I thought I also make a post here specific for the problem I have at the moment next to my general project description.
ATM I need to figure out how to best control this Peltier. http://uk.rs-online.com/web/p/peltier-modules/4901395/
I am planning on using this powersource, but I am flexible on that.
I think I need an H Bridge as I want to use the peltier as heater and cooler.
However I am not sure which one exactly as there are lots of chips and boards and motor drivers out there.
Just remember that you will need a way to remove the heat from the hot side when your using the cooler because it will quickly stop being cold as the heat cancels out the affect if it has no wear to go!
Are you sure it is a 15V peltier element ?
Are you sure it can be reversed by reverse polarity ?
You should test it first. Perhaps you can try it with 5V so it won't get too hot.
As Drew Davis wrote, the hot side needs a heatsink.
Not if you need to be able to make one side both hot and cold.
I have used many pelter elements and all have been reversible by switching the polarity. Thats not to say some can't be reversed, I just haven't had that experience. I do agree that it should be first tested. If you have a variable power supply you could start at a low voltage like Erdin suggested and slowly bring up the voltage while monitoring how much current your using.
I am not worried about cooling the hot side too much. I will have a rather large copper block on the "hot side".
I am only interested in a postive temperature gradient (heating of 1 degree per second or so ). So it is running in a heating mode most of the time, where it is activly pumping heat out of the copper block (into the sampleplate) and only if the heater wood overshoot I would need to cool a bit, so then I will always have a rather large cold reservoir (the copper block). the actual cooling down to 77 K will be done by liquid nitrogen
I was wondering if anybody has experience with using an old computer psu for testing purposes?
Or knows where I can find info on that. that would save some money in the initial tests. I would want something like 12 V, 6A.
Heating without overshoot ? What about a simple heater ?
Perhaps a PID to prevent overshoot.
A PSU from a PC can be used.
Some old ones requires a small load, so you have to check the output voltages without load.
There is a sense wire, and you have to connect it to something. http://reprap.org/wiki/PCPowerSupply
I have such a prepared PSU and it's very handy to have.
I have read some of the peltier threads but surely not all.
thank you for the input on the pwm atenuation. I was planning on putting in a simple low pass filter, with a resistor in series and a capacitor in parallel to the peltier. However I have not yet looked into the specifics. do you have any experience on that part of the story and can make suggestions?
actually on second thought I don't think that PWM needs to be filtered as long as the power suply doesnt exxed the maxium ratings of the peltier and the PWM cycle is fast...
see point 41 here: http://www.tellurex.com/technology/peltier-faq.php
Sedeca:
I was planning on putting in a simple low pass filter, with a resistor in series and a capacitor in parallel to the peltier.
That's no good, you will waste a huge amount of power in the resistor. You need a series inductor (capable of handling the maximum Peltier current without saturating or overheating) and a Schottky flyback diode. You will need to increase the PWM frequency in order that you can use a low enough inductance (otherwise the inductor would be huge), and because of that you will probably need to use a mosfet driver chip (e.g. MCP1407) to make the mosfet switch fast enough.
