custom tanks

[dean.collins]

I've been using a reserator for about a week now and have been really impressed with how easy these things are to use.

I'm cooling the cpu, video card and hard drives (using a dual silent star box).
(now looking for a northbridge cooler and also looking for water cooled power supplies)

Unfortunately i have too many pc's in my 19" rack to cool each of them with a reserator so I'm thinking about having a custom tank built that will mount on the side of my 19" rack.

The dimensions of the tank will be 1" x 26" x 60"

The cubic volume of the tank will be about 25 liters plus the piping etc.

This compares with the Zalman reserator at 2.5 liters.

I'll probably be cooling 4 pc's (using 4 independant pumps) thinking of using submersible pumps but then I'd need to drain each time I replace a pump so maybe not worth it - might be better to mount them on the outside with a stopcock feeding each pump about every 7 inches apart for each pc (they are mounted in 4 ru cases).

Has anyone seen something like this before?

TIA,
Dean

[dean.collins]

what are zalman resrators made from?

I've been reading about galvanic corrosion but dont know what material I should be specifying these tanks to be made from.

also is there any reason why no one replied or do most people here no own 19" racks so it's an unusual design.

Cheers,
Dean

[bobkoure]

Originally Posted by dean.collins

what are zalman resrators made from?

Aluminum

I've been reading about galvanic corrosion but don't know what material I should be specifying these tanks to be made from.

If you use aluminum, be sure to use a corrosion inhibitor in your coolant. If you're having a tank built, why not use brass or copper sheet? I'd guess you'd be better off with fins folded from the tank material rather than soldered on (solder's pretty bad for heat transfer)

also is there any reason why no one replied or do most people here no own 19" racks so it's an unusual design.

Most folks here do not use reserators (AFAIK). It's tricky and we mostly know we don't know. For instance, air flowing up the side of your tank as it warms will likely be moving very slowly, which means it will probably stay in a thin layer alongside the tank ("laminar" flow). How important is this? I dunno - maybe a lot.
I keep thinking that the right passive design would look a lot like those pipe and fin arrangements in hot-water-heat registers (but wider fin spacing - that stuff is for a 30C temp differential).

[dean.collins]

why corrosion inhibitor, the zalman manual just specifies distilled water and no addidtives, is this because they are anodised?

if I have a tank built from marine aluminium but continue to use the same cpu / gpu blocks?

Dean

[bobkoure]

Because aluminum will corrode, even in the absence of an electrolytic loop. You could have the surfaces anodized (essentially boiled in acid and then in a filler substance - usually colored). There will be corrosion anywhere the anodizing didn't cover (or gets worn/scratched away from).

The zalman manual didn't specify anti-corrosion because, well, it was wrong (and the anodizing will probably keep things just fine in the short term). Do a search here (or over at the Silent PC review watercooling forum ) for reserator and corrosion - or simply post a question here like "reserator OK without corrosion inhibitor?". I think you might have to go to the silent forum for folks who have actually used 'em without corrosion inhibitors - and either were fine or suffered some consequences - go see if it's OK.

BTW, the warmer your coolant is the more effectively a passive radiator will be able to shed heat. All pumps add heat to the loop (friction, mostly) so you want to stay with as low powered a pump as you can (there's a good discussion in this thread about pumps, heat, and radiators).
Both of these factors mean that you are going to want waterblocks that are as efficient as possible (expressed as (c/w)/flow) to keep your CPUs as close to coolant temp as possible for the least amount of flow/pressure. Currently, that's the Swiftech 6000 series - and these are all copper (or maybe some brass sheet in the top).

You could also have a look at Airspirit's posts here. He's watercooling a number of PCs (don't remember if they were in a rack) with a single cooling system. He's gone the high-performance, high-noise route, but you may still find useful information...

[Brians256]

Originally Posted by dean.collins

what are zalman resrators made from?

also is there any reason why no one replied or do most people here no own 19" racks so it's an unusual design.

Cheers,
Dean

I just bought a small 19" rack for mounting equipment, but I feel that I'm fairly unusual. Other than the IT/IS professionals, it cannot be common to put rackmounted hardware into your home. I'm doing it for structured wiring in my house, not really for any heavy duty computing (although I hope to put a media server in my garage eventually).

Also, it seems that many people here are more interested in fan-based systems than in pure passive. Passive systems are almost non-portable because of the bulk required to make an effective cooling system for a high-powered system.

However, for passive systems, I think it was Bladerunner that made the first REALLY effective passive cooling system. He buried a tank in his back yard and had that as his heat exchanger. It's not that it isn't done. It just isn't common to make that kind of space/noise tradeoff.

The people that are focused on noise reduction tend to congregate at SilentPcReview primarily, but there is a decent amount of exchange.

The reserator product also didn't do well in these forums (I think) because it is so expensive for what it does. The product *should* cost more like $100, not $200. It is just a transmission cooler with a small pump. Build your own and save 50%, eh? Put a nearly silent fan on it and get drastically increased performance at almost no cost in terms of noise. You'd be surprised at how much difference a little bit of air movement can make.

[bobkoure]

Originally Posted by Brians256

You'd be surprised at how much difference a little bit of air movement can make.

Not surprising at all if you think about laminar flows. Ever left your unwashed car parked in a rainstorm - and wondered why it was still dirty afterwards, even all that rain ran down its sides? Same thing - only the warming air is washing upwards...

[BillA]

no Bob, here in so cal it DOES rain soot
lovely place

[bobkoure]

Sounds just wonderful
Of course, our Boston temps (for the last month or so) have been lower than the ones in Anchorage...

[DrCR]

Hey dean. I was going to do something similar by using an large uninsulated direct hot water cylinder (link), but I never did track one down (though I didn't try too hard).



The setup was also going to have heatercores so I could detach the tank and go to a lan party or something. But budget said a new rig would have to wait another year so the project was shot down. A bummer too cause G4 Storm at the time was I believe about $52. I did get my hands on a MD20Z before the cut off though so I'm not too upset.

Let me/us know if you ever get the setup you described together!

DrCR

__________

[dean.collins]

a bit hard to stick into the side of the 19" rack but I get what you mean.

I was working with those dimensions because it would bolt perfectly into my rack without even needing to drill if the 4 ears were located correctly.

[maxSaleen]

If you are looking for passive cooling in a fashion similar to the reserator, head over to ebay and search for "finned transmission coolers". You can get your hands on decent 18 inch single pass models for around $45 last time I checked.

Four of these will run you $180... with shipping... say $200. For the same price as a reserator you will get much better cooling (I cannot back this with empirical data as I'm too lazy to run a controled experiment). The difference is that you get more internal surface area, as there are fins inside the tubes, and more external surface area (rough estimate).

Your block selection will be a different story. Pure copper+pure aluminum transmission cooler+distilled water=a dead loop in no time. You could have the internals of whatever blocks you have gold plated, making them inert. I have no idea what this would cost you, though.

Another option is making your own radiator. Try this: head on over to home depot and pick up some 1/4" Cu tubing (buy 1 coil 50ft in length and split it into three equal lengths). Wrap it around a few 2" pipes (of whatever material you can find) to achieve this design:

The inlets and outlets for the coils can simply be epoxied into place in the reserviors. More seperate lengths of copper will yield lower flow resistance. Sorry for the crumby image but I put the design together in about a minute. If you need clarification just ask.

[Brians256]

Just add a corrosion inhibitor, maxSaleen.

I just checked ebay, and there are many different models available. If you wanted single pass finned transmission coolers, you can get 12" models in the mid $30 range, and double pass for $49. There are 18 inch, 24 inch and 30 inch models, too.