Hello Fellows,

I am in designing phase of a project with app. 800W system. The load is not inductive type (not motors) but it is of Thermoelectric Device. I am planning to build a PWM H-Bridge type application that I can change the direction of voltage and also vary the voltage to the thermoelectric device.

Input Voltage : upto 50V
Current: ~20A

Questions:
1. N-FET vs P-FET - I was thinking of using four N-channel FET instead of two P-types and two N-types FET's. I am still selecting a reliable FET driver (high and low side). I am selecting N-FET purely because of less heat wasted on N-FET vs P-FET.
2. PWM - This will be my first time designing PWM since college days. I was thinking of just pulsing the low side FET and keeping the high side ON. Any benefits on pulsing both (high side and low side) versus pulsing just the low side? I am thinking of operating at around 10 KHz to keep the EMI noise to minimal.
3. Safeties - Came across some fully protected H-Bridge but not rated for my application. Are there any rated for that high current? Mainly interested in OC, OV, UV, etc.

Looking for some feedbacks.

Happy Near Year

 

If you don't need the instant polarity change, a DPDT relay and a single FET will do what you want.

Except for this part -

vary the voltage to the thermoelectric device.

That is one thing PWM does not do.

There must be hundreds of FET's that can handle 50 volts and 20 amps. At 10 KHz, there are also many IGBT's to consider. They drive just like FET's.

 

High current Variable Frequency Drives use high current 3 phase bridges, a vfd for 480v will have a 1,000v bridge. Allen Bradley 1330 has such a bridge and I believe the ones I worked on were rated at 80 amps.

 

Have a look at ST Microelectronics' VNH2SP30-E - Automotive Fully Integrated H-Bridge Motor Driver.

It's rated for up to 30A at up to 41V. Not quite your stated voltage rating, but it's close, and worth considering.

Using N-ch MOSFETs for both high and low side does reduce losses in the high-side MOSFETs. However, there is always some leakage in the caps on the high-side drivers. If you're going for a discrete or even a packaged driver solution, plan on using PWM on the high side to keep the caps charged, and keep in mind that 100% ON won't be attainable, but 99%+ will be.

As Beenthere implied, you won't be able to regulate the voltage using an H-bridge. However, if you use a large enough inductor, you can keep the current through your TEC device (I'm assuming a Peltier device) reasonably constant.

 

If you don't need the instant polarity change, a DPDT relay and a single FET will do what you want.

Except for this part - That is one thing PWM does not do.

There must be hundreds of FET's that can handle 50 volts and 20 amps. At 10 KHz, there are also many IGBT's to consider. They drive just like FET's.

"Except for this part - That is one thing PWM does not do." I guess you mean thats the one thing PWM can do. Polarity is not changed instantaneously in my application. I would say a delay of ~ 1 sec.