Amplifier water cooling project!

Status
Not open for further replies.
Disabled Account
Joined 2003
If you're interested, here's how I'm putting my water cooling into practise. Cheap :nod:

I bought a (my95-97 or something) Nissan Micra's heater element for 15€ (15$ or so) and i'm going to use it as a radiator. The whole system is gonna be cheaper and more effective than those computer water cooling kits (I think :scratch: ). The class A amp I'm building dissipates around 60-80 watts. The radiator doesn't heat up at all when it's heated with a ½ kilowatt hairdryer, so it should be effective enough.

I made a water cooling block from 45x45mm steel profile. The end parts are 5mm aluminium and are going to be glued in the ends of the steel part. The copper pipe(s) are going to go through the alu end parts, which isn't exactly good because of the corrosion 🙄 I'll make some holes in the copper pipe(s) so that the water spreads evenly in the steel block. Pvc tubes are then connected to the copper pipes and the water is pumped around by a 14€ garden pump (you might think that it's noisy but it isn't. Amazingly silent :spin: ).

Ok. here's the block. More pics click the links.

An externally hosted image should be here but it was not working when we last tested it.


http://kotisivu.dnainternet.net/esmatak/coolingblock2.jpg

http://kotisivu.dnainternet.net/esmatak/radiator1.jpg

http://kotisivu.dnainternet.net/esmatak/radiator2.jpg
 
Yep, convection. If needed, I'll put a 80x80 fan in it. But it won't be necessary. The system is going to be tested tomorrow, when all the parts are together.. I'll heat up the block with a pencil torch and see how the fluid cools down.
 
Test went ok.. I tried to warm the block up with a 1200 degrees (or something) bottled gas torch, and the result was that the block heated to something around 50 degrees celcius (I think). Dindn't go any further.

Unfortunately, the water tank is made of plexiglass and didn't stand the stress of the tube joints which I screwed in.. :bawling: Must do a new one.

Some pictures are coming when the system is running again :headbash:
 
Had a little break.. 😱 Now going on again, if someone is still interested 🙄

Now all the components of the system are tested to fit in the case and I'm finally welding the amp itself. (My homepage is full of stuff so I put the pics straight here until I manage to find more space):
 

Attachments

  • aclass1b.jpg
    aclass1b.jpg
    39.4 KB · Views: 485
Sellaista todella yksinkertaista (tuntuu välillä, että itse vahvistin on sivuosassa tässä projektissa 😉 )

kytkissivu

Suunnittelija on Petri Pykilä, ja osat hain Tricompista. Jäähdytyksen hinta lienee syyläristä ja vesiblokin kalleudesta riippuen jotain 15(syyläri) + 14(pumppu bauhausista) + 10 € = about 40€.
Täytynee tosiaan laittaa kaikki kuvat ja jutut jossain vaiheessa kotisivulleni jos tilaa löytyy..

Millä nimellä muuten liikuskelet Hifilehden foorumilla?
 
Yes I will write in english also when necessary 😀

BTW, I managed to test it few minutes ago.. It works! :bigeyes:

I quickly measured 0,01 volts (R) and 0,018 volts (L) DC-offset (or whatever it's called), does this sound normal?

Man, does it need water cooling.. Fets heat up in couple of minutes quite hot, I had them installed on the cooling block only and no water going trough yet.
 
Now there's some water circulating! It has played music for about half an hour and I (roughly) measure that the radiator is about 40 degrees celcius, cooling block about 45-50 C and FETs around 55 C, maybe a bit more.

Cold sweat on my forehead waiting when it starts to leak! :bawling: :dead:

Test setup:
 

Attachments

  • aclass3b.jpg
    aclass3b.jpg
    61.7 KB · Views: 366
Hi Otherwise,

The cooling system looks very good. I really like it. One question though, is it a totally enclosed system? If it is then IMHO there may be some issues regarding the choice of medium and the way how the supply tank is assembled. You may like to consider chemical reactions or gas build up or water turning green in the system. How about installing a pressure relief in the tank. I do not think there is a need for a pressurised system in order to lower the boiling point in this kind of application. No pressure means less chance for leaks.

Good work.
Regards
Chris
 
Chris ma:

It's not pressurized. Good point anyway! Mostly I fear water escaping through the threads of the FETs' fastening screws. You see, the screws must go through the block, because it's only 2mm thick. I've put the screw holes full of silicon heat transfer paste and coated the screw heads with paint. It seems to be OK for now.

And corrosion of that copper/alu/steel combination.. We'll see :scratch:
 
Hi,

Brass screws, teflon tape and plumbing grease were the tools I used when I made my wet/dry filter for my fish tank. I am using two submersible pumps rated at 800gph with no leaks, although my system is not pressurized but I have to deal with 6 foot gravity pressure of the water column of 2 inch diameter against the power of the two pumps combines.

One other thing to consider is a way to make sure no air bubbles trapped in the core of the radiator otherwise the effeciency will be degraded.

Chris
 
Otherwise said:
Chris ma:

fastening screws. You see, the screws must go through the block, because it's only 2mm thick. I've put the screw holes full of silicon heat transfer paste and coated the screw heads with paint. It seems to be OK for now.

And corrosion of that copper/alu/steel combination.. We'll see :scratch:


The other way may be you can attach a small but thick piece of stock on the inside of the block first (use screws and contact cement as gasket), so that you can tap a blind hole for mounting the chip on the other side.

Chris

PS. There is no need to drain the system when changing chips
 
If you are going to have the radiator in the same box as the amp, you might as well attach the transistors directly to the radiator.
Still might need the circulation pump , just to circulate the fluid in the radiator itself. It might work without a pump if you
tilted it slightly so the "tank" parts of the radiator are on the top and bottom and the transistors were attached to the bottom tank. Then the warm water would rise through one set of tubes, and return, cooler on the lower set. Probably you would need to solder thicker copper to the tank for mounting. and the soldering might mess up the radiator, but that is the only downside. One advantage of the water coooling is that you can have a remote radiator that is big and could have a fan if required.
 
chris ma said:


The other way may be you can attach a small but thick piece of stock on the inside of the block first (use screws and contact cement as gasket), so that you can tap a blind hole for mounting the chip on the other side.

Originally posted by Variac
If you are going to have the radiator in the same box as the amp, you might as well attach the transistors directly to the radiator...

Chris:

I should have been thinking that in the first place.. I first thought that I'd put some glue on the screws inside the block but I wasn't possible (at the time 😉 ), so I decided to make up something when the block is fully built. Teflon tape on the threads might also have been good, although the thread size is quite small so.. I don't know.

Variac:

Yes that could be an option, but then it would not be so technical and exciting 😎 😀
 
Status
Not open for further replies.