|
|
Water Block Design / Construction Building your own block? Need info on designing one? Heres where to do it |
Thread Tools |
04-19-2004, 09:29 AM | #1 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
#rotor block construction - pump flow and silence
Hey everyone
First time wb builder (as we all are at some stage i guess), ive gotten most of my supplies. 50x10x1500 length of copper stock stainless/plastic mix res... im hoping this will be ok, thoughts? its a stainless airtight sealing coffee style container with a clear top. Half inch OD barbs with 3/8'th threads, taps etc Solder/silicon/gasket paper? Can you get away with gasket paper with gasket paste or will there be adverse effects? Heatercore (brass/copper mix i believe) - its a smaller one, will fit either 2x 92mm fans or 1x120 - its around 135 long, 130 wide (the fin area that is). All of the above cost me less than 70nzd, which i think isnt bad This is mostly for learning - its going on an HTPC duron, basically its cheap hardware so i dont have to risk anything on my main rig until im certain it will work. I realise could more than likely get a quieter set up out of air, but im wanting to experiement for watercooling my main box, which is overclocked quite a bit and im wanting a little bit more I know that the #rotor style block will probably be for me in a lot of cases - i was thinking of going with a slight variation on this theme, with a central inlet, and a 'round' style drill pattern - im thinking about playing with jet plates etc as well - as i said, this is a learning project more than anything, and so long as it does better than air cooling im happy The other main question i have is - for this style of system (i may include another block later on) what pump would be needed? Anything other than Eheim is pretty hard to get in New Zealand. Ive done some research (of course) and found that people are saying they're outputting 35db... how much quieter can you get these pumps? I'm not wanting to hear the pump over 7v fans. Of course, im not sure how realistic this is, basically i want a higher flow pump with very little noise. Is the 1048 a step in the right direction? if anyone has a better recommendation im not above going international and having it shipped in. Of course, the cheaper the better. Excuse my long post thanks for any help. |
04-19-2004, 10:10 PM | #2 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
wow, 55 views and not one post...
Im off to buy a pump now, its a waterwerks 1000LH pump with 1.6m head, for around $32USD or $54NZD. |
04-19-2004, 10:30 PM | #3 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
A high flow, low pressure pump? You can do better...
|
04-19-2004, 10:37 PM | #4 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
hmm what do you recommend ? Im not sure on the ratio's of the pumps flow rate to head pressure, i was going on eheim pumps -
Code:
1048...Centrifugal pump Pump output 600 l/h 132 lmp. gal./h. 158 U.S. gal./h. Delivery head 1,50 m/wat.col. 4 ft. 11 in./wat.col. Power consumption 10 W Hose connection (suction side) Ø13mm (1/2") Hose connection (pressure side) Ø13mm (1/2") Dimensions: lenght x width x height 147 x 75 x 118 5.8 x 3 x 4.7 in Code:
1250...Centrifugal pump Pump output 1200 l/h 264 lmp. gal./h. 7317 U.S. gal./h. Delivery head 2,00 m/wat.col. 6 ft. 7 in./wat.col. Power consumption 28 W Hose connection (suction side) Ø17mm (3/4") Hose connection (pressure side) Ø13mm (1/2") Dimensions: lenght x width x height 180 x 95 x 120 7.1 x 3.8 x 4.7 in. |
04-19-2004, 10:41 PM | #5 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Pressure (a.k.a head) is more important than flow. I don't generally recomend Eheims, except to those who require a quality product. In this case, I'd go with the 1250, just because you're making a new block, and you may want the ability to try different flow rates. Then again, you might want to look at other pumps too...
|
04-19-2004, 10:48 PM | #6 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Sadly theres very few pump brands in New Zealand - Hydors (sp?) are really hard to come buy, pretty much eheim is the most common used brand - i cant bring myself to spend $150 on a pump though. An eheim 1046 has much less head than this pump, and from the design of a rotor style block, im imagining that it would be fairly unrestrictive?
Im a total newbie to watercooling, unfortunately - I understand each part and why they need the specs they do, just putting them together and the effects are something im still learning. The products available to us in New Zealand for watercooling are pretty restrictive (only 4.2million people, in a sport orientated country - you can imagine that watercooling isnt the number one hobby here ) http://www.overclockers.co.nz/produc...ool/pump.shtml Our overclockers website is drastically in need of an overhaul, but as you can see, theres not much choice. The res pump there is 400L/h with a head of (i think) 1.2m... the same pump i have just found in a garden store for 37.50. It was beside the pump i was looking at, quite funny to find things like that together |
04-19-2004, 10:55 PM | #7 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Yikes!
You may try OCAU and see if your "next door neighbour" can hook you up here... (assuming that import duties are insignificant). |
04-19-2004, 10:59 PM | #8 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Exactly. Its not the best situation for watercooling over here. Thanks for that, if i cant find anything that suits my needs at OCAU ill probably end up getting it anyway - im after a cheap learning experience as well as semi-decent temps, so maybe ill get both with a little luck
edit - im finding more and more pumps similar to the ones ive seen the garden shop - ie, the pondmaster pumps they sell at below-0 in brisbane, which has 1000LH flow and 1.9m head. Apparently 'ideal' for watercooling, but hey, i never trust what a site has to say Perhaps Cathar will recommend a pump since hes from aussie seems like if anyone has had experience with a wide range of pumps, its him. Last edited by Etacovda; 04-19-2004 at 11:16 PM. |
04-19-2004, 11:25 PM | #9 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Via Aqua 1300
Water turnover: up to 1100l/hr (with tap) Power consumption: 16W Filter dimensions: 11.4cmx5.5cmx8.5cm Height max:1.4m Code:
HYDROGARDEN L20 L30 L40 Q MAX (L/H) 700/185 1200/320 2800/740 H MAX (CM) 135/50 195/75 230/90 Last edited by Etacovda; 04-19-2004 at 11:31 PM. |
04-20-2004, 12:16 AM | #10 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
How about two Via Aqua 1300's?
|
04-20-2004, 01:35 AM | #11 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
What kind of head do i really need for good performance? I guess motor torque comes into it quite significantly - im assuming under restriction the pumps output can drop pretty hugely and the quality of the pump here is where it comes into play. Ive asked at the place i found the other pump if they can source me a hydor L20 or L30, I should find out within a few days, I hope. I'm wanting to be cheap, but not at the expense of too much performance or reliablity - I dont mind another couple of degrees under load, but I dont want the pump to fail too prematurely.
I guess its a risk i have to take with a cheap pump tho |
04-20-2004, 03:28 PM | #12 |
Cooling Neophyte
Join Date: Apr 2003
Location: Auckland, New Zealand
Posts: 42
|
I bought two Via Aqua 1300s, I have got them so I can easily change them from series to parallel so I can see which is best. I got one from the Switched On Gardener, slightly cheaper than Coolcases where I got my first one. They are the same except the one I got from Switched On Gardener is not earthed which I didnt particularly like, but I had already chopped the housing up!
|
04-20-2004, 06:07 PM | #13 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
I figured i may as well mash this into one thread - first revision of the wb design.
Basically, i figure im going to not make it diagonal as it is now, i was thinking of the most water flow, but have realised that it probably doesnt matter. 1> Now, its not cooling a TEC, isnt ever likely to be, so i probably dont need that many drill holes and can save myself quite a bit of time? 2> Does the symmetry of the design come into play when overclocking? im taking it that an even heat accross the core is better than a one sided heat... will the flow in either direction matter? Im thinking it probably wont, but I figured I'd ask anyway, experience is often better than presumption. 3> Will grinding out the middle four pins above the core, and having my barb sit very low in the block help cooling via extra water pressure directly onto the core? 4>To restrict flow a little bit more, ive considered when grinding through the gaps that are in a directly line with the outlet barbs only say 3/4 of the way to the base. This is illustrated by the link indicated by red spots on one of the block plan views. I think this will force the water to have a bit more turbulance before it leaves the block, but I take it that a lot less cooling happens here and removal of water is more important? Obviously im a learner and you've all heard this before, but some help would be appreciated before I destroy a perfectly good block sized piece of copper and several drill bits Ozzy - how do you find the via pump? what price are they ? are they very quiet? thanks for your help. Code:
Hydor L30 320GPH (1200 l/h) Pump - 110/120V & 1/2" OD Fittings. The high performance pumps for aquariums, ponds, flow control system, and fresh water tanks. It features a one-piece rotor assembly with filter foam suitable for biological and wet-and-dry(trickle)filter. The pump is water circulation equipped with quick-tube connectors. Retail box. Features: Delivery Head 75" (135cm) 1200L/H One last thing.... using 1/2 ID tubing, does that mean you go for 1/2 OD barbs or 5/8ths? The solidworks barbs are 1/2 with 1/2 threads. These are the barb sizes available to me - Code:
Tube Thread 1/2 3/8 1/2 1/2 1/2 3/4 5/8 3/8 5/8 1/2 5/8 3/4 Edit - reading JayDees thread below this one now, if you'd like to post and give me pointers (would be appreciated) you can but i guess i can figure it out from that Last edited by Etacovda; 04-20-2004 at 07:22 PM. |
04-20-2004, 07:30 PM | #14 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
When choosing a pump, you should get a copy of the pump's P-Q curve if possible. Look for a pump with a high head rating that keeps a fairly high flow volume over as wide a range as possible. This is FAR more important than the free flow rating. It is probably safe to bet that the average system will have between 1 and 1.5 meters of head loss so look for a pump with decent flow at that point, preferably with a max head rating of 2-3 meters minimum.
In terms of barbs, you want at least 1/2" OD barbs if you'll be using 1/2" tube (or the metric equivalent) Many folks will argue in favor of going for the next size larger barb and forcing the tube over it for a theoretically tighter fit. (Good idea with Tygon tube, not so good w/ vinyl which tends to split) either way you should get a thread that doesn't restrict the flow (1/2" npt?). Some folks even ream out the insides of their barbs to reduce their flow restriction. In terms of design, bear in mind that nearly all the cooling will occur in the area immediately over the die, and you should maximize the flow and turbulence in THAT area ONLY. Getting a jet action right onto the area of the die can be good, as is anything that increases turbulence in that area. Once you have cleared the die, you should concentrate on getting the water out of the block smoothly with as little additional flow restriction as possible. The general idea is that flow restriction roughly equates to turbulence, and you only want either where it will help you cool. Away from cooling areas, you want to maximize the flow and minimize the turbulence, thus the large bore hoses, single pass rads, and other restriction reducing design choices. Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
04-20-2004, 09:49 PM | #15 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Thanks very much Gooserider, that clarifies pretty much everything I need to know.
Out of interest, what pumps are everyone using? After reading spec's from here there and everywhere, theres not that many pumps that people are actually using with that kind of head? Last edited by Etacovda; 04-20-2004 at 09:57 PM. |
04-20-2004, 11:35 PM | #16 |
Cooling Savant
Join Date: Feb 2002
Location: Dione, sector 4s1256
Posts: 852
|
that attached picture looks like it has the pins over the core area removed.... why is that?
__________________
There is no Spoon.... |
04-20-2004, 11:52 PM | #17 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Actually the plan was to make them a bit lower, so i could get the barb closer to the core. They're not meant to be entirely removed, just taken down to around 1mm in height - my barb will be able to get a lot closer to the core that way, as im thinking of adding a small jet assembly to it (barb > brass end cap with holes drilled - of course, this is totally still in development, im still sourcing parts)- ive actually had a completely different thought for the block which I will solidmodel and put up here tonight (which is in about 4 hours here )
Is getting the jet plate closer a good idea or am i just off my nut ? Also, thanks for designing such a newbie friendly block rotor, im sure heaps of people (me included, of course) appreciate it |
04-21-2004, 08:32 PM | #18 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
#Rotor has far more insight into how his block works than I do, but my understanding is that the rotor block is not designed to use jet impingement. The Rotor method and the Jet impingenment method are two different approaches to the cooling problem and aren't really intended to be combined...
At least thats the way I've been led to understand it Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
04-22-2004, 12:37 AM | #19 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
hah, yup i gathered as much, but since i dont really have access to a mill, the drill press method is more through neccessity than design.\
I do, however, have a brother who works at SKF bearings here, so I will look into getting side load bearings for my drill press and give it a shot - im not after 100% accuracy here (obviously, the vibration from a drill press is far more than that of a decent mill) but it will allow me to do flat bottomed channels with a mill bit. My understanding of rotor's design is that of a tumbling turbulence - the water 'rolls' accross the bottom of the inverted cone shaped holes that the drill bit has left, up the sides, hitting the ridge, and rolls back down the other side etc. The pins help in removing the head by adding extra surface area? The rolling allows more water molecules to touch the surface of the copper and adsorb the energy. Well, thats my theory at any rate The pump I am using is very similar to an ehiem 1048 in pHaetus's graph of pump flow vs. head (I think, cant remember the exact title of the graph). |
04-22-2004, 01:19 PM | #20 |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
The standard #Rotor design is intended to be made w/ a drill press and a dremel, that is one of it's biggest virtues, that it doesn't require a mill, so it is easy for a minimally equipped user to build.
Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
04-22-2004, 11:48 PM | #21 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
thanks very much for your help goose, its really appreciated.
Im still working on the design, looking at a pressure type cooling situation still. Im not sure how it will cool, but hell, ive got enough copper and enough time to experiment, and thats what this is all about right? http://www.brendons.orcon.net.nz/wbrev2.jpg edit- this is an older version, ive got two more channels heading to each outlet now. Also, the centre hole is drilled to the depth of the other ones now. All this design requires is a 3mm drill bit, a 10mm and a 1/2". And a dremel, of course. |
04-23-2004, 07:12 PM | #22 | |
Cooling Savant
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
|
Quote:
I believe it works because as the water flows down each little channel between the pins it collides with the water coming down the cross channels, creating major turbulence at each intersection point. The conical hole bottoms are actually a bit of a mixed bag if I remember some of the stuff I've read. A flat bottom hole (such as you'd get with an end mill) is slightly better in theory, but the difference is so small it isn't worth while to worry about, especially given how much harder it is to get a clean hole w/ a mill bit (they tend to wander) Maybe #Rotor can clarify for us. Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c |
|
04-27-2004, 08:57 AM | #23 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Well, i got another heatercore today, so thats a bonus - its half inch barbed, and a better design with more fins, so hopefully it'll be more efficient. Cleared it out with the new pump I also got today, 800lph, 1.6m head pump. total cost of pump and rad was 65nzd, 40US or 27ish BP.
This is the design I ended up going with - www.brendons.orcon.net.nz/block.jpg And this is a rather bad scan of the first run at drilling it. My very cheap 100NZD drill press did NOT like this much, I can tell you its definately a budget drill. www.brendons.orcon.net.nz/blocky.jpg Well, I should have everything else set up shortly, should be interesting to see how it performs. And yes, i was in a hurry ill be dremeling the gaps tomorrow, next time i do it (if this block performs well) i'll be making sure im a bit more accurate. |
04-28-2004, 01:05 AM | #24 |
Cooling Neophyte
Join Date: Mar 2004
Location: Auckland, New Zealand
Posts: 50
|
hi, first.. woohoo another NZ'er
i was just wondering where you got your copper from and how much it was ? also, for a pump, i dunno if there still is but there was someone on trademe.co.nz who sold pumps, theyre ahh Resun brand but i got one, it was $65NZD and i havent had any problems with it.. well actually i have, i broke the impeller thing off it but that was my fault (its not good to drop your pump ) its ahh 1000LPH and got a 2m head. but since you got one it dosnt really matter now. |
04-28-2004, 07:26 AM | #25 |
Cooling Savant
Join Date: Feb 2004
Location: Dunedin NZ
Posts: 735
|
Yup, NZ'er here
How loud was your pump? Silence is the main concern - id rather pay 150 on a silent eheim than 80 on a no name brand pump thats loud. I got my copper second hand at a scrap metal dealers, total steal and a good find at that. |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|