Re: full h bridge theory

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Anthony,

Others may have different ideas on this, but here are mine ;-)

The H-bridge is basically four IGBTs or FETs.  The voltage each IGBT is 
rated at should be considerably higher than the voltage across the entire 
bridge.  If the voltage across the IGBT ever goes higher than its rating, 
it will explode...

For the current,  usually the "peak" current rating is used for pulsed 
applications like Tesla coils.  In the case of IGBTs, they seem to be able 
to take 2 to 3X that rating in low duty cycle resonant applications like 
ours, so we push them pretty hard.  The gate voltage should be about 25V to 
run super high currents though.  The IRG4PF50WD can take 750 amp hits!!

We tend to use spice modeling like MicroSim to find the RMS current through 
them.  That current multiplied by the voltage drop across them gives the 
power dissipated.  It is a little messy since there is a reverse diode and 
such but that data sheets give the voltage drops "somewhere".  If the 
current is say 10 amps RMS and the voltage is 5 volts we get 50 watts but 
each IGBT only works 1/2 time so that is 25 watts...  The spice model might 
have already considered that.

Every time the IGBT switches, it burns up heat too.  These are usually 
listed as "switching losses" vs. collector current.  For the IRG4PF50WD at 
say 50 amps it heats about 14mJ (fairly linear curve) for each 
switching.  If it switches 20 times each cycle at 200 BPS we get 0.014 x 20 
x 200 = 56 watts.  If we switch at 10 amps instead that goes down to about 
10 watts.  Thus we try and switch near zero collector current.  Once one 
knows the general amount of heat they put out, you bolt them onto a big 
heat sink and blow air on them like crazy ;-))  Reports are that they don't 
get too hot...  Most people are using really big heat sinks.  I am using a 
small but very efficient one with temperature monitoring.

It can all get pretty messing and there is no good formula for the whole 
thing.  It is often a very rough guess until it is actually built and 
tested.  When in doubt, go with big heat sinks and fans.  You really can't 
run them too "cold" ;-))

Hope this helps...

Cheers,

        Terry

At 12:44 PM 1/18/2005, you wrote:
> hi guys i was hoping you would be able to explain something to me about 
> full h bridge how do you work out how much power they can take and how the 
> power and current is distrobuted if you could answer these questions i 
> would be great full thanks