Supply Power to a TEC?

[LazyBeam]

How does one supply power to a TEC for cooling purposes?

Looking at powering a couple of these (one on each side of the laser) for great cooling and the TEC shows max current and voltage as 2.1A/3.7V... so how do I ensure it does not go over 2.1A and 3.7 volt?

Do I just give each one an LM338t set to 3.7V and hope it doesn't suck more than 2.1A???

 

[Coherent Light]

I saw where someone was using a wally-wart plug-in charger

Coherent:

 

[JLSE]

If you are looking for the coldest setting and not controlled temp, you can simply use a PC P/S.

For small TEC's I use the 5v lead, and for larger, the 12v leads. I have run even small TEC's way over their spec'd limits and never blown one...

Though I wouldn't recommend this unless you are willing to take the chance.

I think the PC P/S 5v gives 2-3amp @ 4.85v, but cant remember the max amps for sure. Probably 5A at most. Even the smallest TEC I have took it with no problems. Just don't run it like this unmounted, or without thermal compund on the hot side.

 

[LazyBeam]

So I only need to regulate the voltage? I'm thinking of using 2X LM338t to drive 2X peltiers each (4 total TEC). The LM338t can be set up for voltage regulation. So the peltier should draw it's own power? I assume the rated current is what it draws at that particular voltage?

 

[JLSE]

Im not too sure, but what I had noticed was even putting 12v into a 3v tec, the voltage would drop to 3v (approx). Depending on available current I believe is the deciding factor on what may be too much...

Are you looking to set a certain temp? Or are you just interested in ice cold temps?

You have to regulate the current to keep it level. But again, it depends on what your goal is.

If you run it at its max or beyond, all you need to do is keep flowing the heat away from the TEC.

To play it safe you could simply use a lm350 (5A max) and use 2 x 1.2R 1watt resistors in parallel.

@.6ohm you should see 2.08A You should be able to run the power off the PC power supply still with the 5v leads, or the 12v, but the 12v will add heat to the circuit. Id try the 5v leads and go from there.

If 2.08A is too much, you could move up and use 2 x 1.4ohm resistors in parallel for 1.78A

or...

You could use amc7135 chips. They are low drop out and 6 chips paralleled will give you 2.1A on the button.

 

[LazyBeam]

So the stragegy is to limit the current with a simple IC regulator huh?
If I power an LM338 with 12VDC input, for example, and set the output to 1.5A will the voltage output be whatever?
Would I just use diodes or resistors to drop whatever voltage happens to come out of the regulator safe levels?

Well, the labby is getting 12V DC input. I'm hoping for a 445nm @ 1500mA with enough cooling for 100% duty cycle.
The current strategy I have drawn up is water cooling using a water-jacketed diode pocked machined myself.
Then a small pump and 80mm radiator. A simple setup like that would still run ~$150 new. That's pricey but unique.

I recently started investigating TEC cooling. I found a site selling 20.5x20.5mm peltiers for ~$2.
Basically <$10 to for 4X TEC cooling. I have a feeling 4 peltiers can cool better than liquid cooling.
Not as cool as liquid cooling but the price is hard to pass up.
I guess I could make the TEC cooled labby and sell it to fund the water cooled labby.
The decision boils down to... is water cooling awesome enough to justify +$125 for the build. :undecided:

 

[JLSE]

I actually just updated my avatar... That is a 2.8W 445nm with a 12v 100+W TEC mounted below it. 100W is overkill and would probably form ice on the head

TEC vs water cooling.. id go with the TEC. You can get the temp much lower if desired, and the change in temp is just shy of instantaneous. Though keep in mind the whole condensation issues... There is a chart that shows temp vs relative humidity and the dew point (lack of a better term) that will keep you away from droplets forming on the laser head.

The voltage will more or less adjust itself. As long as you use a current regulator (or a voltage regulator in current regulation mode), you should be fine. You may find that the full 2A is too much and 1A is too little. You may want to make a TEC driver that allows you to sweep through the range of applied current to find your best temp on the head. IIRC I read somewhere that these LD's operate best a 29c or so..


Sounds like a cool setup, be sure to post pics as you go..

 

[LazyBeam]

So what your saying is that dry nitrogen purge and flow is needed across the LD head to keep it <0C with a heaters on the optics to keep them clear...

j/k...j/k lol...
but no, really, I could set up dry nitrogen flow if needed.

Thanks for all the help! +1

 

[JLSE]

So what your saying is that dry nitrogen purge and flow is needed across the LD head to keep it <0C with a heaters on the optics to keep them clear...

j/k...j/k lol...
but no, really, I could set up dry nitrogen flow if needed.

Thanks for all the help! +1

LOL, give it a couple of years when projectors bring about 24W single diode packages :drool:

Till then glycol or just a plain TEC should do just fine :beer:

 

[HIMNL9]

Powering up TEC plates is relatively easy, and the circuits depends mainly if you want some sort of regulation feedback, or if you just want to keep them powered with a stabilized current (they work in current, not in voltage), regardless to the effective temperatures.

If you want an extra-precise and extra-fast response feedback and temperature regulation, you can also buy a custom module or use an Arduino board, if instead you want just to experiment, or have a working, cheap circuit, here are some

First one, no feedback, is my version of the current sink LD driver ..... it's designed for laser diodes, but can be used for any load that need to be powered in current ..... simply don't mount the red components, that are planned for the LD modulation, connect the TEC plate in place of the LD, and use a power supply voltage a bit more high than the one required from the TEC plate (as example, if you have a 4,5V TEC plate, use 6V, if you have a 10V plate, use 12V, and so on, but never go under the minimum working voltage of the op-amp):

Second, my old PWM driver based on the 555 IC and with a temperature feedback, place the NTC on the cold side of the assembly (preferably on the object, not directly on the TEC plate surface), and it simply shorten the distance from the pulses when the cold side temperature increase (increasing the medium value of the voltage), and elongate them when the temp decrease (decreasing the medium value), keeping the temperature relatively constant (not a true thermostat, but is the more cheap PWM alternative that i've designed til now ) ..... for change the frequency of the PWM pulses, if you want it more high, the 10uF capacitor connected from the GND and the pins 2 and 6 can be reduced, til 1uF, without too much problems:

And, if you want also to place some sort of control on the hot side, you can use one of these cheap solutions for change the speed of the fan placed on the cold side heatsink of the TEC (they increase the speed, so the airflow, when the temperature of the heatsink rise, proportionally ..... i made 2 versions, so regardless if you have N-channels or P-channels mosfets, you can use that what you can find more easily)

Just as suggestions


EDIT: i changed the 555 TEC driver a bit, for improve it. My version was working good, but another user had problem of instability, replicating it, so i made a pair of tests, and improved it, making it working more "smooth" and making the "on" and "off" parts of the PWM signal independently controlled ..... need just to add an 1N4148 (or any other silicon signal diode) between the pins 2-6 and 7, with the cathode connected to the pins 2-6 (is the one drawed in orange) ..... now, the trimmer control the "on" part of the signal (active pulses), and the NTC control the "off" part (pauses between the active pulses) independently, making the circuit working better also with critics and/or unstable versions of the 555 ..... also, changing the capacitor from 10uF to 1uF, make it work more stable