Peltier unit enviromental control

[dr_deathweed]

Hello all. Please forgive my newbie status, I am much more adept in the biomedical field than in electronics, but am always open to learning something new. I would like to run my project past some experts to ensure that I am on the right track, and possibly get some ideas on improvements.

The project: I have a 50L cooler that I would like to install a peltier unit in to maintain a temperature between 65 and 68deg F. This will be stored in a location where environmental temp can exceed 80F. I would like to be able to run this entire project off of house AC so I can set and forget the system for months at a time. Also having an external display of interior temperature and relative humidity would be a plus, but not necessary if cost or space requirements are exceeded.

In the past I have built larger enclosures using a household thermostat on a relay to switch a fridge compressor unit on an off, but due to space that will be over kill in this situation.

My questions:
- Cooling 0.05m2, I roughly figured I needed 2-4W of cooling to allow the peltier to run long enough, and prevent short cycling. Please correct me if I am wrong in this.
- Currently I am looking at building a thermostat unit like the one at http://www.craig.copperleife.com/tech/thermo/ and using that to control the power supply to both the peltier and the DC fans on hot and cool sides of the unit off the relay pictured. Is there any problem running the high amps of the peltier through the same circus as those fans, or should I place them on a separate circuts with a SPDT relay?
- Powering this peltier and the circuit at a decent cost has been a thorn in my side. I am currently thinking of using a peltier with the ratings range of ~15-20W @~15v and 2-3A with a 12v 1A power supply to get what I need. Is this reasonable, and can I run my thermostat circuit off this as well?
-As for display, I am thinking of using a cigar hygrometer, mounting on the lid of the cooler, and placing the thermsistor and humidity sensor on leads through the lid. I have done this with various household thermostats, but have not taken one of these hygrometers apart. Any foreseeable issues, or is it possible to build this circuit as well with my beginners skill set?

This is all I have at the moment, I am sure more questions will arise, and thank all in advance for all your help.

 

[daddles]

You can estimate (note I said estimate) the heat load by a simple back-of-the-envelope calculation using Fourier's law -- heat flow in watts per square meter of surface area equals the thermal conductivity times the temperature gradient (I'm ignoring the usual minus sign too). Pick a thermal conductivity of a likely material you'd use as an insulator. For argument, assume k = 0.1 W/(m*K), as this is a fairly typical number for thermal insulators (expanded plastic foams can be a third of this or so). I'll use SI units.

Assume some thickness of insulation; let's pick t meters. You want a temperature differential of 80-65 deg F = 15 deg F or 8 K. Then the heat flow is k*(8/t). If t = 1 cm, then this produces 0.1*8/0.01 = 80 watts/square meter. If you know the area of the insulated container, you can then calculate the heat load. You can then estimate the electrical power needed by dividing by the efficiency of the Peltier device (you also account for less-than-perfect thermal joints by lowering the efficiency some).

Note this ignores convection and radiation effects, but they won't be too significant for this problem anyway.

 

[dr_deathweed]

Thank you very much for correcting my horrendous estimation. I had figured the wattage needed to cool my volume and not factored in heat losses. As is, my cooler is roughly 24”x14”x14” with 1 inch foam insulation, or (very) rough rounding up for error 1.2square M external surface area. With a compromised k value of 0.05 for foam and a thickness of 2cm, that gets me at a 20W heat load.

Browsing Digikey for what I am willing to spend I have found peltiers from 33.4W 4A, 50.5 W6A, and 75W 8.5A (all at 15.4V). What considerations should I have looking for a power source for these? I think I remember a switching source is better for peltiers, but may be misinformed. Off to Google for more reading….

 

[daddles]

I haven't applied Peltiers personally. I'd recommend picking a power supply that is rated roughly 1/2 to 2/3 of the needed power. I suspect the type of power supply is irrelevant. I'd recommend using a switching supply because it's more efficient than other choices. If you live in the US, check places like All Electronics, Goldmine Electronics, or MPJA -- they often have power supplies at good prices.

 

[ChrisQ]

I used to run a peltier on a waterblock hooked up to a computer chip back in my extreme overclocking days. I have no useful data to give you except that if it's not insulated it will look like an old freezer in about 10 minutes, However the power supply was the Meanwell S-320-12 which was at the time (and evidently still is) a standard among TEC enthusiasts.

The power supply plugs into the wall and then the Meanwell supplied 12v to the TEC.

Haha just noticed that it's out of stock Maybe it has been replaced by now after all.

 

[dr_deathweed]

This next question may just demonstrate my misunderstanding of peltier units and basic electronic principals in general, but as I understand it, it is ok to under power these peltier's voltage wise, you just will not get the full rated wattage benefit. However, where I am confused is when looking at performance curves, if a 15.4V 6A peltier is run at 12 V, it will draw about 4.18 amps or so. If my power supply is only rated at 3 A or 4 A, how will that affect other loads on the circuit (like hot side fans). Now I understand draw will be less as delta T is less across the module, but as it heats up it draws more, possibly impeding fans, increasing draw etc. in a positive feedback loop.

I swear this seemed simple at first, and the more i look into it the harder this becomes. I may just stick to medicine here....

 

[daddles]

If your TE (thermoelectric) module draws more current than your power supply can deliver, then that's a bad thing, as it can make other things not work. There's no real reason for confusion -- think of it like e.g. a blood flow situation at a given pressure. If a pump is putting out X cc/s of flow and you've got n blood vessels hooked in parallel, each allowing Y cc/s of flow at this pressure, then if n*Y > X, you've got a situation where you won't have the flow you want. Since current is a flow of electrons per second, it's an analogous situation with voltage analogous to pressure. This is a bit of a hackneyed analogy and can fail because of real-world nonlinear effects, but at least regarding the two fundamental laws of conservation of mass and conservation of charge, it yields the same thinking. I guess that's a fancy way of saying your medical intuition should be fine for this problem...

Power supplies are pretty cheap, so I don't think you need to scrimp -- buy one with more current capacity than you need for all your loads together. I like to buy roughly twice as much capacity for whatever I'm buying, as a design almost always changes down the road (for electronics, this means other functions get added, requiring more current). It's not a good idea from a reliability standpoint to run things at their designed ratings -- lowering the running vs. rated ratio usually lowers component temperatures, which increases reliability. And you'll know from your medical background that a lot of things have Arrhenius-type rate equations, so little changes in temperature can make big differences (reliability engineering for electronics is strongly attentive to temperature because so many failure modes are temperature-dependent).

 

[dr_deathweed]

That makes MUCH more sense. I was confused on the 1/2 -3/4 power rating, and was taking that too literally as a product of voltage and current. As is I will look for something in the 12V 6-8A range, put together a parts list, and try and draw a simple schematic of what I am planning so far.

Thank you again daddles, your help has been invaluable.