Hi,
Just a couple of notes on the 'inverter' previously linked to in this thread...
First, the LM324 isnt a great driver for transistors. If anything, the output of
the CMOS would be better as the rise amd fall times will be better.
Second, and probably even more important, there is no 'dead time' built into
the inverter. The dead time is the time delay between the switching of opposite
transistor groups and works out as a time delay to turn on and no time delay
to turn off the transistors.
What happens is the transistors on one side turn on and when the drive signal
tries to turn them off they dont actually turn off for some 10us or more. During
this time if the opposite bank turns on it will cause a very low impedance short
from V+ to ground through the output transformer even though the collectors
go to opposite sides of the transformer. This will cause a huge surge which is
considered very destructive and usually blows the transistors.
By delaying the turn on only and not the turn off, the transistors that are
turning off get a chance to turn off fully before the opposite bank is turned on,
thus eliminating the near direct short.
The actual delay time depends on how fast your particular driver can turn off
the transistors and should be slightly longer than that. For fairly fast transistors
5us is enough, but 10us is better and due to the poor drive method used for
this inverter maybe 20us would be even better.
Just some things to think about before turning this 'inverter' on.