The big TEC driver thread!

[Solarfire]

Depending on price and when, i'd take about 3-5..

cheers!

[LaNeK779]

depending on price and when i'd also be wanting a couple

thanks

[dnar]

I am still working on the beer TEC controller, the micromatch connectors arrived Monday but my AVR-ISP programmer is not working. I ordered another and it will arrive in a few days.

[Solarfire]

I am still working on the beer TEC controller, the micromatch connectors arrived Monday but my AVR-ISP programmer is not working. I ordered another and it will arrive in a few days.

No rush, I already got a shit load of those lying around here and they’re not going anywhere. I just like to be flexible and stay supplied with such components.


Something else though, ya think one of those cheap 30$ ISP programmers on ebay are up to the job of flashing these controllers? You got any recommendations?

[LaNeK779]

I am still working on the beer TEC controller, the micromatch connectors arrived Monday but my AVR-ISP programmer is not working. I ordered another and it will arrive in a few days.

no problems at all lad. i also wait for the irf540, and would like to toss a couple of die4chills in the mix just for flavour

got swamped at work lately, no time to fiddle with my lazors...

[dnar]

No rush, I already got a shit load of those lying around here and they’re not going anywhere. I just like to be flexible and stay supplied with such components.


Something else though, ya think one of those cheap 30$ ISP programmers on ebay are up to the job of flashing these controllers? You got any recommendations?

I have never tried the cheap eBay programmers. Always used the AVR-ISPII. Works a treat, having said that my old work horse simply won't respond to avrdude.

I tried the Arduino Mega-ISP firmware and that does not work, or I can't get avrdude to recognize it. In the end I just ordered another Atmel unit @ $60. It ain't worth messing around. Changing subjects, I purchased a cheep Freescale programmer the other, not bad not bad.

[dnar]

well, i can't think of the use of a 60w tec to cool a lazor

something small, around the lines of solarfire. my tecs are no more than 30w, closer to 20watts

I had some spare time today, I managed to get the Arduino AVR-ISP code working and have been able to re-program these Beer Keg TEC board.

The design is a little lacking, the ADC is only 10 bit, the current NTC divider (100k/100k) provides approx 0.1c per ADC value. Unfortunately not enough precision to make PID effective.

I have re-written the code to provide PWM Proportional control, and with a 33W TEC under a 1/4" block with 445nm @ 1250mA, heatsink and fan on the other side of the TEC the regulation and response time is pretty good, considering. PWM ranges from 0% to 100% over 10 ADC values (1c) and in steps of 10% due to the low res NTC readings.

Of course the downside to P only control is the offset error, this is why PI and PID where invented. It's not too bad though, with my 445nm test setup the temperature of the plate was with 0.5c of the setpoint (22c) even after hours of CW with the LD @1250mA.

I could alter the hardware to provide a expanded range and get a few more bits worth of resolution over the temp range 18-24c for example, it all depends if people want to use there boards "as is" or are prepared to modify them.

The high impedance of the NTC circuit means that is a little susceptible to noise as well. I can see the ADC jitter increase when I power on an LM317 current regulator for the LD. A bit of C across the NTC terminals helps.

I'll do a bit more playing around in the following week.

BTW, the original code is klunky, providing OFF below one setpoint, 50% PWM within a range and then ON fully above another setpoint. The TEC is either OFF, 50% or ON. Yuk.

[lasersbee]

Well done dnar... for the 10bit ADC that variation of 0.5C of set point
is not bad...
Like you said if the total range wanted can be reduced the 10 bit ADC
would be fine.


Jerry

[LaNeK779]

excellent work mr dnar, i think that you are really on to something here!!! 0,5c of "play" is pretty much ok for me!! i am not going to use this to cool or heat crystals in a dpss or for thermal lensing of a dpss beam

my only question is this. will the diode temp be kept within a 22-23c window if the ambient temp is 40 degrees c? and will it be kept within 22-23c if the ambient temp is 15 degrees c?

if you want to discuss the hardware moding part for better resolution, it is a simple matter of what we want to do. do we want a simple and cheap (the cheapest around) tec controller that anybody can get, reprogram and install or do we need a more sofisticated solution that people should have the necesary parts/knowledge/equipment to make it work according to "tighter" specs

i am happy the way things are at the moment, but this is a subject open to discussion

i also like the fact that a beer keg part will be inside my projector

dnar iz da man i tell you!!!!

[dnar]

Well done dnar... for the 10bit ADC that variation of 0.5C of set point
is not bad...
Like you said if the total range wanted can be reduced the 10 bit ADC
would be fine.
Jerry

I am not happy with how it looks on a DSO however. Mostly the loop sits between 2 ADC values, the PWM jumping between 2 x 10% increments. Pretty clunky. In realitity, I am only using 3 bits or resolution when the loop is locked.

This really needs 20dB gain on the NTC, an opamp *could* be glued to the board, but I am thinking "use as is" or design something better. The old strawberry jam comparison.

excellent work mr dnar, i think that you are really on to something here!!! 0,5c of "play" is pretty much ok for me!! i am not going to use this to cool or heat crystals in a dpss or for thermal lensing of a dpss beam

my only question is this. will the diode temp be kept within a 22-23c window if the ambient temp is 40 degrees c? and will it be kept within 22-23c if the ambient temp is 15 degrees c?

This all depends on:

1. The amount of heat generated by the diode
2. Size/performance of TEC
3. Thermal insulation between hot/cold plates
4. Thermal resistance of the TEC hot side heat sink and cooling

Performance at the conditions you state above are not dependent on the controller.

The new code is just:

Program size: 195 words (390 bytes), 38.1% of FLASH

So if I can obtain more ADC range, I'll add an Integral component to reduce offset errors. I have plenty of program space left.