I tried doing this just this past weekend when I balanced out my 7.5
hp idler. Adding the PF caps didn't decrease my line current, it
increased it.

I tried a multitude of run cap and then PF cap combinations while
balancing my converter to run my 5hp lathe. I found my best
combination for loaded balance was 120 mf between L1 and L3 and 80 mf
between L2 and L3. L3 being the manufactured leg. This yielded 243v
between L1 and L3 and 245v between L2 and L3. Incoming line voltage
is 244v. Likewise my currents for L1 are 8.1 amps, L2 7.8amps, and L3
7.7 amps. This to me is pretty close to generating real three phase
power. Unloaded L1 to L3 and L2 to L3 voltages are up around 270
volts.

Idling the converter measures 3.7 amps on the lines feeding it. I've
added 10 and 20 mf caps across the incoming lines and the incoming
idling load increases to over 5 amps. It also increases the incoming
current while loaded. On top of that it throws the balance out of the
generated leg voltages and currents a bit. So in this instance it
doesn't appear PF caps are going to help me any.

Any thoughts as to why?

"Robert Swinney" wrote in message ...
Gary sez: "Power factor correction caps connect between L1 and L2. The
value
is whatever reduces the 1 ph line current from the service entrance
to the lowest value. Again experimentation will be required. Do this
*after* you balance the converter, if you're going to balance it. The
value won't be the same for a balanced converter as it will be for
an unbalanced converter. (Usually, it'll be smaller for a balanced
converter.)"

Right on, Gary! The 2 key points in Gary's closing paragraph are "do this
after you balance the converter" and "reduces the 1 ph line current from the
service entrance to the lowest value" Important here is that the current
from the service entrance is measured with a clamp-on type of Ammeter.
Unlike the voltage measurements for balancing this is a current measurement.

Bob Swinney