Driving a 0.8V peltier from 5V using PWM and H-Bridge

[riccardo]

Hello,

I am wanting to control a small peltier module to regulate temperature in both hot and cold modes.

I am thinking to use an H-bridge motor driver such as the MAX14870 so that I can feed a PWM signal from an Arduino, and another pin will select hot/cold using the DIR input.

However, the peltier is rated for 0.8V, 3A, but the h-bridge drivers like this seem only to go as low as about 5V.

Would simply adding an inductor in series with each of the driver's output pins, and the parallel caps with the peltier serve as a sort of buck-converter/filter? (see attached)

 

[Pommie]

Do you have a link to the peltier device. I've never seen one that runs at 0.8V.

Mike.

 

[riccardo]

35.53US $ 5% OFF|10*10mm 0.82v 3.3a 1.6w Tec1-00703 Thermoelectric Cooler Cooling Cool Peltier Plate Tec - Tool Parts - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

www.aliexpress.com

 

[Pommie]

That is not going to cool anything. 1.6W of cooling is nothing. The device takes ~2.5W to power it which will be converted to heat which will overwhelm the 1.6W of cooling effect. I've used the **broken link removed** that produce 50W of cooling and they are very inefficient.

Mike.

 

[riccardo]

It is for something very small. I am not asking about the peltier, I am asking about the driver. Please could you keep replies on topic.

 

[Pommie]

It is for something very small. I am not asking about the peltier, I am asking about the driver. Please could you keep replies on topic.

Just thought you might like to know the probable pitfalls of trying to use this device. I'll just keep quiet from now on.

Mike.

 

[dr pepper]

I think that would probably work, for protection from kickback you could put a transient suppressor diode across the inductor.
Another way you could do this is use a smps Ic like a uc3843, and construct a current source, then you might have better control of heating or cooling, but then polarity reversal would be an issue.

 

[crutschow]

I suggest a PWM driver circuit with a series inductor (and free-wheeling diode) for best efficiency, with a MOSFET H-bridge at the output to control the polarity to the Peltier.

 

[riccardo]

I suggest a PWM driver circuit with a series inductor (and free-wheeling diode) for best efficiency, with a MOSFET H-bridge at the output to control the polarity to the Peltier.

I was considering this way, but I can't find a suitable h-bridge that will work at such low voltages. I'd rather not use separate MOSFETs as I need to make this as small as possible. (There will be four independent drivers on one board.)

 

[dr pepper]

If your working below 5v you'll need low threshold mosfets.

 

[alec_t]

I can't find a suitable h-bridge that will work at such low voltages.

Although Vgs needs to be >> 0.8V I don't see why the drain voltage supply can't be 0.8V; so various H-bridges should be suitable.

 

[rjenkinsgb]

The bridge supply does not need to be particularly low voltage. 5V or even 12V will work.

With a series inductor (with the load), plus the diodes in or across the bridge FETs, it will act as a buck converter and the PWM duty cycle will control the load current.

You just need to limit the duty cycle proportionally to the supply voltage.

 

[Diver300]

I don't think that there's much point in having two inductors. The overall inductance and the saturation currents are the important considerations. Similarly, there's not a lot point to having the capacitors in parallel with the Peltier cell, as the current is kept reasonably constant by the inductance. You need lots of capacitance on the supply to the H-bridge, when using PWM, as the input current is discontinuous.

The important consideration is to make sure that the H-bridge is driven in a way that shorts the output during the off part of the PWM cycle. On the MAX14870, that is done by using the PWM input, and not a direction input.

The MAX14870 has a maximum current of 2.5 A so it won't drive 3 A into the Peltier cell. You need a larger H bridge, but you might be able to parallel two bridges.

 

[rjenkinsgb]

A VNH7070 or VNH7100 would do - operation down to 4V and rated 15A.

VNH7070 STMicroelectronics | Mouser Europe

STMicroelectronics VNH7070 are available at Mouser Electronics. Mouser offers inventory, pricing, & datasheets for STMicroelectronics VNH7070.

eu.mouser.com

One inductor plus a good low ESR cap on the DC bus side, as Diver300 says.

 

[ronsimpson]

With a series inductor ….. it will act as a buck converter and the PWM duty cycle will control the load current.

4.7uH=4.7uH=9.4uH, 4 volts across the coil, = 425mA/1us, or in about 7uS the current will go from 0 to 3A (max IC current). Then another about 7uS for the current to get back down. You will need to run the PWM pretty fast to keep the current under control. Maybe larger inductors.

There will be voltage loss in the IC so maybe it takes 10uS to hit 3A. Take care you don't go off and leave 5V on the 0.3 ohms load.

 

[alec_t]

Here's the basis for a simple PWM switcher, running at ~33kHz, to give a controlled ~3A of selectable polarity. The polarity switch is speeded up in the sim just for the demo.

M1-M4 represent an IC H-bridge. The downside of simplicity is that if the current-sense feedback fails the Peltier gets the full 5V.

 

[rjenkinsgb]

4.7uH=4.7uH=9.4uH

I was basing it on calculating an appropriate one for whatever frequency was appropriate for the bridge, not the 4.7uH ones.

 

[riccardo]

Thanks for your suggestions. I found a nice little module that can do it.. PL-TEC 1-0305. However, they are asking for almost £200 each for them. Fortunately theres a photo which shows the part number for the IC they use which is a MAX1978 at only £20. It looks to me like their board is basicaly a breakout for that IC with the components reccommended in tha datasheet.

I want to be able to set the temerature using a microcontroller, so will need to add some extra circuitry.

I think with this IC, I can use a PWM signal from my microcontroller, through a low pass filter to make a voltage for setting the temperature. I'm not quite sure how to integrate this with the built in thermistor feedback. Hopefully another look after getting some sleep will provide some inspiration!