"Jim Thompson" wrote in message
...

My specialty is circuit design, primarily chips, so I can't help on
transformers. I was specifically wondering what you were trying to
accomplish with the SCR... particularly wondering if there wasn't
possibly a better way to hold current thru the zero crossing.

That may best be reserved for another thread. But we have discovered that,
for highly inductive loads, applying a gate signal on the positive excursion
of the voltage waveform means that it will be removed at the point of peak
current, and relies on the SCR's current latching characteristic to maintain
conduction through the zero crossing. But there is a point where the current
drops below this level, and I surmise that the current still flowing will
create an inductive voltage "kick" which causes a distorted waveform at the
zero crossing. At this point, the opposite device's gate drive will have
already been ON for a quarter cycle, so the distortion appears only at the
end of the current waveform.

Sometimes this seems to cause erratic behavior and latch-up, where the SCR
may be turned back on, and sometimes this happens only on one SCR, which
causes a net DC component and correspondingly very high input current which
trips the mains protection instantly. On a 200 amp source this means the
fault current is at least 2000 amps, and probably much higher.

I don't know exactly why the continuous gate drive works to solve this
problem, and it may not do so completely, but it has been shown to be
effective and I don't know of any downside. I don't think maintaining 300 mA
on the gates does any damage. We've had SCRs in the field using this method
for at least ten years, and most failures seem to be due to an open gate
wire or customer abuse. We have dealt with the former by using the new SCR
trigger board which checks for gate continuity before triggering, and we
have reduced the latter by using a "programmable overload device", or POD,
which limits the duty cycle of operation.

Last I heard from you it seemed you were interested in this phenomenon. But
I didn't realize you had no experience with transformer construction and
high isolation DC-DC converter design. Now that I review your areas of
expertise I see that the closest relative is switching power supply design,
and they are typically only rated for common mains voltages up to 265 VAC.
It would be really helpful to find one that works on 480 VAC mains, but I
have searched in vain. However I recall seeing something recently that might
be suitable for this purpose.

Thanks,

Paul