Higher current TEC Drivers (>3A)

[Pack34]

I'm looking at experimenting with a secondary TEC system. I'm working with a low power laser and currently have a TEC that's regulating it's temperature directly. Now, as the laser is cooled using a TEC the opposite side is heating up the chassis that the laser is embedded in. I'd like to see what the effect would be in regulating the temperature of the chassis.

The end goal is to see if output power stability of the laser itself is effected by regulating the case temperature of the instrument. If I can cool and regulate the temperature of the case then the ambient temperature around the laser will be equivalent to the temperature at the base of the laser.

This all may be overkill, but I'm trying to trim off a couple mW of power fluctuation in the thing.

For my question, I can find a few TEC drivers on DigiKey that should run a 3A TEC just fine but I'd like to look at picking up a large one (40mm x 40mm) since it'll be cooling a large chunk of metal.

What would be the best way to go about doing this? What I'm trying to find is essentially a standard TEC driver that uses some external FETs in order to handle the extra current but I'm having an issue finding some. Can anyone help put me in the right direction?

I'm hoping on avoiding just using a generic constant current supply since I'd like to involve a thermistor to provide feedback and set a specific temperature.

[awallin]

take a look at wavelength electronics
http://www.teamwavelength.com/
we use a lot of them in the lab and they work well. around 100-200 $/eur maybe so not worth building yourself if you only need one or a few.
AFAIK all of these are PWM drive.

I did DIY a linear TEC driver (no feedback just an open-loop CC source) based on OPA569:
http://www.anderswallin.net/2014/09/tec-drive-v2/
this takes feedback from a multimeter that reads pt100 sensors, the system is very slow so changing current every 5 minutes is OK.

For your TEC-system ideally you want a large heatsink (like an optical table) where you can dump the heat of the TEC hot side.

HTH

[MarkL]

The LTC1923 is a TEC controller with external MOSFETs.

If you want to get a microcontroller involved, Infineon has a number of high-current PWM bridges in their automotive line (BTM7752G: 8 to 12A, IFX9201SG 6A, TLE8209-2 8.6A).

If you need to keep it linear for noise reasons, a Kepco BOP could be used with just a thermistor and a reference resistor in a simple proportional control loop.

[ajb]

If you only need to move heat in one direction, you could use a standard switching controller in constant current mode and use the thermistor to trim the set point, either directly or via MCU.  Look at external-switch LED driver ICs, some have an external setpoint input that can be used to control current without having to fiddle with the current feedback. 

Or, since you don't necessarily need closed loop control of the current, you could implement your temperature PID or whatever in an MCU and drive a half-bridge power stage directly via PWM.  As long as you're getting adequate heat movement to keep the temperature in regulation and your power stage doesn't cook itself you don't particularly have to care what the TEC drive current is.  This also makes it easy to go full-bridge on the TEC if you need to provide heating as well.