[QUOTE=tocket;78522]That's a brilliant idea! Makes it possible to control the TEC with just a microcontroller and H-bridge; no need for a thermistor.
Dr L, just pmed me with the other problem. Switching noise! Draht! foild again....
Steve
[QUOTE=tocket;78522]That's a brilliant idea! Makes it possible to control the TEC with just a microcontroller and H-bridge; no need for a thermistor.
Dr L, just pmed me with the other problem. Switching noise! Draht! foild again....
Steve
You can't get a much better linear fade with direct injection than current sense control, assuming the diode threshold is set right. What I see optical feedback is useful for is compensating the changes and nonlinearities in conversion efficiency in DPSS lasers at the different modulation levels.
go over to avrusb and add the usb. With a small atmega, that gives you 6 analog in.
Steve
What I wanted to see gone is some of the lag from the 800 microsecond ND upper lasing state storage time.
Steve
Right, exactly- by closing the loop, you can get a much smoother result.
My current design uses two of the analogue inputs on an AVR- one to take the modulation signal from the controller card and one to take the output from the photodiode. The photodiode's linearity isn't perfect but it's well characterized and the software in the AVR executes a self-calibration when it starts. At this point the 15-second time delay I have on my projector still hasn't opened the shutter so the few milliseconds (it's not timed accurately, since it doesn't matter) pulses it uses to determine what the minimum and maximum points of the output curve equate to on the diode reading aren't a problem. It doesn't need to communicate with the PC at all during this process but I have it spitting the numbers out on the serial port for testing purposes.
I haven't got it all nailed down yet (for example, spill light from the other lasers in my rather small optics box interferes with it) but I'll let you know when I do
I would like to see versions of big 300+ watts TEC cooling for water chilling,
the heat exchanger,TEC array,power supply.
Horribly inefficient though. For that amount of cooling it's probably better to use refrigerant, and a fraction of the energy input.
Re that red diode and optical feedback idea, I wasn't convinced it was a goer, hence the 'if'.. but, what if it mode hops like Rohm's do at levels over 150 mW. That hopping changes power out by around 20%, so might it help with that if you want to push for 200 mW?
Well, I'm not convinced that the reds really need it, but it turns out when I thought a little more, the advantages are pretty substantial:
- automatic setting of gain curve, max and min points
- tunable delay on a per-channel basis (so the slow DPSSes come on at the same time as the fast red, reducing the mismatch)
- linearization on the same scale as the other colours
This could be (assuming it freaking works... right now it's just sort of sitting there sulking) just the thing to get my RGB to make real greys! It turns out that EyeMagic have already built a similar, but much more expensive and sophisticated thing:
http://www.eyemagic.gr/laser-iris-colorsafe.html
I was instantly convinced by their nice smooth greyscales.
Nice to see some action here!
Maybe the optical feedback project should get its own thread?
Does anyone have good sources for small TEC's they want to share?
The ones on ebay are usually big ones, 40x40 mm or so.