Peltier current control

Other Hardware Motors, Mechanics, Power and CNC
41

Thanks dc42, ajofscott.

Start again :~

  1. LED chain not behaving as it should, even when the Peltier is disconnected. As you haven't said what controls the power transistor that switches the LED, it's hard to say why, although if it only occurs with the switching PSU connected then ground loops may have something to do with it.

Power transistor HIGH, LOW digital pin, sequenced with the camera shutter, also HIGH, LOW digital pin through an optocoupler (no issues with shutter circuit as ground is through optocoupler). I notice that the LED array is operating out of sequence with the camera shutter and think an optocoupler on the arduino pin to LED switching transistor Base might fix these problems.

  1. Peltier PWM affecting the temperature reading.

Reconfigured circuit per dc42 and ajofscott instructions.

Not able to post images or Fritzing files at the moment - complains about being full.

42

This thread is getting a bit long and tedious. I did start a separate thread relating to TEC PWM using a KA3525A but since deleted it. Still can't post attachments - linked to 7.5kb png image.

The plan from here is to implement the suggestions so far and make a PCB to fit the side of the heat sink fan. That's about the only place it will fit. I will invert the FET and have it sitting in the heat sink fan airflow. The new batch are higher rated and have much lower rds on, so they are not getting as hot as the others.

The circuit is adapted from a pwm motor control circuit I found on line - one of the few. I think that there should be some sort of frequency adjustment from PWM in -> DC ripple out to the FET gate. I can't imagine that the KA3525A will automatically provide the correct output. This will mean a variable resistor on pin 6.

The other thing I intend doing is routing the temperature control and LED array circuits through some 8 pin optocouplers that I have. At the very least, have the GND's returning via the optocouplers.

If this approach is flawed please let me know. At the moment it seems to be the way ahead.

43

geoland:
At the very least, have the GND's returning via the optocouplers.

Each signal drives a photo emitter and is recieved on the isolated side as light detected then by phototransistor. Linearity is going to be a trick on analog signals.

44

While this thread has been somewhat mixed, starting with sensors and ending with electronic issues, this is what I have learned from this exercise as far as sensors and other I/O functions are concerned, when mixed with switched mode power supplies.

It seems that there are SMPS's and there are SMPS's.

Besides having a linear wall adapter, it occurred to me that I have been using another 12v 1A adapter for years with the Arduino on another device without problems. It turns out to be an SMPS, which completely escaped my mind. So I guess, the filtering/noise suppression on the SMPS wall adapter is designed to be suitable for domestic use.

The 12V 12.5A big SMPS that is driving the TEC, obviously, is not intended for general use, is probably low quality and from what I can see its noise output is difficult to control. An oscilloscope will be needed to measure its output frequency, typically somewhere between 100khz and 1Mhz for SMPS's, and a I think, though I am quite probably wrong, a Butterworth filter designed specifically for it.

dc42 has provided a circuit diagram for the same purpose and I haven't had time to try it out as yet - but from recent reading it looks like a low pass filter.

I may strip down the SMPS wall adapter and co-locate it with the big SMPS, tapping its 240v supply. That way I can dispense with a separate adapter and retain just two cables - no power boards and general untidy c**p lying around to get tangled in feet and hands. The lead to the Arduino from the big SMPS is the shielded type, so I'll use that.

The other thing that has escaped my attention altogether, is that the power supply on the motor shield that powers the stepper motor is also SMPS - a mini one - to cut out the heat associated with linear power supplies driving heavier loads - it has not caused any problems.

So, I guess, somewhere in the back of my mind SMPS's are OK and didn't expect problems. Introducing the big SMPS, the effect on the Arduino is really disruptive and totally screws up the timing of the digital I/O, while the analog is just unreliable.

Thanks to everyone. I realise that you guys stretch yourselves around the forums, so it is much appreciated when you take the time to answer.

45

Having implemented the changes suggested by dc42 in his circuit diagram further back, except for the choke, everything is working as well as can be. I opted to convert the old SMPS wall adapter to a dedicated supply attached to the top of the open frame SMPS, tapping off the mains power. Finally, I need to address the MOSFET set up due to over heating. I'll ask across at the electronics forum.

Temperature readings are very stable which has stabilized PWM, a function of ambient temperature and set point temperature. Variation is half a degree +/- This is pretty good and there should be enough thermal inertia in the system for these variations to be indications only.