AVR Push Pull converter

After playing with the Buck converter, I wanted to try my hand on a full push pull converter. These experiments use a common mode choke, this idea was gotten from this article by Electronic Design.

Setup

Below is the general setup of a Push pull converter (image taken from wikipedia article).

These are the parts I used:

Switches: IRF820 mosfets (I got loads of these from some electronics lot)
Gate drivers used are the IR2301 low and high side drivers.
Diodes: 1N4148 diodes
Transformer: Common mode choke from broken ATX powersupply
Input and output capacitors are 330uF rubycon low ESR capacitors
Vi: variable
Vo: variable with feedback resistor, using a 5V reference.

Microcontroller used is the ATtiny2313 running at 20MHz.

tn2313_pushpull.c

isrs.S

TODO: code description

Breadboard prototype:

Closeup of the microcontroller section:

Power output:

Oscilloscope traces

Single cycle ripple:

There is quite some spikey ripple when new energy is coupled in the transformer. Most of this can be attributed that the whole is prototyped on a breadboard.

Multiple cycle ripple:

Multiple cycles, transformer side.

multicycle transformer side, not rectified

Multiple cycles, half bridge side.

Conclusion

The push pull converter works pretty well with a generic common mode choke. The voltages and efficiency is probably lacking, but not really a reason for this excersise. Timer1 of the AVR microcontroller is pretty flexible and enables to create complementary waveforms with modifyable deadtime.
Prototyping powersupplies on a breadboard is suboptimal, lots of noise and spurious pulses get coupled in the oscilloscope input.

What next?

Now lets apply what I have learned here and modified one very common half bridge converter. The ATX powersupply. The goal is to make a variable voltage and variable current laboratory powersupply.

Back

Last update at: 15-02-2011