Water cooled audio circuitry

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In case the passive heat sinks don't work out for you (I know it turned into a nightmare for me), the heat exchanger is "A" shaped, open at the bottom, closed sides, with each side (the actual coils) being roughly 18" square by 2" thick. Bear in mind that there are several variables that you can play with--water flow rate, size of the coils, forced air, etc. I'm still not using forced air, and I can run the Aleph as long as I want with heat sink temperatures staying about the temperature of a hot shower; even the devices themselves are touchable. Given that my system runs as cool as it does, there's clearly room to 'downsize' one or more of the elements in the system.
And it was cheap. Not counting the vinyl hose to carry the water, the current version cost me less than $100.
I intend to run the send pair of Aleph 2s off of the same heat sinks once I get them running. Still short on time to get the boards etched, but I'll get there, sooner or later.


[Edited by GRollins on 06-21-2001 at 09:49 AM]
Now you got me looking at aquariums in a new way.

Ding, can your formula be adjusted to reflect using the devices mounted to 1/2" glass of a 150 gal aquarium? I looked at the P/D values of some of the 220 & 247 devices and they can run to 150w each. You mentioned 10w amp w/PD of 70w's, is that to be expected of the Aleph 2's power disipation that has Grey's running at 110 degrees? Is the water temp one of the variables derived from your models?

Grey, what type of mounting are you using? How are your to220's mounted to your heatsink? Are you using thermal compound, tight fit? Could something like a aluminum or copper heater core be used?
rol1 said:
Now you got me looking at aquariums in a new way.

Ding, can your formula be adjusted to reflect using the devices mounted to 1/2" glass of a 150 gal aquarium? I looked at the P/D values of some of the 220 & 247 devices and they can run to 150w each. You mentioned 10w amp w/PD of 70w's, is that to be expected of the Aleph 2's power disipation that has Grey's running at 110 degrees? Is the water temp one of the variables derived from your models?

Grey, what type of mounting are you using? How are your to220's mounted to your heatsink? Are you using thermal compound, tight fit? Could something like a aluminum or copper heater core be used?

I wouldn't try to mount the devices to 1/2" glass. You can't drill/tap it, as fish tanks normally use tempered glass. All you can do is clamp the devices to the glass. Also, I think you'd find that glass has too much thermal resistance to work well. Finally, I have a fish tank (salt water), and I know that sooner or later, everything around it gets wet. This wouldn't be too good for an amp.

As to using the fish tank as a water-to-air heat exchanger, I think you'd end up cooking the fish, and fouling the heatsink. The suggested size for a fish tank heater is 2-3W per gallon of water. Also, running fish tank water through the heatsink would create a nice warm place for stuff to grow.

This reminds me, ALWAYS use antifreeze in a water-cooled heatsink. It kills the nasty things that like to grow, and keeps corrosion down.

As to the water-to-air heat exchanger, Grey's original post said he was using an 'A' shaped element from a heat-pump. Based on his description, I would guess that it's the part from inside the furnace (as opposed to the part from outside the house). These units are typically designed to transfer tens of thousands of BTUs from FREON to forced air.

In his current implementation, Grey's only trying to disipate ~700W. This is so low that a fan isn't required (convection is adequate). Once he gets the second set of amps hooked up, I would expect the temperature to rise almost to the normal running temp for Class-A, STILL WITHOUT A FAN.

If, due to space or parts limitations, you'd like to use something smaller, a heater core, a transmission/oil cooler, or even a radiator, would all work OK. All of these devices are designed to transfer heat from a liquid to air. If using a smaller device, you may need a fan for it. Also, a heater core is normally rather thick, with lots of resistance to airflow. A fan would almost certainly need to be used on this.

As an example, I feel that Grey could probably run one channel of his amp to a good transmission cooler, without a fan.

Finally, I like a margin of safety whenever I deal with cooling issues. For this reason, I would suggest that a cooling fan ALWAYS be used, and that it's power go through a thermal switch mounted to the outlet side of the water-to-air heat exchanger. This would allow the fan to do nothing, until it's needed, but it would also keep the amp from frying itself, on a HOT day.
I dunno...just make sure that it's flounder that are in the tank, and supper will be ready at the end of your listening session...
I agree with Thoth that the thermal resistance of glass is probably too high for mounting semiconductors to.
As for mounting--nothing special, just the standard stuff. Device/heat sink grease/mica insulator/heat sink grease/1/4" x 1" copper bar soldered (silver solder) edgeways to 1/2" copper pipe, like this: --O Water connections at each end. Piece of cake, once I got the conceptual bugs worked out.
I considered not using mica insulators, as that would improve heat transfer even further and the drains are all at the same voltage, but decided that the upper and lower banks of output devices might see a connection *via the water* once enough copper ions got dissolved in the water. Distilled water doesn't conduct all that well, but water with stuff dissolved in it conducts. If I were to use a separate circulation channel for each bank, I could return to this idea, but the increase in thermal efficiency isn't necessary.
I'm not using anti-freeze, as there's no corrosion problem with copper. The pipe I'm using is standard water pipe from Lowe's. The heat exchanger tubes are copper also (yes, the fins are aluminum, but they're just press-fitted over 3/8" [I'm working from memory on that diameter] thin-wall Cu tubing). The only things that are in contact with the water are copper, brass (the barb fittings), and braided vinyl hose, none of which have any real trouble with water. Mark (mefinnis) expressed reservations about the aluminum I was using in an earlier incarnation; he, like you, was thinking that anti-freeze would prevent corrosion. I'm not using the aluminum in this version, so I'm not worried.
Yes, you're correct, the A-shaped heat exchanger is from the air handler (the box that is inside the house).
I've got a fan or two handy if I should need them, but so far, there's been no need. I haven't etched the boards for the second Aleph yet, so I don't have any idea on whether I'll need a fan if I use the same heat sink for two amps. I will report as soon as I get that going.

About the 12 gallons of water to cool the mosfets...
Even though the thermodynamic calculations look alright within the limits of the approximations that were made, shouldn't ANY cooling system be designed to reach rapid equilibrium temperature and then remain at that temperature?
The tank idea is that of an "overcooled" system. Even without considering the thermal gradient in different regions of the tank (if the water is not stirred), T would constantly be changing, and with that the bias current, DC offset and everything else. Make sense??
As for a tank, convection currents would mix the water pretty decently, although equilibrium might take a while to reach. How long it would take to get there would be hard to estimate, given various incidental losses, but it would eventually stabilize.
In my case, I'm running roughly five US gallons (separate and distinct from Imperial gallons, something I got into a debate about when I was visiting Scotland a few years ago). The system itself holds approximately three gallons, with the remaining two sitting in a reservoir. With my system, thermal equilibrium is reached in about 30-45 minutes. A 12 gallon system--I don't know, make it an hour or hour and a half. In a non-forced circulation system, it might take a bit longer, although convection should do a pretty good job.


I thought that we decided that fish tanks wouldn't work. In any case, check the surface area of the tank, and it's ability to radiate heat. I think you'll find that you need something else to get rid of the heat.

Rol1 was talking about 1/2" glass. This thickness is commonly used only on large tanks (100-120 gallons; larger glass tanks are hard to find). This size of tank might have adequate surface area to work as a radiator, but it would be too hot for fish to live in. If you can get the heat into the tank, it should allow the heat to leave, but why waste a perfectly good fish tank on this?

Grataku mentioned a 12 gallon tank. The surface area of this is so small that, without an external radiator, it would soon be boiling.

You mentioned a 'non-forced circulation system'. Were you refering to a tank setup, or a heatsink/radiator setup, like you have?

You also mentioned that you used deionized water, and so you didn't expect to have corrosion problems. The problem with deionized water is that it tries to suck ions out of whatever it touches (in your case, copper, brass, vinyl and, I think, solder), until it reaches equilibrium. Pure water has a resistance of 18.2M Ohms/cm (properly stated, it's conductance is 0.055 umho/cm). If you check the resistance of your current water, I think you'll find that it's dropped considerably.

Good luck.
Sorry for the confusion. If I don't start writing more clearly, they'll revoke my author's license...
As for fish tanks, to me the concept can be split into two sub-possibilities (not including a way to cook supper...a third possibility):
1) Mounting the devices to the glass of the tank itself. You and I are in agreement that glass, being a poor conductor of heat, not to mention hard to work with (carbide bits are available to drill glass, but why go to the bother?), would not be the best way to go.
2) Using the tank as a large radiator. Again, we are hampered by glass not being a good conductor, but don't overlook the evaporative possibilities. How many gallons (and, by extension, how large a tank) would be needed? Hell, I dunno. But it would work if it were large enough. Sooner or later you'd find a size where the system would radiate enough heat to get the job done. In my case, running Alephs, the fish would be pretty thoroughly cooked. Someone could, in theory, run a class B (yuk) or AB amp and perhaps make some species of tropical fish very happy, indeed. However, when all is said and done, you and I are in agreement that it would be an inefficient & wasteful way to go at things. I was just treating it as a purely theoretical exercise in thermodynamics. I am *not* an expert on fish. What temperatures fish like, and how many watts it takes to heat a tank to that temperature...I'm at your mercy. I was just thinking of an empty (of fish) tank.
My clarity, or lack thereof, reached its nadir when I was speaking about non-forced circulation. Again, apologies. I must be shorter on sleep than I thought. In that case, I was imagining a fish tank with the devices attached to the tank. Set aside concerns about thermal conductivity of glass for the moment--as we've agreed, that's going to be a major factor against us. I was just thinking of convection within the tank itself carrying the heat from the devices and radiating it through the walls (the glass again), plus evaporative losses. If it weren't for the low conductivity of the glass, we'd have it knocked.
I believe I have that part knocked.
Diamond is a near perfect conductor of heat (infrared). Now, don't go bothering me about price...or about how to get the devices mounted. Just think how well you could see the fish! And the thermal properties would be just wonderful.
Remember, you heard it here first...
Um, I don't recall saying that I *used* deionized/distilled water--I didn't. Just tap water. (Which around here means that the water is equivalent to a dead short, as it's so hard [lotsa dissolved ions already] that you could use it for a hammer.) I didn't bother measuring the resistance/conductivity of my water--I just put in the mica insulators. Cheap insurance. Someone with better water, or who was willing to use distilled water, might not need to worry about it. Like I said, I'm getting rid of plenty of heat as it is. Running without insulators was just a possibility.
Once I get the Alephs reconfigured and a second pair built (Real Soon Now...promise), I'll have them connected via copper elbows and pipe in a few more places where there is only vinyl hose right now. At that point, the mica will be necessary anyway.
Hopefully, that's clearer. If not, prod me again, and I'll take another whack at it.

Amp cooling problems solved!

Why have I not read about these before?

1. Make sure ALL joints are soldered. No connectors at all.
2. Spray coat against possible water bridges.
3. Stick the amps in a small bar ****.


If you live in a house, put the amps in the crawl space under, drilling holes in the floor where needed for cables. Cool them with one of those $12.99, 20" box fans. Mucho air flow, mucho cooling, mucho noise free, mucho cheap.

Don't laugh.
The only problem with crawlspace air is that it smells stale & musty. (Though perhaps a return duct back to the crawlspace would cure that problem...) In my current house, I don't even have a crawl space; in my previous house I did.
Now one thing I think I mentioned somewhere regarding this water-cooled thing (it may not be in this thread) is the possibility of putting the heat exchanger in the crawlspace to get cooler air flow (same thing you're describing) and to put the heat somewhere other than your living space. Obviously, this would require holes to get the water hoses through the floor. However...in my previous house (the one with the crawlspace), the heating/air vent right next to my stereo was so poorly cut that I could get cables through next to the boot for the air. I don't remember whether there was enough room for 5/8" line such as I'm using now, but perhaps several smaller lines running in parallel would do the trick.
The crawlspace is just so much wasted space in most houses. I agree that it could be put to good use.

I don't think amps have olfactory receptors so the musty smell shouldn't be a problem for them. Crawlspaces are usuall cool in the summer and temperate in the winter and there is no limit to the amount of forced air cooling you can provide to the amps. And the engineering problems are nil; oops, sorry, takes all the fun out of it, doesn't it?

Crawlspaces have other uses as well.

I buried a 2" diameter steel post in concrete in my crawlspace. The post comes up through, without touching, a hole in the floor. Atop is mounted a platform for my turntable. Instant isolation from structure born vibrations! Works like a charm. Damn, takes all the fun out of it again.

My woofer system uses 1, 10cu.ft. enclosure per channel. Good grief, talk about obtrusive in the living room!
My solution was to cut 2 rectangular holes in the floor and drop the major portion of the enclosure into the crawlspace. The results are a much smaller visible portion of the enclosures and any remaining vibration from the enclosure walls serves to entertain the mice when the cats are on vacation.
I had a similar solution for turntable problems in my old house. I put a cinder block on the ground, then jacked up the floor about 1/16" or so, slid a 4x4" post underneath the joist where the turntable was, then let the floor back down. Low tech, cheap, and effective. (I always like the cheap part.)
Your solution about putting part of the woofer enclosures under the floor reminds me of the folks who've built the entire enclosure under the floor. Horn guys, in particular, are infamous for putting *huge* horns under the floor. Transmission lines work well there, too.
I once built a transmission line into an unused chimney (a different house). None of this quarter-wave. This was going to be a half-wave. The downfall was that I never got the thing stuffed, as it was a rental property, and I wasn't sure I'd be able to get all the stuffing back out when I left; I lived there for less than a year. Without the stuffing, the sound was less than inspiring. Still, those of you with unused chimneys (or crawlspaces) might be able to use these ideas.

I currently have the water cooling system broken down as part of the ongoing process of pulling my stereo up by its bootstraps. The second Aleph is pretty much ready to go in, but I decided that I'd tidy things up a bit while I had the various pieces pulled apart. In particular, the heat exchanger was rusty. Not from using it as a heat sink; it was rusty from its previous life as a heat pump. The condensation during the summer tends to rust these things and they just use a cheap galvanized frame.
While I was getting the rust off, I discovered that the heat exchanger is stamped:
COLEMAN 6904-501
or at least that's what I think it says, the rust makes it difficult to tell. Since the major variable in someone being able to analyze or reproduce what I've put together here is the thermal capability of the heat exchanger, I thought this might help.
If I come across any more information, I'll post it.

you might also want to think about adding nitrates to your cooling systems if you ar using a closed loop system, little critters live in the water and really thrive in warm water wich after a while will leave you with a system full of slime and leaks. hard water will eat the copper to pieces in a short time as will electrolisis(exscuse my spelling if i got that one wrong) i would use deionized water, antifreeze or a coolant that is made for closed loop systems. a valved vent at the highest point in the system might also be a good idea to remove any trapped air from the system. you will also need a good (clean, spill proof) way to add water to this from time to time. if you use two different types of metal in your systems they should be isolated from each other as not all metal like each other and they can start to corrode badly, use a short piece of tubbing or di-electric uniouns.
a radiator with an electric fan from a small car would make a good heat exchanger that could be used with or with the fan. add a thermal switch in the cooling line to turn the fan on only when its needed. another good heat exchanger would be a heater core from a car or truck, most of these are copper and if treated right will last a long time.
remember that if your going to be sweating(soldering) copper fittings to clean the pipe and the fittings with sand paper and then apply flux before you sweat them. yall probably already know this but i just thought i'd throw it in there.


Good advice on a cooling system.

Years ago I built a water cooled system for an Elektor Crescendo in deep AB. I used a slim 2U case, an aluminium welded channel section with a large flange for the semis, and a purge pump from a dishwasher. A level sensor detected water shortages and switched a solenoid tap from the mains.

For heat exchanger I buried a ten foot length of black plastic pipe about 6" underground. This proved entirely adequate over the three years it was operating. While you are correct about critters living in the warmish water, in practice I had no problem as evaporation from the reservoir (just one pint) replenished the water regularly. The only problem I ever had was the noise of the dishwasher pump - a shaded pole mains affair - which although it was 20 feet from the amp and outside the house could still be heard. But it was reliable.



ok heres a kind of off the wall idea. you build one of those indoor water fall things, ya know the ones where the water runs down a bunch of rocks and stuff. then you connect it to your amps and pump and pump the heated water out the top of it so it cools as it runs down the rocks. but the best part it that you can tell your wife that your building it for her and she'll just be thrilled.


More Ideas for watercooling:
Motorcycle oil/watercoolers, comes in a lot of shapes and sizes

aquarium pump or the kind of circulation pumps you got in heating system for houses they are very quiet

Put the pump, cooler and some quiet (papst?) temperature regulated fans in a dampend tower computer chassi

Wild idea: Maybe a microphone controlled fan with electronic braking when it gets silent (got to try that out)
water cooling - BravO


I am new here so I missed most of the earlier posts. I think water cooling is a great idea. I used to work for a company that built industrial power supplies. From that experience I know that if you do it right you can dissipate twice as much heat with a given device with liquid cooling as opposed to air cooling. I thot about building a water cooled Zen amp, but scrapped the idea when I found a large heat sink. Here are a couple ideas I had left over from that experience. 1) Go to a hardware or hobby store and get some rectangular brass stock. This stuff is hollow. You can run water thru the inside and clamp the transistors to the outside (somehow). The problem here is getting from rectangular stock to tubing. Maybe you could cut a slot in the side of a tube and stick the rectangular stock in that. Or maybe you could block of the ends of the stock and drill holes in the stock and solder tubing into that. (I hope that makes sense.) 2) Take a piece of tubing a convienient size, a piece of wood and a hammer, hold the wood next to the tube and whack the wood with the hammer so as to make a flat spot on the tube. Somehow affix the transistor to the flat spot.
Hot tub, eh?
You must be from California.
Them things is illegal in Alabama, Georgia, and six counties in South Carolina. Contributes to a decline in morals, ya see. Least-ways, that's what the Southern Baptists claim...
Just kidding.
I hope...
(You don't know what it's like to live with those people. They've tried to stop just about everything else.)
As most of the suggestions in the last few posts were covered earlier in the thread, I'll skip on to the waterfall...which kinda tickles my fancy. There's a house up in North Carolina with a (natural) waterfall in the living room. Yes, I know about Falling Water (famous Frank Lloyd Wright design for those who aren't familiar with it). Been there, in fact. No, I don't intend to mention it to the wife-unit. She might want one.
I thought aquarium pumps pumped air. (Shows how much I know about keeping fish.) Do they have some that pump water? I'd think--could be wrong--that the flow rate would be a bit low for this kind of thing.
I am open to suggestions for pumps. That's my brewing pump in the system now. I'd like to recover it for its intended purpose, replacing it with something else. Qualifications should be obvious: cheap, quiet, reliable (i.e. high duty cycle), decent flow rate. Hugh's dishwasher pump might be a good idea except for the noise he mentions, although if it's pulled out of a defunct dishwasher it might be had for free--an attractive point.

...<i>Hot tub?</i> That man's trying to get me arrested.
Guys, remind me not to do any more Mini-A circuits, I don't think it went over well with Nelson...

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.