The big TEC driver thread!

[weartronics]

I wonder whether all this effort is really necessary if one uses the laser just as a lamp, essentially....

The power output of common 445 nm and 640 nm diodes is not significantly affected by small temperature fluctuations. Fast control loops of any type are purposeless here, because the thermal mass of a typical diode mount acts as a big low-pass filter anyway. Fortunately, proportional control is perfectly sufficient for this application, does not risk overshoot or undershoot, and does not require tuning any loop parameters. I posted the schematic for a very simple proportional TEC controller earlier in this thread.

A PID controller like the one being discussed here would be appropriate for things like temperature control of DPSS crystals, which are sensitive to slight and rapid temperature fluctuations. Fast control loops are only possible with a small thermal mass, which is why the crystals are often nestled in indium foil directly on the TEC rather than using a bulky mount. I'm impressed by the open collaboration going on here and I think that the design so far is a valuable community contribution.

[Solarfire]

Have you factored peak current in the inductor chosen? Peak current will be higher than average current.

Shouldn’t be any issues here, as long as the 3A of the H-Bridge are not excessively exceeded.


I added another trimmer on the free IO of the tiny24.


Updated schematic and PCB (see attachments).



I’ll wait a few days after posting before building the PCB’s in case anyone can find some form of improvement which should be implemented.


Cheers!

[dnar]

Shouldn’t be any issues here, as long as the 3A of the H-Bridge are not excessively exceeded.


I added another trimmer on the free IO of the tiny24.


Updated schematic and PCB (see attachments).



I’ll wait a few days after posting before building the PCB’s in case anyone can find some form of improvement which should be implemented.


Cheers!

Looks good mate. My only comments are as follows:

1. Why not use 0.1% resistors in place of R9/R22? Reduces cost and also the need to calibrate these values.
2. Why not simply use test point loops for X2-3 and X2-4? http://au.element14.com/vero/20-2136...100/dp/8731195

Otherwise, good job!

[Solarfire]

Why not use 0.1% resistors in place of R9/R22? Reduces cost and also the need to calibrate these values.!

Yeah.. I thought about that, but a simulation isn't 100% if the values are verified on the prototyp then these could be substituted.

Why not simply use test point loops for X2-3 and X2-4?

Thought that maybe the one or other would prefer permanent monitoring of the temperature via a panel DVM..

[Badpip]

I’ll wait a few days after posting before building the PCB’s in case anyone can find some form of improvement which should be implemented.
Cheers!

Looks good, but it's good practice to add a 100nF decoupling cap to an opamp.

/Thomas

[ogoun]

Hope this hasnt already been asked, but how about adding support for RS232 comms? For those of us who dont need it, it can be left off/not populated. Otherwise, a max232 or similar could allow remote monitoring as well as serial loading of operational parameters. For such a flexible tec controller, this seems to be a useful option.

Cheers,

Pete

[dnar]

Yeah.. I thought about that, but a simulation isn't 100% if the values are verified on the prototyp then these could be substituted.



Thought that maybe the one or other would prefer permanent monitoring of the temperature via a panel DVM..

Good ideas.

Hope this hasnt already been asked, but how about adding support for RS232 comms? For those of us who dont need it, it can be left off/not populated. Otherwise, a max232 or similar could allow remote monitoring as well as serial loading of operational parameters. For such a flexible tec controller, this seems to be a useful option.

Cheers,

Pete

That's taking it to another level again. This would allow those that know how to baseline a process and tune a PID loop maximum flexibility.

[ogoun]

Good ideas.

That's taking it to another level again. This would allow those that know how to baseline a process and tune a PID loop maximum flexibility.

Agreed. The hardware is simple enough, so I would suggest adding it to allow the design/pcb to support it without needing to kludge up a hardware sub board. Should be no more than about 6 components (max233, 4 caps and a DB9F)

For those who dont care, they can leave this bit unpopulated. Also, the firmware can stay as is, with the addition of serial comms as and when needed (at a later stage?)


My main interest in this is for a TEC controlled variable temperature water bath, and I want to control everything via PC (liquid level, temp, pumps, etc), so serial control is important to me, more so than local control/display. I am also working on a TEC based fridge, and want to remote this as well.

I imagine there are a HEAP of projects/designs for stuff that needs to be temp controlled, and would benefit from a small TEC controller module that is serially controlled. This could be used as a building block for more complex projects, such as the ones I mentioned above.


Cheers,

Pete

[Solarfire]

Looks good, but it's good practice to add a 100nF decoupling cap to an opamp.

/Thomas

Yes.. That's why more heads are better than one.

[Solarfire]

@dnar

The standard for the onboard ISP Interface, is that male or female? Looking at pin positioning in AVR Datasheets it would be male?

What do you think on the RS232, you're the softy on this one? I have no Idea how much more work that would be on your part or if you even have time for such. Not to mention things like code space. This would be excellent in the developing process of lasers with such functions as PID tunning, Temperature readout for logging, TEC on/off?

cheers!