In article ,
karel wrote:

"Robert Bonomi" wrote in message
...
In article . com,
wrote:
I've got an ARU 44 gyro and I'd like to power it off my 24VDC system.
It needs 115VAC 3 phase, and about 75W to 100W. Is there an off the
shelf unit that can do this, or maybe someone knows of a schematic to
build my own? I'm an electronics engineer so I can do this, but
inverters are not my specialty so any guidance would be greatly
appreciated.


WAG says it's probably 400 cycle, as well. not 60Hz.

Design can be fairly simple, particularly if you use a micro-controller
for the 'logic' of phase generation.

start with a clock that is 2 X phases X frequency,
Feed it into a counter, that counts from 0 to (phases-1).
Hang a number of comparators on the counter outputs, one that
triggers for each value.
Send each comparator output through a flip-flop.
(this gives you a 50% duty-cycle output for each phase,
and at the right frequency.)

[now for the "analog stuff

low-pass filter to sine-wave.

Amplify to get the needed voltage/current levels.

Voila!

And how does this manage to produce 110 volts from 24?

Note the last line: "amplify to get the needed voltage/current levels".

Original poster claimed to be an E.E., so I didn't feel it necessary
to go into all the gory details -- amplifiers that produce larger output
voltage swings than their supply voltage _are_ "well known" devices.

Some kind of transformer wil be required.

Indeed. the 'output transformer' of the powered-by-low-voltage amplifier.

Step-up rectifying might be an option
but will be hard at these power levels,
i.e. require huge condensers.

But I'm afraid I am no more able than you
to really fill O/P's request for a schematic

PS does anyone know how important the true
sine wave is for this kind of equipment?

Depends greatly on the device (I'm not familiar with the specific one the OP
names), and whether there is anything more to it than just the drive motor.
*MOST* motors could care less about the quality of the wave-form. 'smarts'
for attitude read-out are a whole 'nuther story.

There are various ways to build the voltage/current 'amplifier' stage:
1) use a switching-type power supply to 'up-vert' to a DC level adequate
to allow the use of an output-transformer-less power amp circuit.
2) take the square wave output from the phase generator directly into
a 'class C' amplifier, and feed that output into a step-up xformer.
(very good efficiency on the amplifier, lousy efficiency through
the transformer. corollary of lousy transformer efficiency -- need
a bigger transformer, and noticeably more power input to it.)
3) filter out (some/most/all of) the high-frequency components of the
waveform _before_ amplifying. then current-amplify, and transform.
(lower amplifier efficiency, but also lower losses in the xformer.)

trade-offs *galore* between the approaches.