On May 5, 2021, at 12:22 PM, Jon Elson <elson at
pico-systems.com> wrote:
On 05/05/2021 10:37 AM, Paul Koning via cctalk wrote:
Incidentally, a way to get three phase power at a
frequency of your choice is to use a "variable frequency drive". That's
basically a high power solid state inverter intended to drive three-phase motors with a
chosen frequency resulting in the RPM of your choice. Depending on the model, those can go
up to 120 Hz or so, or all the way to somewhere around 400 Hz. I have a very cheap one at
home that runs on single phase 220 volt power, producing up to 3 kW or so at anywhere up
to 120 Hz. (Made by a company named Teco, amusingly.) Most of these and especially the
larger ones want three phase mains input, though I'm told that even for those you can
typically just connect them to single phase power (between two of the three inputs,
leaving the third unconnected) at reduced power ratings. These devices are surprisingly
cheap, in particular they tend to be cheaper than "rotary phase converters"
which is how machine shops traditionally produce three phase power when their mains is
just single phase. paul
Note that VFDs are designed to run motors exclusively.
They approximate a sine wave with pulse width modulated 400 V pulses. DON'T EVER try
to run electronic gear with a VFD, at least without a massive smoothing filter to convert
to a true sine wave. With a motor, the winding inductance solves the problem, and the
coils see nice sine-wave currents.
Jon
I've heard that. But why? It's not like the electronics we're talking about
actually runs on AC. Instead, it goes right into a transformer (an inductive load not
much different from a motor) and after that into a ripple filter. That filter IS the
"massive smoothing filter" you're talking about.
It's certainly worth a test especially before connecting something rare and ancient,
but I see no theoretical reason for it to cause trouble.
The voltage issue is a different one. I've never seen a VFD that offered anything
other than frequency change -- indeed, it produces the same RMS output voltage as what you
feed it, and it isn't insulated. Again, for the application we're talking about
that is likely to be fine. For example, the CDC 6600 wants 208 volt three-phase (in the
US model) both at 60 Hz for the motors and at 400 Hz for the main power.
Chuck mentioned an "idler motor", that's what is sold commercially as a
"rotary converter". Those are nice and can be had for quite high power, but
often a VFD does the job more cheaply (and with the added benefit of speed control) for
much less. I think my Teco converter cost about $100, way less than a commercial rotary
converter rated at 3 hp as that one was, and for a lot less effort than home-building a
rotary converter would be.
paul