Water Cooling
Had some teething problems with my own cooling setup that was used for sub-zero cooling of an overclocked computer - principles exactly the same though....
Ensure you use a dilute anti-freeze solution to avoid reactions between the copper/aluminium/steel parts in the water cooling loop, helps to resist things growing in the water as well.
You can use a circulation pump from a heating system - they work efficiently and quietly and cost little (less if salvaged!).
For cooling you can of course use any kind of heat exchanger like an old car radiator, an oil cooler, etc, but i have seen a miniature water cooling tower on an overclocking website (it basically mimiced the cooling towers used in power stations) - it consisted of a length of 6" pipe (waste pipe) with a Y-piece near the bottom of its length like so;
W
| |
| |
| | X
| | / /
| /
| /
| |
Water enters at point W (in spray form, say via a shower head - this is called the spray deck) and a fan extracts air from pont X. The performance is in excess of anything you will achieve with heat exchangers/oil coolers, but there is evaporation to consider, so the water supply needs to be topped up regularly, and the falling water can be noisy - but it can be built from scrap and is therefore very cheap!
Had some teething problems with my own cooling setup that was used for sub-zero cooling of an overclocked computer - principles exactly the same though....
Ensure you use a dilute anti-freeze solution to avoid reactions between the copper/aluminium/steel parts in the water cooling loop, helps to resist things growing in the water as well.
You can use a circulation pump from a heating system - they work efficiently and quietly and cost little (less if salvaged!).
For cooling you can of course use any kind of heat exchanger like an old car radiator, an oil cooler, etc, but i have seen a miniature water cooling tower on an overclocking website (it basically mimiced the cooling towers used in power stations) - it consisted of a length of 6" pipe (waste pipe) with a Y-piece near the bottom of its length like so;
W
| |
| |
| | X
| | / /
| /
| /
| |
Water enters at point W (in spray form, say via a shower head - this is called the spray deck) and a fan extracts air from pont X. The performance is in excess of anything you will achieve with heat exchangers/oil coolers, but there is evaporation to consider, so the water supply needs to be topped up regularly, and the falling water can be noisy - but it can be built from scrap and is therefore very cheap!
evaporative water cooling [aka bong cooling] is a good solution. I use it to cool my computer, a 1.2 athlon MP @ 1.7ghz. The effeciency is extremely high and goes up the higher the temperature. They are very easy to build but require some research to learn exactly how they work. There is a lot of physics and math involved that you should learn first. An enginneering reference book should have the required material. The previous poster was correct in his topical explanation of how one works except for the airflow. The fan at point X blows toward point W, if it didn't you would have a lot of drift losses. As for sound the only part of my design that makes any sound is the fan. Through proper design you can't even hear the water.
The major downfall of the design is you have to pour water in it from time to time. It works on the principle of "the heat of evaporation", the energy required to make water evaporate. It keeps my water temps at 2ºC below ambient at all times, no matter what load.
If you are interrested in a more detailed explanation of how it works and the design just ask. If you are going to use evap water cooling then you can not use anti-freeze for obvious reasons. If you decide to use a closed loop system I would reccomend that you also add a small amount of jet dry and do not use a lot of anti-freeze.
The major downfall of the design is you have to pour water in it from time to time. It works on the principle of "the heat of evaporation", the energy required to make water evaporate. It keeps my water temps at 2ºC below ambient at all times, no matter what load.
If you are interrested in a more detailed explanation of how it works and the design just ask. If you are going to use evap water cooling then you can not use anti-freeze for obvious reasons. If you decide to use a closed loop system I would reccomend that you also add a small amount of jet dry and do not use a lot of anti-freeze.
aquarium pumps
Hi
I justed chekced out some aquarium pumps. its not the flow you should worry about, its the prices huh!
Is there any DIYfishpump.com out there?
Link: http://www.thepetstop.com/fish_shop/Water_Pumps/
Keld
Hi
I justed chekced out some aquarium pumps. its not the flow you should worry about, its the prices huh!
Is there any DIYfishpump.com out there?
Link: http://www.thepetstop.com/fish_shop/Water_Pumps/
Keld
For those who are concerned about bacterial growth in a cooling system, I can say that I've had absolutely no problems whatsoever after the better part of a year in service.
I looked things over when I took the system apart and have not found any evidence of corrosion. The only metals in the system are copper, brass, and silver solder (the plumbing kind)--all pretty well behaved with respect to water.
Grey
I looked things over when I took the system apart and have not found any evidence of corrosion. The only metals in the system are copper, brass, and silver solder (the plumbing kind)--all pretty well behaved with respect to water.
Grey
HOT TUB!
Nelson: you are Brilliant!
Grey, I used to live in East Tennessee - I know the mindset to which you refer. This is a friendly site: we won't go there!
Anyway, earlier in this thread you (Grey) mentioned how great this liquid cooled Aleph sounded. (yes I have read the previous posts. I appologize if I repeated anybody's ideas.) I am wondering if you have done any more exploration as to "where the 'better' sound comes from"? Is it the higher bias current, or maybe the fact that the junctions are cooler?
Nelson: you are Brilliant!
Grey, I used to live in East Tennessee - I know the mindset to which you refer. This is a friendly site: we won't go there!
Anyway, earlier in this thread you (Grey) mentioned how great this liquid cooled Aleph sounded. (yes I have read the previous posts. I appologize if I repeated anybody's ideas.) I am wondering if you have done any more exploration as to "where the 'better' sound comes from"? Is it the higher bias current, or maybe the fact that the junctions are cooler?
The cheapest pumps you can find will be supplied by pet stores. Get a mag drive style pump, they are inexpensive and very reliable. Eheim and Danner make good mag drive pumps. Mag drives are essentially silent so they wont cause noise but a magnetic field goes out over 6" around the pump. It is relatively easy to shield them though. Never use a bilge pump or the like, not only because the noise but because they are only rated for a few hundred hours and aren't meant to be operated constantly.
As for bacteria and mold my open loop evaporative water cooler doesn't have much of a problem. Additionally, I have not seen evidence of corrosion. If properly done they are not a problem. If I were to ever water cool an amp I would definitly use an evaporative water cooler or an underground style cooler. The point of standard water cooling or any cooling for that matter is to transfer heat. No matter what you'll have to have a large radiator/heatsink somewhere in the system to dissapate the heat therefore, in my opinion, water cooling is not worth it unless you plan to achieve below ambient temperatures. Below ambeint temps can only be acheived by a physical change [such as water evap in my cooler] or something along the lines of thermoelectric cooling [peltier.] In a water cooling setup you have copper/aluminum then water then copper/aluminum again acting as exchangers of heat. It is very inefficient to have all these interface materials. It is good in computers because the radiator can be outside of the system and be huge. Additionally since the heat is dissapated outside the case the case temps are lower. In an amp I don't see this as an issue unless you just really want to hide your cooling setup.
Water cooling may seem to cool better than regular air convection but this is only because water's heat capacity [which can be saturated] and the added surface area of the entire system.
As for bacteria and mold my open loop evaporative water cooler doesn't have much of a problem. Additionally, I have not seen evidence of corrosion. If properly done they are not a problem. If I were to ever water cool an amp I would definitly use an evaporative water cooler or an underground style cooler. The point of standard water cooling or any cooling for that matter is to transfer heat. No matter what you'll have to have a large radiator/heatsink somewhere in the system to dissapate the heat therefore, in my opinion, water cooling is not worth it unless you plan to achieve below ambient temperatures. Below ambeint temps can only be acheived by a physical change [such as water evap in my cooler] or something along the lines of thermoelectric cooling [peltier.] In a water cooling setup you have copper/aluminum then water then copper/aluminum again acting as exchangers of heat. It is very inefficient to have all these interface materials. It is good in computers because the radiator can be outside of the system and be huge. Additionally since the heat is dissapated outside the case the case temps are lower. In an amp I don't see this as an issue unless you just really want to hide your cooling setup.
Water cooling may seem to cool better than regular air convection but this is only because water's heat capacity [which can be saturated] and the added surface area of the entire system.
The other end of the heat exchanger is stamped:
COLEMAN 6904-502
This side is more clearly read--not rusted--but whether the first end read 501 or 502 is open to question. If I get time, I'll roam around on the web and see if I can find any specifications on such a part. Since Coleman doesn't make heat pumps, I'm assuming that they are an OEM supplier to Carrier, Fedders, et. al. This may make it difficult to get information, particularly if they custom design parts to the customer's specifications.
Grey
COLEMAN 6904-502
This side is more clearly read--not rusted--but whether the first end read 501 or 502 is open to question. If I get time, I'll roam around on the web and see if I can find any specifications on such a part. Since Coleman doesn't make heat pumps, I'm assuming that they are an OEM supplier to Carrier, Fedders, et. al. This may make it difficult to get information, particularly if they custom design parts to the customer's specifications.
Grey
i have a watercooled computer and felt compelled to add to this post.
in all of the pc's we use Ehiem pumps, they are more expensive but last a very long time, are quiet and they do not have a large interferrance.
for the heat exchanger the best and cheepest one avalbe are only heater cores (or matrixes) for cars. the waterblock for the mosfets can be made out of a waterproof aluminum project box. to improve heat transferr and prevent bacteria build up use water wetter in the water.
in all of the pc's we use Ehiem pumps, they are more expensive but last a very long time, are quiet and they do not have a large interferrance.
for the heat exchanger the best and cheepest one avalbe are only heater cores (or matrixes) for cars. the waterblock for the mosfets can be made out of a waterproof aluminum project box. to improve heat transferr and prevent bacteria build up use water wetter in the water.
phage, have you ever noticed that the efficency of you evaporative cooling circuit decreases after running for a while?
I was thinking that since you are increasing the relative humidity of the air in the room , it will eventually lower the evaporation rate , unless you have good ventilation?
I was getting the idea to build one myself for my computer , silencing it totally , since one of my collegues just spend a few working days designing one at work (cooling power = 10 MW)
( I design chemical plants for living)
power semiconductors allow for a higher temperature of the cooling water than a CPU,hence higher efficency but if you are running a class A amplifier, it lead to massive evaporation (@ 500W, almost 1 kg/h) wich probably would saturate the air if placed indoors and give rise to moisture damage in you house.
/rickard
I was thinking that since you are increasing the relative humidity of the air in the room , it will eventually lower the evaporation rate , unless you have good ventilation?
I was getting the idea to build one myself for my computer , silencing it totally , since one of my collegues just spend a few working days designing one at work (cooling power = 10 MW)
( I design chemical plants for living)
power semiconductors allow for a higher temperature of the cooling water than a CPU,hence higher efficency but if you are running a class A amplifier, it lead to massive evaporation (@ 500W, almost 1 kg/h) wich probably would saturate the air if placed indoors and give rise to moisture damage in you house.
/rickard
The water temperature of my cooler actually drops when it is on. It levels off at about 2C below ambient after a few minutes. This figure (2C below) seems constant despite the load which makes sense (higher dt of water to air = more evaporative losses and more temp drop) but the water temp would go up with time as the air saturates. I usually have ok air circulation in my dorm room but it is very humid, humidity isn't as big of a factor as you would think. If you think about it despite humidity the surface area of the water being exposed to air is much better than that of a radiator and there is no interface material so there is better dissipation/no thermal gradient.
My computer dissipates about 100W. The cooler evaporates 7oz/hour of water with negligible drift. My amp dissipates about 900W, I would never try to cool it with an evaporative water cooler... Losses would be massive. It would likely saturate the air of just about any room very quickly and have massive drift losses (if not just steam comming out of it)
My computer dissipates about 100W. The cooler evaporates 7oz/hour of water with negligible drift. My amp dissipates about 900W, I would never try to cool it with an evaporative water cooler... Losses would be massive. It would likely saturate the air of just about any room very quickly and have massive drift losses (if not just steam comming out of it)
phage, you can actually calculate the humidity in the air, given that the water is 2 C (wet bulb temperature) below ambient temperature(dry bulb temperature) but for that you need a air-water-enthalpy diagram, which can be found in most chemical engineering books( one I recommend is Perry´s)
but like you already said before, you need to do some calculations for that.
but if you want to cool a amplifier, I think you have to place it outside since like you said, you will saturate the air with water and then cooling drops off very quickly, since you only have conductive heat transfer from the waterdroplets, and not evaporative cooling.
and actually, the cooling surface of the drops arent that big, considered that they fall very quickly. It is just that evaporative cooling is so efficent. A better way would to put filling in you cooling tower , hence increasing surface dramatically.
but it seems like your cooling is pretty good though, to go all the way down to the theoretical wet bulb temperature would take a
LARGE tower.
/rickard
but like you already said before, you need to do some calculations for that.
but if you want to cool a amplifier, I think you have to place it outside since like you said, you will saturate the air with water and then cooling drops off very quickly, since you only have conductive heat transfer from the waterdroplets, and not evaporative cooling.
and actually, the cooling surface of the drops arent that big, considered that they fall very quickly. It is just that evaporative cooling is so efficent. A better way would to put filling in you cooling tower , hence increasing surface dramatically.
but it seems like your cooling is pretty good though, to go all the way down to the theoretical wet bulb temperature would take a
LARGE tower.
/rickard
cheapo amp cooler
Years ago, I saw a cool set-up at a nightclub and wondered if anyone else had tried it. The amp they used was overheating on "loud" evenings due to poor placement etc. The owner used an old water fountain chiller, some plastic line(draft beer supply line) a circulating pump, an old pony keg and some copper tubing to cool the amp. All this stuff was free or cheap! He ran the chiller outside the club and attached the plastic line to the potable water outlet. It was wrapped w/ insulation. It made its way to the amp where it coupled to copper tubing which was woven in and out of the HS fins. Alumunim foil was crammed into the gaps to help heat transfer. From there, back to plastic, outside to the pony keg (help to dissipate alittle more heat and buffer the loop) to the pump and back to the water inlet for the chiller(air cooled). It was ugly but he claimed it worked great. No thermastat just a switch for the pump.
By the way, anyone remember the late 70s Sanyo freon cooled amps?
Good luck, Taylor
Years ago, I saw a cool set-up at a nightclub and wondered if anyone else had tried it. The amp they used was overheating on "loud" evenings due to poor placement etc. The owner used an old water fountain chiller, some plastic line(draft beer supply line) a circulating pump, an old pony keg and some copper tubing to cool the amp. All this stuff was free or cheap! He ran the chiller outside the club and attached the plastic line to the potable water outlet. It was wrapped w/ insulation. It made its way to the amp where it coupled to copper tubing which was woven in and out of the HS fins. Alumunim foil was crammed into the gaps to help heat transfer. From there, back to plastic, outside to the pony keg (help to dissipate alittle more heat and buffer the loop) to the pump and back to the water inlet for the chiller(air cooled). It was ugly but he claimed it worked great. No thermastat just a switch for the pump.
By the way, anyone remember the late 70s Sanyo freon cooled amps?
Good luck, Taylor
I have used a radiator (15 *12,5*50 cm, no fans , only natural draught) from a heat pump as a cooler for a AC-DC power supply. And I had a water block similar the once in previous poster + an aquarium pump to pump the water.
I ran the water at about 50 C in to the radiator and 42 C out. With a flow of 1,5 l/min that equates to 840 W cooling power!
With a fan I estimate that I could have put out 3 times as much power with the same temperature. (50 C is a good temperature of the power block) or used a much smaller radiator.
These radiators can be found in places that sell fridges and
cooling ventilation and various heat pump resales. And dont forget the A/C units. Or car-radiator.
/rickard
I ran the water at about 50 C in to the radiator and 42 C out. With a flow of 1,5 l/min that equates to 840 W cooling power!
With a fan I estimate that I could have put out 3 times as much power with the same temperature. (50 C is a good temperature of the power block) or used a much smaller radiator.
These radiators can be found in places that sell fridges and
cooling ventilation and various heat pump resales. And dont forget the A/C units. Or car-radiator.
/rickard
I am going to use a water cooling system for my 7200W amp. I have already constructed the lower part of the chassis, with the radiator and fan under it. It uses a radiator, water pump and fan from a car, and a nice, quiet dishwasher motor for the pump. The fan is above the radiator and sucks air up through it. I haven't been able to test it yet because I haven't figured how to get the motor started under even the fairly slight load of the pump, but It should have plenty of cooling capability. The plan is that there will be MOSFETs mounted on all 4 sides of each of the aluminum blocks, which are of course hollow, with 1/4" walls. There will be a board on the top and bottom of each end, for the 8 channels. The power supply will be in a large compartment on the back that I have not yet built, and the main chassis will be 7 1/2" high, for a total height of 15 1/2". The speaker terminals will be on each side, to keep the speaker wiring away from the rest of the circuitry.
So far, it looks somewhat (no, exactly) like this:
So far, it looks somewhat (no, exactly) like this:
Attachments
UrSv - DI water in a water cooler is a bad idea unless you are about to add antifreeze to it. DI water would actually promote galvanic corrosion as it becomes unpure over time. The antifreeze you put in the system is only there to prevent corrosion. Antifreeze actually causes water's heat transferring ability to be less so you should not use much.
Also I've noticed the water block designs on this board to be poor. Most of the designs promote lamnar flow of the water. Basically it is where the outer column of water becomes an insulator due to the high flow rate of the water. The inner part of the water column therefore stays cool while the outside gets very hot. In your block deisngs you should include something to cause turbulence. Check out the best water block desings for computer cooling and you'll see they all promote turbulence. Of course this causes more head on the pump which is also not desired, it's a tradeoff.
Also I've noticed the water block designs on this board to be poor. Most of the designs promote lamnar flow of the water. Basically it is where the outer column of water becomes an insulator due to the high flow rate of the water. The inner part of the water column therefore stays cool while the outside gets very hot. In your block deisngs you should include something to cause turbulence. Check out the best water block desings for computer cooling and you'll see they all promote turbulence. Of course this causes more head on the pump which is also not desired, it's a tradeoff.
What I have now for my blocks is just a plastic elbow epoxied into a hole in the end if the blocks. I suppose I should put something in the system to add turbulence. It shouldn't take much, as the water will be moving very fast through my system.
Does anyone think I might have problems with coolant flowing unequally between the two blocks?
Does anyone think I might have problems with coolant flowing unequally between the two blocks?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.