OK, bear with me here, this crazy idea has been haunting me for a couple of years. I keep thinking it would be such a fun project to bring to Burning Amp.
Here's the idea: use a pot (one with a thick aluminum bottom disc) as a heat sink for an Aleph amp (the forthcoming Aleph J boards look square-ish enough?). Mount the semis without penetrating into the interior of the pot. And fill the pot with water. Mount the transformer/power supply in a circular enclosure underneath, so the whole contraption looks kind of like a pot sitting on a camp stove.
Yes, I know it's crazy. Yes, I know having water close to electricity is inherently unsafe. Let's just think about this theoretically for a minute, shall we?
Is it even thermally feasible? Does anyone know what the dissipation would be for an aluminum pot with a half gallon of water in it?
Here's the idea: use a pot (one with a thick aluminum bottom disc) as a heat sink for an Aleph amp (the forthcoming Aleph J boards look square-ish enough?). Mount the semis without penetrating into the interior of the pot. And fill the pot with water. Mount the transformer/power supply in a circular enclosure underneath, so the whole contraption looks kind of like a pot sitting on a camp stove.
Yes, I know it's crazy. Yes, I know having water close to electricity is inherently unsafe. Let's just think about this theoretically for a minute, shall we?
Is it even thermally feasible? Does anyone know what the dissipation would be for an aluminum pot with a half gallon of water in it?
As long as there is still water in the pot the "heatsink" temperature should not exceed 100C. However, 100C is a lot higher than the 50C-55C target for most of the amplifiers that we build.
TL;DR: Cooling with convection is better than cooling with conduction, in this case.
Let's see. The real question is whether the pot of water will provide the thermal transmission needed to keep the transistors from self-destructing.
Initially: the water is 20 degrees C, like the room. Turn on the amplifier, and the transistors naturally heat up to some temperature well above 20 degrees. Over time, the heat from the transistors is conducted to the water, and heats up the water. At the same time, however, air currents carry away heat from the pot and from the surface of the water. As the water heats up, these convection currents increase, taking away more heat.
Depending on how much water there is, at some time later the water temperature will reach the equilibrium point where the heat leaving the pot equals the heat being dissipated by the transistors. However, as @lhquam said, it is possible that the temperature could reach a maximum 100 degrees C, at which point your devices would likely have already stopped operating. "Ah!", you say, "But I can use a bigger pot of water!" Yes you can, but doing so increases the surface area that the air around it is cooling. You are thus creating a bigger heat sink, which will have a lower equilibrium temperature.
Your idea, I think, is that the water would be able to absorb more heat than air, and this is true, if we forced our amplifier to operate in a sealed box, making it use a similarly fixed amount of air. Over a much shorter time, the air inside would become very warm, like the pot of water.
While the specific heat of water is indeed much higher than air, cooling with heat sinks works because of convection, whereby the effectively inexhaustible supply of cooler air is readily available to remove excess heat. We use heat sinks with fins because they increase the surface area exposed to air, and this increases the volume of air removing heat.
Let's see. The real question is whether the pot of water will provide the thermal transmission needed to keep the transistors from self-destructing.
Initially: the water is 20 degrees C, like the room. Turn on the amplifier, and the transistors naturally heat up to some temperature well above 20 degrees. Over time, the heat from the transistors is conducted to the water, and heats up the water. At the same time, however, air currents carry away heat from the pot and from the surface of the water. As the water heats up, these convection currents increase, taking away more heat.
Depending on how much water there is, at some time later the water temperature will reach the equilibrium point where the heat leaving the pot equals the heat being dissipated by the transistors. However, as @lhquam said, it is possible that the temperature could reach a maximum 100 degrees C, at which point your devices would likely have already stopped operating. "Ah!", you say, "But I can use a bigger pot of water!" Yes you can, but doing so increases the surface area that the air around it is cooling. You are thus creating a bigger heat sink, which will have a lower equilibrium temperature.
Your idea, I think, is that the water would be able to absorb more heat than air, and this is true, if we forced our amplifier to operate in a sealed box, making it use a similarly fixed amount of air. Over a much shorter time, the air inside would become very warm, like the pot of water.
While the specific heat of water is indeed much higher than air, cooling with heat sinks works because of convection, whereby the effectively inexhaustible supply of cooler air is readily available to remove excess heat. We use heat sinks with fins because they increase the surface area exposed to air, and this increases the volume of air removing heat.
If you decide to do this I have the new square boards, you’re welcome to them. 
Also worth note, IRFP150 has a max allowable junction temp of 175C, and the on-resistance graph goes to 160C, so 100-120C, though incredibly hot by our standards, seems like it’s not going to cause problems.
Also worth note, IRFP150 has a max allowable junction temp of 175C, and the on-resistance graph goes to 160C, so 100-120C, though incredibly hot by our standards, seems like it’s not going to cause problems.
Last edited:
New square Aleph J boards? Is this the reason the boards have been out of stock for awhile? Feel free to start a new thread to share the secrets.If you decide to do this I have the new square boards, you’re welcome to them.
Also worth note, IRFP150 has a max allowable junction temp of 175C, and the on-resistance graph goes to 160C, so 100-120C, though incredibly hot by our standards, seems like it’s not going to cause problems.
OK, so it doesn't sound completely thermally inconceivable then? Next steps, I'm slowly going think about parts. Like a pot. And think about finding a drill press...
@6L6 if you've got an amp board, I'd love to have it. No rush, I expect it'll be some time before I learn all the skills that I'll need to undertake this crazy idea, but it will help me figure out what size pot I'll be needing. 🙂
@6L6 if you've got an amp board, I'd love to have it. No rush, I expect it'll be some time before I learn all the skills that I'll need to undertake this crazy idea, but it will help me figure out what size pot I'll be needing. 🙂
There are water-cooled vacuum tubes, and a very few evaporative (vapor) cooled vacuum tubes. They seem to have been squeezed-out between fan-cooled and water cooled.
History of Vacuum Tubes, Sōgo Okamura
And FWIW, swamp-coolers really do cool, west of the Mississippi.
Hopper cooling was used on many-many small farm engines. Pour in a bucket of water and it might steam all day.
100 years ago the Still Engine took steam from an IC engine's cooling jacket for more efficiency.
The WWII Rolls-Royce Goshawk was not evaporative cooled, but managed its evaporation. Seemed like a good idea. Led to better ideas......
Many steam-based power generating plants are evaporative-cooled via cooling towers (symbol of Atomic Energy but really used anywhere they don't have a river).
History of Vacuum Tubes, Sōgo Okamura
And FWIW, swamp-coolers really do cool, west of the Mississippi.
Hopper cooling was used on many-many small farm engines. Pour in a bucket of water and it might steam all day.
100 years ago the Still Engine took steam from an IC engine's cooling jacket for more efficiency.
The WWII Rolls-Royce Goshawk was not evaporative cooled, but managed its evaporation. Seemed like a good idea. Led to better ideas......
Many steam-based power generating plants are evaporative-cooled via cooling towers (symbol of Atomic Energy but really used anywhere they don't have a river).
you need very clean destilled water -
or you will have some corrosion (different metals...........)
why not mount the semis onto the bottom outside the water kettle?
or you will have some corrosion (different metals...........)
why not mount the semis onto the bottom outside the water kettle?
You could consider a thermal protection of sorts, and I‘d advice to increase the cooling surface (more surface, more water-air-contact = bigger heatsinks)
I love crazy ideas. This is a hobby.
Anyway, for every ten degrees C on the die, rule of thumb is half life.
Industry standard to never exceed is 135C on the die. Good reliability is under 90. Case temps will be a lot lower and some dvices, like MOSFETS have "hot-spot" issues.
Anyway, for every ten degrees C on the die, rule of thumb is half life.
Industry standard to never exceed is 135C on the die. Good reliability is under 90. Case temps will be a lot lower and some dvices, like MOSFETS have "hot-spot" issues.
I have been given flack about how much green tea I drink. I think smoking hot sounding (and operating) amp
that could also heat water as a part of its cooling process, would be at home in my rig(s).
I am ALL IN for the prospect of such technology. Tea kettles are over-rated!
I digress. sorry.
that could also heat water as a part of its cooling process, would be at home in my rig(s).
I am ALL IN for the prospect of such technology. Tea kettles are over-rated!
I digress. sorry.
not sure if my math is good, but the energy needed to increase 1 gallon of water by a temp. of 25C is .11kWh. This of course leads to an enthalpy calculation based on the surface area of the water, the speed of air over the water, and probably something else I'm missing. All of which has long been forgotten since college physics class! Other variables involved in this particular scenario would be the conductive rate between FET and it's mounting surface and the conductive nature of the kettle itself, (preferably copper), and whether this surface is capable of maintaining a delta T of a particular temperature within a time that would represent a safe zone for an electronic part, while heating up said fluid in the kettle. Additionally, would the energy removed from the system by evaporative cooling be enough to maintain a stable state at said temp (25C, a typical rise for a class A power amp). The kettle/vessel itself would also have some convective energy loss to help the system. I can raise a gallon of water by 25C on my stove, but this requires a hot element that is most likely beyond the safe operating temp of a FET, for a given amount of time. It probably wouldn't be too difficult to experiment by empirical means with a 'test jig' to observe results if the math is a hang up.
Last edited:
The thermal storage capacity of water is huge. At a certain point you'll have to figure out a way to cool the water.
Might also be used at the holidays for keeping the glogg warm! 🙂 Not sure it will manage green tea steeping temperature, but I guess there's only one way to be sure... ;-)
OMGosh, off topic but I never thought that perusing this thread would take me back 60+ years to a rural Michigan Historic Farm Equipment show and some of the early eye-popping steam-based belt-drive systems for thrashing machines and literally anything where you needed some kind of motive power. Amazing to see a couple of those monsters fired up! Of course those weren't as efficient and innovative as those this thread.
Then there was the miniature steam engine I got for Christmas in the mid-1950s. Around 12" tall and 4" in diameter with a flywheel and a steam powered whistle and a 'safety' release blow valve. The electric coil heater in the blue hammertone painted base got so hot it turned the paint brownish by the third use. That thing was probably on the top 10 list when they started looking at product safety. (Bah Humbug, survival of the fittest!) It must have had a half dozen issues that now give me pause. Living in a rural area it would have been from the Sears or "Monkey" Ward Christmas catalog. Maybe a retro-science project to put on my list?
Wasn't one kind of swamp cooler a cylindrical metal gadget we hung on the passenger window of our ~ 1962 Olds that caught air from a forward facing inlet and passed it thru some fibrous mesh then into the passenger compartment. Worked 'better than nothing (but not by much!) on a road trip to Buena Vista Colorado.
Thirty seven years ago, Cray Research began selling their Cray-2 supercomputer (link to Wikipedia article). Its circuit boards were immersed in circulating liquid called Fluorinert. You can buy a gallon of it yourself, today, for USD 600 per gallon (sales link)
My friends who worked at Lawrence Livermore National Laboratory, affectionately nicknamed their Cray-2 "Bubbles".
My friends who worked at Lawrence Livermore National Laboratory, affectionately nicknamed their Cray-2 "Bubbles".
Couple of directions this could go:
- The world's best sounding percolated coffee
- Water pump, radiators: A water-cooled amp.