Author Options:

Question on using a Peltier/TEC device for air temperature control - Help please. Answered

I want to have some control over the temperature inside the Orchidium I'm designing and I thought it might be cool :) to use a Peltier Device (device aka module) (Peltier aka TEC or Thermoelectric Cooler). I find I need a lot of help! (Please!) Alright, this isn't a completed Instructable, it's a plea for help, and maybe if the subjects lie in some of your fields of knowledge then we can all enjoy and learn from it. So, the Orchidium I'm designing is an acrylic case 24"W x 18"D x 30"High. It's to grow species orchids indoors in a microclimate, with LED grow lights, proper humidity, air movement and temperature control. (Of course, other critters would like the case, too: poison dart frogs, newts, carniverous plants, etc.. But I'm going to call it the Orchidium.) I've got it all pretty well planned out so that it can be built for a very reasonable price (yes, including the LEDs) and still be aesthetically pleasing and real purdy, too. All planned out EXCEPT FOR THE TEMPERATURE CONTROL. I was looking for some way to cool my case and I stumbled across Peltier devices in eBay. They are CHEAP, costing about $5 or more, depending on the Wattage, etc. The eBay sellers intimated that all you have to do is plug them in and the device gets ice cold. Later, with diligent web-study I learned that actually ONE SIDE of the peltier gets cold, while the other side gets hot. Also, you MUST attach a heat sink and fan to both (?) sides of the peltier. Also, that these devices are not ready to be plugged in; you must attach a DC power supply to them. Oh, another trick that these miraculous devices do is reverse their hot & cold sides when you reverse the polarity of their juice. Ideally, I would like a Peltier device with heatsinks, fans, a thermostat and a DC wall transformer attached... the Peltier/heatsinks/fans would measure about 2" x 2" x 6" and would be mounted in the sidewall of the Orchidium. When the temperature is 65-85F degrees the orchids are happy and the device is Off. But when the thermostat senses the internal temp going over 85F it turns on the Peltier, cold side inside, and so the inside of the case doesn't go up to 90-95F like mine does now; it cools the case a little. Conversely, for someone with chilly orchids or sneezing newts the thermostat would switch the Peltier to hot-side-in to heat the Orchidium a bit. The retail cost for us to buy a Peltier device, 2 heatsinks w/fans and a DC transformer is cheap... roughly $30. The thermostat might be cheap, but I don't know enough about what's needed. If it's too expensive then the Orchidium can do without it. I was hoping I could find an off-the-shelf Orchidium cooler/heater. No such luck. These miraculous Peltier devices are still practically undiscovered -- relatively speaking. People want to use them to cool their computer chips but are hampered by condensation; my orchids welcome condensation. Pathetically, it seems the most common use for Peltiers now is to cool/heat the little boxes on your car seat... they plug into your cigarette lighter and keep your 6-pack cold. Come on! You folks at Instructables can surely help me figure out how to best make an Orchidium cooler with this barely-discovered and poorly-utilized device. I started out a few weeks ago writing to many of the Peltier manufacturers around the world in hopes they might help me in choosing which of their modules I might purchase for my Orchidium. None of them was any help. They wanted to know how many million Orchidiums I planned per year. They told me my basic plan was hopeless or inefficient cost-wise. A Swedish company wanted $800. An American company wanted $500. Some other company wanted $5,000 to $8,000. I wrote back and said I could get a Peltier on eBay for five bucks. The Swedes snottily claimed that their Peltiers were very high quality. No. No way is any svensker Peltier $795 better than ANY other Peltier in the known universe. They both get cold and grow ice crystals on one side. I just need to cool the case A LITTLE BIT, like from 90 to 80 degrees Fahrenheit. I am not trying to make a refrigerator or freezer. The case (Orchidium) is large, at about 7.5 cubic feet, and there is practically no insulation. Acrylic provides a little insulation, that's all. The temp of the interior of the case is derived from the ambient room temperature of your house... and the lights... which is why I designed it with LEDs. There is a constantly-operating muffin fan inside the case to provide air movement for the plants, but it does not provide any evaporative cooling since it's a closed case. So, first off what size Peltier do you recommend... do you think a 40 Watt would be enough, or what? Next, the placement. I envision the Peltier device mounted vertically through a hole in the side of the case. It might be a plan to mount it in the ceiling, but remember that the LEDs take up most of the ceiling. Next, the heatsinks. I confess I'm not totally clear on this, but I "think" that 2 heatsinks-with-fans may be needed, with one sticking out the outside and the other inside the case. I went ahead and got 2 heatsink/fans from Newegg for supercheap ($1 after rebate), but they aren't really what I want. They're actually shaped to fit some AMD chip. What I think I need is a copper heatsink with a flat bottom a little bigger than the Peltier, and fins... and a heatsink fan attached... and some way to attach it to the Peltier, and through the case to the other heatsink. See? Simple... well it should be but I can find nothing. Next, the power supply. I know it has to be DC, but I don't know which brick to get. I did find a bunch of DC or AC Wall Transformers for sale at alltronics... around $10 or so. All that stuff would be enough... at least to test the cooling power. But if we want to go whole hog then the icing on the cake would be thermostatic control of the Peltier. Well, I throw that out in case one of you is sharp in that field.


Here is my experience with thermoelectric units, for a box THAT large you will need a lot more than 40 watts to cool the case especially without any insulation. I have a 1.57" X 1.57" (40mmX40mm) 168 watt TEC that I got off of ebay. It is extremely in efficient as it draws 14 amps max ( a real PTA to find a power supply for) I have never been able to provide the full 14 amps...yet. the closest I've ever gotten was with 10 amps from an old computer PSU, and even when it was underpowered (the PSU extremely overloaded) I had to watercool the Hot side before I got any ice on the cold side. So anyway lets say, hypothetically, that you changed your mind and bought a more powerful TEC You would need a much more beefy PSU for it. An xbox 360 PSU works nicely because it can put out 16 amps on the 12v rail. ( =0) then in order to cool the air in the box you would need a heatsink with large fins so they wont frost up and clog the fan but to heat the box you would need to resist some of the current to the TEC (you dont want to fry your plants do you?) then comes the problem of actually switching the current around, we are talking 12+ amps here depending on the TEC. That means you will also need some fairly powerful relays or transistors and resistors to control everything. So in conclusion, from what I see, this project looks like it could take right off and be a really nice featured instructable. It just needs to be re-planned to be compatible with the TEC idea.

Just as an FYI, 'Peltier Junctions' (aka Thermoelectric Modules) are, as you've already figured out, nothing more than solid state heat-pumps. While they vary in efficiency -- drastically, in my experience -- those in use for Government/Military purposes are EXCEPTIONALLY efficient, being run at anywhere from 3-12VDC, and I've personally seen those which run at the higher voltages and which draw *less* than one (1) amp -- no joke. Availability of a unit remotely *close* to that kind of efficiency, however, is either A) very pricey, or B) Restricted.

Check out this Military unit designed in 1961, with the report finished in 1962. Note the EER, the thousands of BTU's per hour, and that the dimensions were 1' x 1' x 3" -- yes, that's correct: a foot by a foot by three inches (!!!!). Compare the stated EER achieved with that of, say, a comparable refrigerative AC unit with the same approximate BTU hours available to the public *today*.

If that doesn't blow your hair back a little, again, note that this was stuff being done the moment the materials were available, and which the military was already working with/putting into commission in the 1960's (!).

So, to repeat NachoMama, you ARE on the right track; while Tech-King is correct in that the majority of the TECs available (e.g., on ebay, etcetera) are garbage -- they suck HUGE amounts of current, nearly as badly as the early Mil-tech ones did, IME, the fact remains that not *all* of them do....Whether they're publicly available or not is mostly an open secret, if one knows the principle manufacturers, and understands their varying part-numbering schemes (e.g., group XX-XXXX is Govt only; group XX-XXXX is unrestricted).

All comments about them needing to be properly heat-sunk ('cold-sunk') on both sides are correct: without the appropriate amount of area for heat-dissipation/absorption, they WILL be destroyed, or at least have their effective lifespan shortened). Forced convection (fans) are a good idea, as well.

Hope that proves helpful, and let us know your progress :)


I tried the link you have listed and I get a 404 Error. When I try to use just the Navy link nothing there. Do you have any updates?

Although I know even less about Orchids than I do about electricity - I was pleased to read this question since it comes topically close to my own dilemma... I have frequent need to warm/melt buckets of honey. The recommended means (cheap) is to put a 60 watt bulb in an old fridge and leave for 12-24 hours. I'm interested to know if I could mount one or several of these little peltier thingies inside the outside fridge, possibly using the fridge skin itself as a large heatsink (yes??) and then run it/them from a standard car battery recharged with a solar panel. If this CAN actually work.. how does the wattage of the units affect temp/electricity usage? and... could I mount one of the units on the underside of the lid of a honey bucket and plug it into the ciggy lighter of the car to be able to melt during long drives? Any help would be appreciated. tonyreay@kidhugs.com

. It should work. You'll probably want to plan for collecting the condensation from the cold side when you're in the car. . The more watts you have, the faster your 'fridge will warm up - and the bigger power source you will need. I'd install in 2 or more banks and switch them on/off as needed - use a thermostat if you want to get fancy.

Thanks.. how much condensation should I expect - about...? how does this thing function... do I just glue it to a piece of the exposed outer skin by stripping out a section of the interior insulating stuff? How do I know which is the right way to do it.. I haven't been able to find any actually "how to" pages anywhere (that's how I found this site..) I don't understand how to install two banks... or more. I would be willing to pay for actual diagrams and instructions and advice, since this is currently a difficult and expensive task. What do you mean "build a power supply" - regulated or otherwise..? AAAaaaarrrrrrggggghhhhhhh I feel SO DUMB sometimes..

1) expect a lot: that way, you will not be or be pleasantly surprised by the condensation 2) you attach it with thermal adhesive to one side of the skin, cold side touching metal. you use more adhesive and mounting screws to secure a decent size aluminum heat sink/fan assembly to the hot side 3)peltiers are powered of 12 volts. your wall has 110 or 220 volts in it. you need a power supply capable of around 5 amps. it depends on the peltier.

this is also a reply to NachoMahma - since you are both helping me stumble towards enlightenment here... I BELIEVE that I'm actually getting to grips with this system but, as always, have a few more questions... 1. If I attach the cold side directly to the outer skin... does it also need a heat sync and a fan? 1b How do I know which is the cold side - just wire it up and feel it.. or are they marked in some way? 2. How big a heat sink do I use on the interior "hot" side? can I just pull one from an old car cooler/heater thingy? What's a "decent" size? 3. Does a fan on the interior simply disperse the heat around or is it more designed to keep the peltier from bursting into flames? 4. Assuming that I only use one unit (mainly) can it indeed run from a car battery, kept up to snuff with a solar panel? 5. Is there a limit to how long one can continuously run juice through the unit without it ... breaking, I guess, in some way? A little clarification about some other comments in this thread please. It has been noted that one is ill-advised to buy units off'n ebay since they are inferior quality.. where does a neophyte get such things here in the UK then? It has also been noted that one shouldn't use thermostatically controlled switches to turn the current on and off... is this true? It certainly would be great if I could, y'know? Hotdarn this baby might actually work.

1) It may not _require_ one, but it will work MUCH better with one. The skin of the 'fridge may provide enough heatsink. Since you are not moving a lot of heat and it's not a critical application, I'd try it without and see what happens. 2) See t-k's answer. Since you only need ~60W, I'd go with unit(s) designed for CPUs/GPUs/&c - fan and heatsink already attached to the hot side. 3) See t-k's answer, but I'd rate them as pretty much equal. 4) If you are doing this more than a few miles (or whatever comfortable walking distance is for you) from civilization, I wouldn't recommend using your car's starting battery. A car battery should work fine, but as t-k points out, TECs can pull quite a bit of current. You could strand yourself in the boonies if you discharge your car battery too far. Maybe an extra battery in the trunk? If it goes dead, your honey is a little thick, but you can get home. 5) With the proper voltage and heatsinks, they are continuous duty. I had one on an old 68030 Mac that ran 24/7. Keeping the voltage below the max will help extend the life. . . I haven't heard of not using switches/thermostats, but I'm definitely not an expert. My experience is limited to the aforementioned Mac, some industrial "cabinet coolers," and a couple of small units that I fiddled around with on my workbench for a few hours. . > Hotdarn this baby might actually work. . I believe it will. It may not be terribly efficient when you get done, but it ought to work. . . PS: when you select a glue, not only should it conduct heat well, but it should be able to put up with the thermal cycling (expansion/contraction). For your application, neither one is critical, but keep them in mind.

a good glue is thermal adhesive by arctic silver, but is a bit pricey. look for gpu peltiers. cpu peltiers start at 80-120 watts.

ah now then.. we're getting to the real meat and potatoes of my ignorance - electronics in general..... watts, amps and volts and their inter-relationships in particular. I see lots of peltier units advertised from 91 all the way to 320 WATTS... And I have NO idea what that means to my particular project... remember, I'm trying to keep the interior temperature of a well insulated refrigerator around 60 degrees for a few hours, long enough to melt a bucket of honey and warm it thoroughly.... and I'm trying to run it off a 12 volt deep cycle battery. Does wattage equal temperature? will more wattage run the battery down faster? what are GPU peltiers and where would one look for them? Assuming that I get a good heatsink for the inside hot side, can I simply botl it to the fridge skin nice and tightly with the peltier "wedged" between the sink and the skin or does it HAVE to be glued in place? Don't give up on me now guys, we're almost there...

> I see lots of peltier units advertised from 91 all the way to 320 WATTS... And I have NO idea what that means to my particular project... . This is the maximum power that the TEC will draw at its' rated voltage. As with a light bulb, you can lower the voltage and lower the power draw. . > Does wattage equal temperature? . In general, higher wattage TEC will move more heat; so, yes. . > will more wattage run the battery down faster? . Yes . > what are GPU peltiers and where would one look for them? . For graphics cards (Graphics Processing Unit). Get them the same places you get CPU coolers. . > Assuming that I get a good heatsink for the inside hot side, can I simply botl it to the fridge skin nice and tightly with the peltier "wedged" between the sink and the skin or does it HAVE to be glued in place? . As long as there is intimate contact, you don't need glue, but getting the proper metal-to-metal surfaces is probably out of the range of your capabilities. Plus the 'fridge skin will probably warp slightly from the temp diff and cycling. The "thermal glue" that t-k mentioned should be just what you need.

OK, thanks ... the only question I have left (I think) is whether the "thermal glue" will conduct the "cold" from the tec to the fridge skin and thus allow it to be dissipated without the need for an exterior heat sink... oh, and does the interior heat circulating fan have to be right next to the heatsink or can it be anywhere within the fridge compartment. that's it... no more... honest and really you guys, thanks an elephantful for all the help and advice and ether hand-holding. I'da bin lost withoutcha.

the thermal glue will conduct the cold to the fridge wall. the fan need not be on the heatsink, but MUST be close enough to ensure good airflow to the chip. just buy a second fan to circulate air in the fridge. the chip will last longer.

90 watts should be sufficient, and more watts will heat more and use up the battery faster. as nacho said, the heat sinks for gpus are what you want. like i said earlier, cpu heat sinks are a lot more powerful than needed.

1)cold side does not need a heat sink a fan 1B) plug it in, count to 20 fell both sides. keep counting. by sixty, the unit MUST be shut off. you cannot use a peltier for over 60 seconds without a heat sink on the hot side. 2) depends on how many watts your peltier is. the bigger the better. dont use heat pipes 3) more the former than the later 4)you could run it from the car battery, but remeber, these things use a decent amperage 5) yes. remember, the heat from one side just moved to the other. assuming you have a good fan/heat sink combo, this time limit is very long. however, dont leave it on continuously. when you can no longer touch the heat sink, its probably too hot 6) check on computer modding and goldmine electronics web sites.

just checking... the condensation would all be on the OUTSIDE (cold side) of the fridge, right?

yeah, frost and condensation on the cold side

> how much condensation should I expect - about...? . Lots of variables - humidity, temp, &c. I'd plan for about a fluid pint per hour. Guessing that it will be closer to a few floz/hr. . > do I just glue it to a piece of the exposed outer skin by stripping out a section of the interior insulating stuff? . Probably not the most efficient way to do it, but it ought to work. Try to find a glue that conducts heat well. I can't come up with anything better. A heatsink and fan on both sides will help. . > how to install two banks... or more . Just wire your source to two or more switches and then run your wires from the switches to the different Peltiers. Turn them all on the get the temp up fast and then run just enough to maintain the temp. . > What do you mean "build a power supply" - regulated or otherwise..? . If you will get Peltiers designed to run at 12-15VDC, you can use a cheap, unregulated power supply. For your purposes, you don't need an expensive, stable PS . If the voltage swings a bit, it's no big deal - as long as it stays within the limits of the Peltiers. . > I feel SO DUMB sometimes . You're not dumb, just ignorant. Ignorance we can fix; can't do much with dumb. ;)

a battery may not be sufficient for this purpose. you will probably have to build a power supply. on the plus side, it need not be regulated. otherwise, it should work. watch out for condensation though.

the 7812 is definitly not capabel of over 2 amps. the peltiers i work with are for computers. they need at least 150 watts on an over clocked rig. most computer power supply, unless over 500 watts, are either too weak, too cheap or just already overloaded, and cannot handle an added peltier. i have never used, and have no access to, military peltiers. like everything they own, their peltiers are more efficient.

OK! i will to help you!. firts point you need create an 2 stage thermoelectic cooler. how? you may join 2 peltiers modules one more powerfull and the other most weak, ex: tec 1 230w, tec 2 130w. since tecs. absorb the heat from a side to other. one side cold and one heat. the tec 1 is used for absorb the heat production on tec 2 and the tec 2 will burst in eficience. becouse thermoelectic technology employ heat transfer by ambiental diference. on keep the heat side below ambient temperature will burst the tec 2 eficience and you get an temperature diference near 50/60c . but you must add eficient coolin for tec 1. since it will absorb the heat from tec 2 and disipate pluss the own heat production of tec 1. second the heat sink must be large and good air blow on tec 1 heat side. and not so large with a simple fan on tec 2 cold side. third you could make an switching circuit for pwm control. if use an 2 stage ocilator you can control 2 tecs with one load not two load ex: tec 1 and tec 2 conected in paralel will boost the power consuption, but if you switch one and other you will get a load only at time.

circuit ocilator 2 stage = 100hz tec 1= hz1, tec2 = hz2, tec 1 = hz3, tec 2 = hz4, this is a way of switching two tesc with 2 stage ocilator using the power fet irfp054. AN SUPER POWERFULL FET.
i hope this help you!!!!!


9 years ago

Hi speak4nirmal, I am sorry but I am not an expert with peltiers; my article in Instructables was in hope that someone would give me practical advice for my peltier project. But I will give you what little information I have learned... Are you sure your peltier requires 12V? I thought they usually required 9V. I do not know about amps. When current flows through the device, one side should get hot while the other side gets cold (with ice immediately). However, heatsinks are VERY important. You must have a heatsink with fan on both sides of the peltier... the cold side will not get very cold unless the heatsink-fan on the hot side removes the heat from the device. To experiment: place the hot side on a metal block (or something like a "heatsink") and then see if the cold side gets cold, with ice. For a power supply, I do not know what is "best", but I plan to just use a common AC-DC converter "brick" (9V), like those used for many electronic devices. I don't know about amps, but someone told me a peltier will draw much amperage... I (we) will just have to experiment. I do not think running time is a problem. I'm sorry I can not be more help. Best of luck to you!


10 years ago

I am building a rabbit hutch heater/cooler using a TEC/Peltier so thought I would drop in what I discovered while reading up on TEC's.

Reverse polarity and you reverse the hot and cold sides -> cold becomes hot side and visa versa

They are most efficient at 50% rated voltage or there abouts. most efficient when run using a PWM (pulse width modulation)power supply with reduced current (or increase your TEC size and under volt it to use same amps but more cooling)

they generate alot of heat on the hot side, heat taken from the cold side plus the heat generated by the TEC's own inefficiency = lots of heat to move.

You will only cool as much as you can remove the heat from the hot side, a large heatsink and fan is a must on the hot side, maybe even water cooled for better results. The faster you move this heat the colder the cool side will get upto manufacture rating.

TECs like to have presure on them, usually a cold plate (simply a copper plate just a bit larger than the tec) is used to bolt to the hot heatsink, suitably insulated mounting screws required (stop the heat conducting to the cold side through the screws) and the cold heatsink mounted on the cold plate (unless your cold heatsink has mounting holes outside the tec size that can take a bit of force)

thermal grease is used between hot heatsink and tec, cold plate and tec, cold heatsink and cold plate, the type used for CPU coolers.

if the cold sink is to be colder than ambient condenstation will arise and needs to be dealt with, some tecs come pre sealed, its best to get these dispite what plans you have to deal with condensation, if it gets in the tec it will blow.

never run a tec with no heatsink, you will most likly distroy it.

I forgot to mention -- and this applies across the board for all TEC units/Peltier Modules -- something the Military (and Contractors, and any/everyone who worked with TECs for any period of time) quickly came to discover was that Bigger Is Not Better -- period. TECs beyond a certain size (optimum being > 2", which is why the 1.5"/1.4" range is so commonly seen) have a MUCH shorter life, properly sinked or not, simply due to the differences in thermal expansion between where the P/N junctions are connected to the copper/silver/alloy metal bonded to the ceramic housing (lapped or tinned), and, last but definitely not least, the coefficient of expansion of the solder connections in the TEC itself. While thermal cycling is bad for *all* electronic devices, because the Delta is so extreme with solid-state heat-pumps (60-70 deg.) it can and will impact them faster...And the larger they are, the greater the expansion/contraction force(s) become.

Short version: the bigger they are, the shorter they'll last' the smaller they are, the longer they'll last.

Design-wise, it's better to use multiple, smaller ones (e.g., instead of one meant for 300 watts, using three 100 watt TECs is recommended) within reasonable limits, almost *always* provides for higher yield, greater robustness, and (if run at the "ideal" of ~30-40% of what each TEC is rated at) has been repeatedly shown to be MUCH more efficient (hence a higher EER for a device designed with four, 1.4" TECs @ ~150W each than for a device designed with two 300W TECs, with the least-efficient design being one gigantic 600W (!) TEC).

Provided all else is equal, with the appropriate thermal design considerations, solid and well-machined interfaces, proper construction and input power control, and the end-product device could, in theory, outlast *you* with the four-TEC devices in mind (I know of several that have been and still are running, and subjected to thermal cycling, for decades now); doing the same with a single large unit? They'd be dead in a few years, more likely. And, if anything wasn't done *exactly* to spec (e.g., not lapped within 0.001" evenness) they'd last shorter than that, I'd wager.

N.B.: While swapping the polarity is entirely possible and not directly injurious to the TEC, electromigrative elements *will* set in after a period of time, thus aging the TEC, although IIRC, nowhere nearly as drastically as running it without proper heatsinking. Also, the use of the larger TECs _could_ prove viable, although only in a situation similar to a Klystron-tube setup, where they were kept running *all* the time, with no thermal cycling ever; in theory, they could last just as long as the smaller ones...Although that's just a theory.



. You're on the right track. A small DC supply with a simple thermostat should do the job. You will need to size the PJ so that it doesn't cool the Orchidarium too fast. I'm guessing that you will want a heatsink (and possibly a small fan) on the cold side, in addition to the required ones on the hot side.
. One of the great things about PJs is that if you apply a few volts they cool a little; more volts = more cooling (up to a point). Another nice feature is that if you swap the polarity, the heat flow changes direction. Making use of these features will require a much more complex thermostat.

small???? are you joking????? the peltier effect works only with high amperage. at 12 volts, you need around 25-30 amps for a decent peltier. small my foot. plus, no vreg (7812) can handle that current. don't forget, the hot side of the peltier needs a huge heat sink. unless the heat sinks in your supply are solid copper with mirror lapping and arctic silver thermal past, I'm skeptical its effective enough.

The Peltier (aka Seebeck) effect is not subject to the qualifications you've described (e.g., minimum of 12 VDC and at 25+ Amperes for a 'decent' one); in physics, the effect was first noted in the mid-1800's, and was a followup on the Thermocouple effect, which some French scientist (Peltier) had the silly notion that perhaps the Thermocouple effect was reversible (i.e., instead of having a temperature differential using two dissimilar metals create a measurable voltage, what would happen if one applied a voltage -- would it create a temp. differential between the ends?), which it was. This was of little use while materials were limited to copper & iron (and alloys)....Then, along came doped silicon and, well, the rest is history.

Provided one understands that even the "usual" expected voltage drop across a P/N junction doesn't apply to 'Peltier Junctions' (more appropriately called ThermoElectriC modules, or simply TECs), the average TEC being composed of 127 Diodic junctions (that is, 127 doped P/N junctions), one can easily find that most of the well-made modules available from Marlowe, Melcor and other common companies are rated at far less than 12VDC -- I have some units which are rated at (IIRC) a max voltage of 3.37, and having accidentally connected one to a +5VDC regulated supply some years back illustrated to me what over-volting them can do :(

See my comment above, and check out the link: it is entirely legal for me to post it, and it is for public release, yet clearly demonstrates the capabilities of such things even in the most *primitive* designs, and with a superabundance of resources (e.g., doing ideal calculations for thermal designs with the presumption of an "infinite" heatsink/zero-HSR is, obviously, inaccurate to the real world; however, for a Naval Submarine of fixed size in virtually ANY underwater location in the seven seas (hah), well, what better to serve as a possible "infinite" heatsink than the Ocean...?) Not even the unshielded, full-throttle-but-not-going-boom Nuclear reactor on a modern Submarine is going to give off enough calories to boil the very Oceans themselves (though it would make things suck for those nearby, no doubt).

Regarding the 7812, my memory is a little weak at this point....Is it rated at 1.5A or 2A? If the latter, it could easily handle a TEC configuration I made the other day with three non-Military-grade TECs in series, each drawing 4 volts DC, with the max current draw of circa 650 mA each, until they achieved Delta, after which the draw stabilized to circa 560 mA, which meant the total power input was 12vdc @ 2A for ~30 seconds, after which it was 12vdc @ ~1.7A.

The average Delta for a TEC is ~60 degrees Centigrade (though I've seen higher, so 72 degrees C or more is possible), and that's with only a *single* stage TEC -- not multi-stage.

As for your ideas on HeatSink requirements, well...I'm not trying to be argumentative, but I'd suggest doing some research into Heat Sink and Thermal Engineering & Design before anything further is said.

Back to the thought of the "30 amps" needed, see the note on page 7 herehere. If you'd rather not read it, in summary, it is the design of an AC Unit for a Navy Submarine, which had a (for the time) reasonable EER, put out a few thousand BTUs/hour, could withstand high-pressure/corrosive environments, and measure a foot by a foot by three inches, and had an operating current of 35 Amperes, which is admittedly a pretty large draw...But not for something doing what that could do...And again, that was in the early 1960's, and Telluride and other doping materials/processes for Silicon semiconductors have, I'm sure you would agree, improved quite a bit in the last few decades ;)

Now I have to go look up what the ratings are on the 7812 Voltage Regulator are; I can even recall the case style, but little else about that stock 12 Volt regulator. I really *am* getting dumber as I get older :(

Best Regards,


tech-king, do smaller, lower-Wattage peltiers (50-80W) still draw such amps? There's gotta be a way. How do those little car 6-pack coolers manage it? RE heatsinks, can you direct me to where I could find copper heatsinks to fit approx. 40-60mm peltiers? I have hunted & hunted. Newegg sells lots, but not what's right for Peltiers. I found one guy on eBay with some for $3, but only aluminum. Condensation: bad for PCs, but GOOD for orchids. One bright spot in the dark Peltier jungle.

the peltier effect works best with high amperes. those car peltiers are both weak and so-so amperage; the five volt usb beverage coolers draw several amps and need over 30 min to cool a styrofoam cup or can of warm soda i would actually recommend going with aluminum. while copper is a better heat sink, and the overall heat sink size could be smaller, copper would weight a lot more, stressing the acrylic sides.

. Check http://www.the12volt.com/ohm/ohmslaw.asp to see how Volts, Amps, and Power are related. At 15V, an 80W device will pull almost 5 and a half amps.
. Nothing wrong with Al heatsinks. Not as good as Cu, but don't just dismiss them. You're not trying to move that much heat.
. Not sure exactly how many Watts you will need, but the only heat sources I see are the LEDs and the surrounding air. The plants and newts don't put out much heat. :) You're only going to be changing the Orchidarium temp a few degrees and it's not a problem if it takes 3-4 minutes to get there.
> Does the Wattage on a PJ remain constant or vary?
. It will change with the voltage applied. More Volts = more Watts.

. LOL I guess it depends on one's definition of small, but I see your point. . I don't think he is going to need the large panel that is shown in the pic above. A 100W (or less) unit should work for such a small space.

also, although its hard to see, that large aluminum object attached to the peltier is a heat sink; pretty big ehh?? imagine the one needed for a bigger peltier. no wall wart psu could power this thing. don't even try. a problem occurs; peltier is on for long time, frost builds up. then circuit reverses, frost melts, drips down on plant, table etc. plus, lest the condensation issue be overlooked.

. Granted, there aren't many wall warts that put out the current required, but a 15VDC/150W PS is not that big of a deal. Especially if he is a serious hobbyist, which his designs/ideas lead me to believe he is. . I think you are overestimating the amount of heatsink/fan required for a small unit that isn't required to move a lot of heat in a hurry (don't want the temp to change too fast for the sake of the critters). Something the size of a CPU heatsink/fan should work. The one on my Pentium D would probably be more than required. . Knuten claims that condensation is desirable.

true, a linear supply would work perfectly. no vreg would even be needed. im not exxagerating that much; whatever heat you pump out from one side ends up on the other side. condensation on the plants may be good; on the table/desk/psu outside the box its preobably not.

Sorry to weigh in late but I just wrote an Instructable on how to move & delete an iBle. You guys are over my head but I shall study what you say and try to catch up. To make it easier for me, can we take Wattage first? As a stab in the dark shall we say 45.6Watts, 12V, 127 Couples, Umax(V) 15.2, Imax(A) 3, 40x40x5.4mm (this is one on eBay now for $2). I want to know if something of approx. that Wattage would be sufficient to cool my case. We wouldn't mind 46W nearly so much as we would buying electicity for a 350W cooler. If 46W is too small, then how about 60 or 80W? Does the Wattage on a PJ remain constant or vary?