Digital Current/Voltage Control

Hi Kas, I just took a Peltier and slowly increased the voltage across the junction. As I raised the voltage the current followed in a fairly linear fashion. Of course as the voltage and current got higher the temperature gradient from one side to the other increased. The answer to your first question is that the voltage will dictate the current. Depending on the precision that you need you can choose the appropriate temp sensors and link them to the microcontroller and then use the PWM output of the microcontroller to drive a FET or similar device that in turn supplies power to the TEC. My Peltier which is only about 2 square inches takes 2 amps at 6 volts so you will want to size and heat sink the driver FET to handle the amount of power you will be switching. You will also have to heat sink the backside of the Peltier if you want to make the most efficient use of it.

John

Hello John,

If I may ask why do you suggest using a PWM to drive the FET and inturn the TEC as opposed to using some form of regulating transistor that you could increase the gate voltage to increase the power to the TEC, I'm guessing that would be current control which is not what you really want ? Asking more to learn and better understand your thinking here.

Thanks

Kas

Hi Kas!

I think John is saying that although you could measure current and try to adjust the voltage so that the current through the peltier can be adjusted, it appears to be linearly related to the voltage anyway. So by varying the voltage, you are varying the current linearly.

PWM can be considered like an average voltage across the device, so it is a nice way achieving control if it suits your scenario. The reason PWM is often preferred (generally - I have very little knowledge of peltier devices : ( is because it is efficient (little power wasted - if linear control was used, then there would be power loss in the transistor or device used to cause some voltage drop, and that loss could be significant since peltier devices usually pass a fair amount of current - e.g. John's peltier device can take 2A at 6V, which is a lot of power 2x6=12W and then the heat in the transistor would also then need to be taken into consideration). Some control algorithm would still be needed, just that the output of it would control the PWM ratio, not an analog control.

Hi Kas, Shabaz has done a good job of covering the bases. Also PWM is built into the Arduino and I suspect the other boards as well so it is an easy choice. Using it to control your FET will put the power accross the peltier and minimize the power loss accross the FET. The PWM makes the FET act like a switch that is being turned on and off to provide an average amount of power to the load. You vary the power by leaving the switch on for longer or shorter periods of time. The other option is as you mentioned to use the FET or Transistor as a variable resistor. The problem with this is you will then have a much greater IR drop accross the FET which will produce heat the will have to be sunk and disposed of. If you use a logic level FET you could drive it directly off the Microcontroller which would also simplify the circuitry. There will be several variables in how you approach the design depending on how closely you want to regulate the temperature of the metal. Sounds like a fun project.

John

@shabaz @John_Wiltrout How do I mark your answers helpful ?

Hi Kas,

I'm not sure (it may be related to marking the discussion as a question when first posted, which I'm not sure how to see anyway)

but thanks for the comment anyway : )

Hi Kas,

Your original post was listed as a discussion and as such it is not considered an answerable question. Element 14 has a box that you have to check when you first post if you want it to be a question.  I like to respond and I appreciate your intention to mark our answers as helpful, that is enough. Let us know how your project turns out as it is always fun to hear about the finished project.

John

Hello John,

I will try keep you posted, currently I'm waiting for some TECs to come in to do some basic testing with them and a uC. I'm hoping the FET will have no switching issues but if it does I guess I'll just have to switch it out for a different one. Once that is done I'll try add in an aluminum block (cooling unit) with a copper pipe through it. The final idea is to see if water can be cooled quickly and efficiently. Overall I am looking to beat the price point of both the system and power consumption of an always on refrigeration system for drinking water or possibly a general cooling system.

Kas

Hi Kas, If you can find a way to put multiple small pipes through the aluminim instead of one large you will greatly increase the surface area to water volume flow and increase the heat transfer rate from the water.

John

John,

Thanks for the idea, I guess that would depend on how tightly I can bend the pipe so I can maintain a very small form factor.

Kas