Ice cream maker using peltier devices?

[cdev]

Well, my sweetie (yes, we are married) and I like to fool around with food.. not that way.. but recently we've started making keto-friendly ice cream, (using coconut milk, mostly, and other things..
we currently are using a machine that only makes 1 pint at a time, which isnt that much. Since I have the electronics interest, a short time ago I bought a single peltier device, was amazed that it worked as well as it did while heatsinked and how it worked, so convinced I bought five more, hoping to get something built by June of this year.. Basically now I want to take these six and use them to make ice cream. I was wondering if anybody else here had done that. Here it is the middle of the winter now. (its been a very mild one) But in Australia, this must be summer.. Frankly, I know the principle of operation and have some ideas of how I might go about this but I can see its difficult to get it right. My wife is actually the engineer, though. Twice.
The instant cold of the Peltier devices and the cold on one side hot on the other aspect of the thermoelectric device is very cool.  I think that given an adequate current, it should be possible. But perhaps not so easy.
I am figuring 20 amps will do it. I have a converted PC supply. And two bench supplies. What I would love to do is make a "tight" convenient ice cream freezer.. that doesn't take up a lot of space. How should I approach this?
It gets very humid here in the summer so I know care is going to have to be put into isolating the cold area from the ultra moist summer air. Although since its distiilled water being condensed out of the air and a fairly low voltage, I dont think it will be that problematic. Maybe just making a cooled slab will do? I can get nice granite slabs really cheap.  Ive never used one. I have no idea how that works out. Would have to break my diet to try out the commercialized version of it.

[retiredfeline]

Hmm, I would do first some back of the envelope calculations with the specific heat of water, well ice cream mix, the efficiency of the device, etc. to get the power required to make it feasible.

[Ian.M]

Lets run the numbers.  The latent heat of fusion of the water content is likely to dominate.   Coconut milk can be up to 20% fat and Ice Cream up to near 40% sugar, so lets assume there's only 50% water in the mix that needs freezing.  Lets assume you'll be doing a liter at a time, and the target cycle time is half an hour,
The heat of fusion of water is 333 Joules/g, so you need to pump 166.5 KJ of heat out just to freeze the water, + a bit more for the other ingredients and to chill it to a low enough holding temperature.  The approx 15% coconut lipids will push the total over 180 KJ.  You've got 1800 seconds to do so so you need to pump heat at a rate of 100W, at a fairly high temperature differential, maybe 40 deg C on a hot day,  with that high a differential + the need to run the heatsink hotter to be able to reject heat to ambient, the COP could fall below 20%.   That's not a freezer, that's a small room heater, literally tiled all over its exterior with Peltier modules to get enough heat out!

[Siwastaja]

Peltiers Do Not Work.

Well, they do, but COP is just so horrible given any temperature differential that you basically have thermal runaway beyond sensible operating conditions.

If you just need cold temperature, i.e., cool a small amount of air or a few drops of water, but no need to do any serious work (i.e., energy), they do work, because with little power, efficiency does not matter.

If you need to actually transfer energy, then you can't get the temperature differential, because the COP drops as a linear function and reaches zero at maybe 50degC differential.

Huge amount of losses and inability to generate temperature differential is a bad combination; you need to spend some serious effort cooling the hot side, trying to keep that as close as ambient as possible.

This is a shame, Peltiers as simple solid state devices would be nice for DIYing cooling solutions.

They are good for calibration jigs of sensors, ICs, etc. For more cooling energy (i.e., freezing significant amounts of water), you can't avoid playing with real refrigeration systems, which requires quite a set of tools (pipe, brazing tools, flaring tools, vacuum pumps, and of course refrigerants, but you can use propane/butane). That's quite a rabbit hole.

Best would be to find a way to modify/utilize existing freezers.

[cdev]

You make a good point but the fact that I have a bunch of DC power supplies that could be used and six peltier devices that (last summer) I got to "do something with the technology" I'm determined to do something to help us stay cool this summer, with them. Its remarkable how much better a Peltier device works given adequate heat sinking and separation of the sides, isnt it? That really is the key, I can see, to getting your money's worth out of them.

Okay, I have to disclose here that I am hoping to be able to duplicate a keto version of ice cream on a slab which is extremely good. Years ago there used to be an ice cream store in San Francisco that made amazing "cheesecake" ice cream on a slab. It was in the Fillmore district. And it was amazing. Unfortunately I forget their name. They used a large marble slab to prepare their creations on. And fresh cheesecake. No it was not Cold Stone Creamery, but the concept was similar, however the ingredients were all freshly made. I think getting a flat piece of stone cool enough doesnt sound so super difficult, Ive seen it done using Peltier devices and heat sinks.

The magic of it is that even milk (regular milk) makes a delicious dessert - when mixed with hot muffins, cheesecake, etc. And its very fast to make.

Its sort of like this

[PlainName]

A quick search turned up:



That's a lot of hardware for a small cold block!

But... think of the marble slab as a battery and the peltier as a low-power charger. You just need to give the slab time to be charged by the peltier rather than try to use the peltier as the realtime source.

[Siwastaja]

Yes and it's nowhere near 600W, heatpipes are no go because their working principle depends on temperature differential, which you can't have. Power consumption might be 600W, yeah.

100W of cooling power is realistic with what is shown. Given enough time and thermal insulation, that of course can freeze water, concentrated into such small volume.

I did build a fairly successful Peltier-based cooling device which I use to calibrate sensors. 6 * 60W elements mounted to a totally massive heatsink with large radial server fans. I can successfully freeze water with it, but this also relies on room temperature being just 20degC or so. Make it 30 and you are out of luck.

But anyway, this is no rocket science. Basic thermal calculations work. Look up the datasheet, typical performance is 100% COP at dT=0 linearly down to 0% COP at dT=55degC.

What this means is, if both hot and cool sides are at +20degC, you consume 100W of power, cool at 100W, and dissipate 200W of heat. Obviously dissipating 200W means temperature differential RthJ-A * 200W over the hot side alone, making the hot side not +20degC, but maybe +30degC with heroic cooling efforts. With dT=10, COP is already down to around 80%. Which means we are at equilibrium, consuming 100W, cooling 80W, and dissipating 180W.

Now the water slowly cools down, and when it's at 0degC, dT is already 30degC - still assuming air-conditioned 20degC room! - and COP is down to maybe 40%. Power in = 100W, cooling = 40W, heating = 140W.

Latent heat required to freeze water is massive, so with just 40W, it takes forever, but eventually it will happen, of course. At that point, you have dissipated so much energy that the room has likely heated up.

But sure, if you have good air conditioning, turn that as low as possible to keep the room cool, and you have chances of success. Say, if you buy 1500-2000W nominal of Peltiers, derate them to run at say, 1000W, use a car engine radiator block with a massive fan to liquid cool the peltiers, and have that radiator inside an air-conditioned space (or only do ice cream during winter), it likely works out.

Easier solution: get a used freezer, and build the ice cream maker inside of it.

[ogden]

Right. Peltier-based ice cream machine is expensive to build and expensive to operate, if it actually works in the end. I suggest to buy ice cream machine and build something else.

[Marco]

Just get an extra freezer and buy as many of those cheap churners as you want.

The only reason to DIY something would be if you wanted one of those freezing stainless steel tops you can make icecream on the fly with with a spatula, though I'd probably still just do that with a normal freezer with a big bucket of frozen brine as a heat dump with a heatsink in it and a propylene glycol cooling loop.

[PlainName]

do you want to spend your time building a gadget or making ice cream?

 

[tooki]

Also, bear in mind that ice cream is something where you really need to go wild with cooling capacity: you want the water to freeze as quickly as possible, to avoid the creation of large crystals. (This is why dippin’ dots and other liquid-nitrogen–frozen ice creams are so creamy: they freeze instantly before crystals can form.) It’s also why your ice cream mix should be chilled to just barely above melting (so already below 0C, due to the sugar) before churning it.

[cdev]

A. Making delicious ice cream~

do you want to spend your time building a gadget or making ice cream?

 

[Marco]

It’s also why your ice cream mix should be chilled to just barely above melting (so already below 0C, due to the sugar) before churning it.

Getting it below fridge temperature but above freezing is a lot of effort and I don't see how it would make it freeze faster. Almost all of it is going to get just above freezing till it starts freezing regardless, stirring a liquid keeps it at equilibrium pretty well.

[tooki]

It’s also why your ice cream mix should be chilled to just barely above melting (so already below 0C, due to the sugar) before churning it.

Getting it below fridge temperature but above freezing is a lot of effort and I don't see how it would make it freeze faster. Almost all of it is going to get just above freezing till it starts freezing regardless, stirring a liquid keeps it at equilibrium pretty well.

Ummm, the warmer the liquid is, the more heat you need to remove from it before you can get it to freeze. It's not hard to get it as cold as possible: put it in the freezer and check on it periodically until it looks like it wants to freeze, or even begins.

Again, the point is to make the actual freezing process as quick as possible so there isn't time for large ice crystals to form.

[Marco]

Almost all the heat energy is in the phase change. If it stirs from fridge temperature a tiny bit of frost will form on the container till it gets to the same temperature if you put it in the freezer first, does that really make a difference to the texture in a blind test?

[Siwastaja]

Note that the sugar and other ingredients mess up the phase change so that it doesn't happen at a single temperature, but over a range. I haven't made ice cream (at least yet), but I'd guess mixing within that range, during freezing, does definitely affect the structure.

Think about eutectic 63/37 solder which changes phase quickly at a single temperature creating shiny joint like block of ice, vs. lead-free solder which, if moved around during cooling, will create an amorphous mess called cold joint, because phase change happens within a range of temperatures.

Similarly, pure water has a clear freezing point, but water-glycol mix does not, it becomes slurry first and different "pipe burst point" is much lower than the theoretical "freezing point".

[Microdoser]

Myself, I would get a cheap ice-cream maker, the sort where you pre-soak the bowl in your freezer for 24 hours.

I've found they work quite well as long as your ice-cream mixture is nearly at freezing already.

We have a fridge-freezer that can freeze items if they are at the very bottom in the meat drawer, so we use that to prepare stuff like ice-cream etc

Far more efficient and reliable.

The alternative is to spend a lot on a maker that freezes your mixture itself. Not tried one of those due to the price.

[Marco]

Note that the sugar and other ingredients mess up the phase change so that it doesn't happen at a single temperature, but over a range. I haven't made ice cream (at least yet), but I'd guess mixing within that range, during freezing, does definitely affect the structure.

If crystallisation and freeze distillation started inside the bulk before you start stirring you're definitely too late. The only difference between fridge temperature and cooling it to just above freezing could be a tiny bit more hard frost on the lining which the scraper can't remove, but I kind of doubt it's relevant.

The big gains in ice cream making are lower wall temperatures, but that is hard to achieve with a freezer frozen bowl filled with an eutectic. Even if you lower the freezer temperature, the melting temperature of the eutectic will determine wall temperature for most of the cycle. Compressor based ice-cream makers can go faster in that respect. They can also deal with hot custard.

[tooki]

Almost all the heat energy is in the phase change. If it stirs from fridge temperature a tiny bit of frost will form on the container till it gets to the same temperature if you put it in the freezer first, does that really make a difference to the texture in a blind test?

My hunch is yes, since I do recall seeing that commercial ice cream makers chill the mix to near-freezing before churning.

For home ice cream making, the fairly weak cooling capacity of many home ice cream makers (especially the traditional ice and rock salt or  brine-walled “cartridge” types) means it’s easy to run out of cooling before the mix has frozen properly. Hence why it’s good practice to get the mix as cold as possible (while still liquid) to ensure it freezes in time.

[rpiloverbd]

Well, I do not think a Peltier plate will solve your purpose. Peltier plates can at best make a mini-refrigerator for storing one or two chocolate bars.  They are not suitable for ice cream machines. You may end up wasting too much power but not getting satisfactory results.

[Zero999]

Regarding sugar, the original poster wants to make keto ice cream, which will have minimal sugar content. I wonder how it's going to work. Normally the sugar is essential in lowering the freezing point, allowing parts of the mixture to be supercooled. Presumably the idea is to use something else such as a polyol, like glycerine to replace sugar.

I agree with the above advice: don't bother with a Peltier cooler. A vapour compression phase-change cooler is much more efficient. A small old freezer is the perfect device to build an ice cream maker from.

[cdev]

We're using coconut milk and monkfruit for sweetening right now. I have given up on peltier and we're just usig our refrigerator and the heat storage type of freezer. We're also using molds for ice cream bars. Which are great. If you have kid definitely do this, I could see it saving a lot of money on popsicles.