tip about the dremeling by hand id already considered and resolved to just go for it. (bought a dremel today, special offer with 40 piece tool set for 8.99 Euro, 15 000 - 35 000 RPM, 135watt)

the tips about the actual machining, well yeah going for quality is always recommended. the actualy making is going tobe a minor adventure im sure. im not sure how much end milling ill be doing, as opposed to using the various stone tipped things and cutting discs. ill try get a pic up of the current drmel tool set i have. that way one can scrutinise and recommend additions. not to mention ill be implementing excellent webcam time lapse documentation of my work.

what do i mean by stove?

[BGP Spook]

I am not seeing anything but "image hosted by www.tripod.com."

[bigben2k]

It's a standard household stove.

How do you propose to use that stove? Put the block in the oven, or sit it on top of an element?

[Brians256]

Standard electric element stove. Not gas burner or glass cooktop.

[DX2]

To be honest its jut the best picture i could find in google image search.
We have a glass top stove. I recently drowned my digicam by mistake so i can't take a pic. but its a flat ceramic (looks like dark smokey glass).very much like this

How do you propose to use that stove? Put the block in the oven, or sit it on top of an element?

im pretty sure there was a hint of sarcasm in that post, coz putting the block in the oven would be funny.

I havnt thought too much about the teting, but now wrapping my mind around the problem id say.. hm..
connect tube to block and to tap, place dads digital thermometer(range -50 °C to +70°C accurate to 1/10th a degree.)somewhere before the block.
add thermal grease to block, place block on stove, weight down.
turn on water note temperature and time taken to fill jug in sink.
turn on stove , wait 5 min, note input water temp, out put water temp and time taken to fill jug.
vary input water volume, noting temp and time taken to fill jug.

asumptions: that stove heat production remains conctant, hence any difference in output temp can be attributed to the blocka ability to absorb heat from the stove.



I ordered a slab of Aluminium today.

http://cgi.ebay.de/ws/eBayISAPI.dll?...tem=7584009240
Length: 217 mm
Width: 58 mm
Height: 20 mm
Material: AlMg 4.5 Mn
Weight: approx. 0,75 kg

When i asked the seller about the hardness of the material he responded that it was medium hardness and was workable with a hand held dremel and files and such. but not so soft that it would smear.

As for the dremel, i guess the tool im missing is an end mill, they included a toothed mill. i saw end mills for approx 2 euro a pop. dont know how high or low grade it is. For 2 euro its worth a test, so ill be pickup that endmill soon and look which files they have. will update this post.

[pauldenton]

Originally Posted by DX2

Yes, all the same block of Alu.

i havnt gotten as far as radiators yet. but chances are ill want to get a good copper one that i can keep for the full setup. and if it work out well with the alu prototypes. ill make it in copper. hence, eliminating the mixing of metals.

aluminium (car) rads are pretty easy to come by if you want to use the prototype(s) long term...

[bigben2k]

There are a few problems with that setup (but I admire the clever effort).

Ask yourself:

How can you be sure that you're applying the same heat level, each time that you test? Specifically, is the knob that you're turning to turn on the burner, going to result in the same power output? How different can it be, each time you turn it on? Exactly how much power are you applying (in Watts), even roughly?

Then:

Since there's no applied mounting pressure (and I assume no TIM joint either), how do you know that you've made good contact? How do you know that you'll make the exact same contact the next time you're going to re-mount the block the next day, next week, next month?

[DX2]

i hadn't really thought the testing process through thoroughly, and didnt realise how many things i hadn't considered.

The thing i had thought about was making the test results comparable to each other. hence in the early tests i wanted to place the block on the stove , say 3 times, and inbetween reset the setup and run it from scratch. and see if the results are comparable. if not then theres really no point in continuing with the stove approach. If they are then all i really wanted to know if how various block features affect performance. eg, base thickness, fin height, holes Yes/no/number. all id want is to know if the water is being heated better or worse, in comparison to the previous block config.

For proper testing, and thus results comparable to other blocks and have REAL performance data ill surely have to seek other avenues of testing. Who knows, perhpas pHaestus. would be willing to put the block in his test setup . But its a long road till im ready for that.

[bigben2k]

The cheapest method would be to use an Intel CPU with CPUBurn, and follow the Intel specifications for cutting a groove in the IHS, to mount a type T thermocouple. A Fluke 2190A would do the job.

The problem otherwise is that you don't have an accurate measurement of the flow rate.

Yep, ask pHaestus.

[jaydee]

Build first, test later?

[JFettig]

All I can say is good luck and I hope you have a LOT of time and a LOT of patients, even then Its not even reasonable, heck, I dont know how it could be reasonable, maybe lost wax casting at most.


Jon

[jaydee]

PLEASE wear safety glasses at least. A full face shield would be recommended. Not only the copper bits flying about but those dremel wheels will sometimes fly apart. Not good to get a chunk of wheel through your eye!

I agree with Jon, but would love to see you try it. if it fails then you can move on to a different design. Good luck and keep us informed of your progress.

[JFettig]

Your cutting aluminum arent you? MAKE SURE TO WEAR A DUST MASK! Aluminum dust is VERY toxic! I myself have inhaled just a small amount and have felt really weird and gross for a day or two, I cant immagine how bad it would be if I inhaled more.


Jon

[Brians256]

Hmmm.. never noticed any bad side effects. But, I did have odd colored mucus for a day or two from nose and lungs.

[DX2]

Thanks for the concern. Will this do?


Should suffice for dust and splinters. a cutting disc hurled at 35 000 rpm will get through the mask. But as opposed to a full face mask i can wear that get up for several hours at a time.

jaydee I agree with Jon, but would love to see you try it. if it fails then you can move on to a different design. Good luck and keep us informed of your progress.

I admit my approach i far from time efficient. Yes i have alot of time, and patience (most of the time) . im just going to wing it.

JFettig All I can say is good luck and I hope you have a LOT of time and a LOT of patients.

Patience, but from the pic it could be both.

jaydee Build first, test later?

probably what i will end up doing for the first samples. then ill build and test the addition of each feature. but yeah right now im just in the mood to get going.