Changing The Current Direction inside the Peltier for temperature regulation

[Ewinci]

I am trying to use a pair of MOSFETs or BJTs to reverse the current direction entering a Peltier (40mmx40mm) in order to control the temperature of one of the surfaces. question is how do I design/arrange the MOSFETs in such a way I can have current reversal without having a negative supply voltage i.e using Vcc and Ground (0V)?

 

[duke37]

You will need four devices not two.
Look up H bridge.

 

[Ewinci]

You will need four devices not two.
Look up H bridge.

I did, and it was my first option to substitute but I couldn't use the H-Bridge because
1. They support low current (1A max)
2. The ones which can handle larger currents are expensive
3. I needed fewer connection to my microprocessor

Or did you mean that I should use the theory behind H Bridge to join the four MOSFETs/BJTs??

 

[Supercap2F]

1. They support low current (1A max)
2. The ones which can handle larger currents are expensive
3. I needed fewer connection to my microprocessor

1. Actually they can support very high currents.
2. You can get a 10A transistor for USD 0.62.
3. How does two sound?
Dan

 

[Ewinci]

1. Actually they can support very high currents.
2. You can get a 10A transistor for USD 0.62.
3. How does two sound?
Dan

I wish it was that cheap here, where I am, but it is not. Either way 2 sounds good . I just have to improvise, at the moment I'm considering designing my own H-Bridge using 4 mosfets, any better ideas?

 

[(*steve*)]

Do some research on driving Peltier devices. I believe they don't take kindly to what I think you may be trying to do.

 

[Ewinci]

Do some research on driving Peltier devices. I believe they don't take kindly to what I think you may be trying to do.

i'll get right on it. Thank you

 

[duke37]

A Peltier device will have a significant thermal time constant so there is no point in switching the polarity quickly so a two pole change over relay may be adequate.

Mosfets are easy to drive but need quite a high voltage so you will want a 10V supply. Use two N channel and two P channel mosfets, also four small transistors to match your microprocessor.

 

[hevans1944]

I have a Peltier module that I was thinking of using to make a humidity sensor based on the dew point temperature. To do this requires mounting a smooth flat mirror on the cold side and then lowering the temperature until a light film of condensed moisture forms on the mirror. An IR LED reflects light off the mirror surface into a photodiode and thereby detects when the moisture forms. A feedback loop controls power to the Peltier cooler to maintain the temperature at the point where the moisture film just forms. This means a constant cycling of power because the film must be warm enough to evaporate and then cold enough to condense again for the control loop to work properly. A clean, dust-free, mirror surface is also required. Commercial versions tend to be a little pricey IMO, but the method can be quite accurate in measuring humidity and seems ideal for hobbyist implementation using an Arduino for control, data acquisition, and humidity display.

My point is this: you may not need to switch polarity back and forth from heating to cooling to accurately control the temperature of one surface. All you may need to do is control the power input, not reverse it. It is true that Peltier devices will pump heat in either direction, depending on polarity of the applied voltage, but they are significantly less efficient in moving heat to the cold side than the are in moving, and generating, heat on the hot side. In moving heat to the cold side, a Peltier device also generates heat that must be dissipated on the hot side. If all you are trying to do is maintain the cold side at a constant temperature below ambient temperature, then simply controlling power may be all you need to do.Turn the power input off, or lower it, and the cold side will begin to warm. Turn the power input on, or increase it, and the cold side will begin to get cooler.

Temperature control is never as easy as it looks. Things usually get hotter faster than they get cooler, which creates a non-linear and potentially unstable temperature control loop.

@Ewinci: Could you tell us exactly how you plan to deploy your Peltier device, the expected temperature difference between the cold and hot sides, the quantity of heat you need to move, and why you think it is necessary to operate the Peltier device as a heat pump in both directions?

 

[Ewinci]

I have a Peltier module that I was thinking of using to make a humidity sensor based on the dew point temperature. To do this requires mounting a smooth flat mirror on the cold side and then lowering the temperature until a light film of condensed moisture forms on the mirror. An IR LED reflects light off the mirror surface into a photodiode and thereby detects when the moisture forms. A feedback loop controls power to the Peltier cooler to maintain the temperature at the point where the moisture film just forms. This means a constant cycling of power because the film must be warm enough to evaporate and then cold enough to condense again for the control loop to work properly. A clean, dust-free, mirror surface is also required. Commercial versions tend to be a little pricey IMO, but the method can be quite accurate in measuring humidity and seems ideal for hobbyist implementation using an Arduino for control, data acquisition, and humidity display.

My point is this: you may not need to switch polarity back and forth from heating to cooling to accurately control the temperature of one surface. All you may need to do is control the power input, not reverse it. It is true that Peltier devices will pump heat in either direction, depending on polarity of the applied voltage, but they are significantly less efficient in moving heat to the cold side than the are in moving, and generating, heat on the hot side. In moving heat to the cold side, a Peltier device also generates heat that must be dissipated on the hot side. If all you are trying to do is maintain the cold side at a constant temperature below ambient temperature, then simply controlling power may be all you need to do.Turn the power input off, or lower it, and the cold side will begin to warm. Turn the power input on, or increase it, and the cold side will begin to get cooler.

Temperature control is never as easy as it looks. Things usually get hotter faster than they get cooler, which creates a non-linear and potentially unstable temperature control loop.

@Ewinci: Could you tell us exactly how you plan to deploy your Peltier device, the expected temperature difference between the cold and hot sides, the quantity of heat you need to move, and why you think it is necessary to operate the Peltier device as a heat pump in both directions?

Let me thank you in advance for giving the insight prior to your point because most of what you explained is almost similar to what i am trying to do. So what am I doing? I'm going to control,perform both heating and cooling, the temperature of one of the surfaces of the peltier using a microprocessor (initially i will use either arduino UNO or arduino Mega) while measuring the temperature using a thermistor for feedback, which I'm going to set it up so that it will have a precision of about 0.01 C. But to answer your question why do i think it's necessary to operate the Peltier as a heat pump in both direction? because it's important for me to either heat or cool by the turning of an optical encoder where by turning it in each of the direction would correspond to either cooling and the remaining direction to heating.
@hevans1944 how would you have done it??

 

[hevans1944]

@hevans1944 how would you have done it??

Hard to say how I would do it, since I don't know the heat load on the cooled surface of your Peltier device, nor the heat pumping capability of the device. Is the controlled temperature considerably below ambient? One side of your Peltier device should be heat-sinked to ambient temperature, perhaps also with a small fan to circulate air at ambient temperature over the heat sink. The heat sink will get cold when you reverse the current to heat the "temperature controlled" side of course. This is likely to only be a problem if you operate the temperature-controlled side above ambient temperature for extended periods of time, long enough to cause the heat sink to accumulate frost.

If that were the norm, operating above ambient, I would add a thin-film heater to the Peltier surface and cool the back side of it with the Peltier, while controlling the power input to the thin-film heater to regulate the temperature on the outward facing side. The idea is to make the thermal time constant as small as possible, so the control loop responds quickly to changes in power input, whether that be to the Peltier or the thin-fillm heater.

I can think of no reason not to use the Peltier as a heater, other than the thermal time constant will be quite long because of the mass of the Peltier elements. So go for it!

What is the range of temperatures that the temperature-controlled side must control? What thermal mass is attached? What are the external mechanisms (besides the Peltier device) for heat loss or heat gain on the temperature-controlled side? If I needed to control temperature very precisely, and 0.01 C is fairly precise, I would probably try to do what you are doing: active cooling and active heating, maybe even both at the same time if that would tighten and speed up the control loop.

Are you going to thermally insulate the temperature-controlled side of your Peltier device? That will be a big help if you are applying both active cooling and active heating of the surface. You probably want to keep the mass of the temperature-controlled surface as small as possible. As for temperature sensing, you really can't beat a four-wire platinum resistance temperature sensor for accuracy and stability. Thermistors are easy to deploy, and generally provide a high-level signal, but I have never trusted them to provide an accurate reading over long periods of time. Of course you can always "upgrade" your temperature sensor later if you don't like the results.

The thing to get out of the way for now is driving the Peltier with a high-current H-bridge, with or without PWM for control. I will let you know if I find something suitable at a reasonable price. I think something in the range of 10 to 20 A at about 60 V DC will do just fine for both Peltier and motor applications.

 

[(*steve*)]

The problem with reversal of current through a peltier is that it induces a large change in the thermal profile through the device. This can place a large strain on the elements which make up the device and speed its failure.

I would use the peltier for cooling only, and perhaps place a large power resistor near the surface for heating. The idea would be to keep the peltier at an almost constant current, varying it slowly in an attempt to keep the resistor current at some low nominal figure. Change in current (both up and down) through the resistor could be used for more rapid temperature changes.

Using low thermal mass surface mount resistors between the peltier and the surface with the measured temperature would probably provide the fastest response.