Rich Grise said:
Well, for heaven's sakes, don't try to run it up in air, or it _will_
be junk. In fact, before you power it up at all, you should probably
bake out it under as hard a vacuum as you can get, and keep it in
vacuum. (you can do this with a valve. ;-) )
But it sure sounds to me like the 54V is from phase to phase, i.e.
delta, i.e. your multimeter will read 54VAC (if it goes to that freq)
from A to B, from B to C, and from A to C. I'm pretty sure you did
the arithmetic right, i.e. if you add a neutral in the middle, and
simulate it with a Y, then, yes, each leg would be 31VRMS to
neutral.
It sounds like you're trying to design a power supply for the thing.
Would that be an accurate guess?
I don't plan to try running the turbo pump beyond short tests until it's
under vacuum. I've carefully cleaned it and baked it at atmospheric for now.
The real bake out will have to wait until it's running.
Yes, I'm trying to build a power supply for it. I really want to try to keep
it as low cost and easy to build as possible, but I don't mind spending time
on the design and I'm patient. My original plan was to use the 3 phase
variable frequency digital sine wave generator I've designed (with help from
this group - see thread: Digital Sine Wave Generation) to drive three audio
ic power amps. Unfortunately I now realise that the required 3 phase 54Vac
means I'll need an 88Voutput swing, which is beyond the IC's I've been able
to track down. There are plenty of high power audio amp chips and power op
amps able to provide far more than the 35W max I'll demand from each, but
none at sensible prices that will work with 90V-100V supplies :-( I'm a
little concerned too about how these devices would behave when driving a 3
phase induction motor anyway.
The alternative idea I've been considering is to use some logic to produce
3 square wave signals which overlap in the required 120 degrees (3 phase)
fashion. I'd connect these to something like the HIP4086 3-Phase Bridge
Driver and 6 FETS.
The HIP4086 can't quite cope with the required 88V either, but I should be
able to find something given the amount of mains powered switch mode stuff
around these days. I'm concerned that this approximation to a sine wave
might be just too poor, especially given that it seems that the limiting
factor with turbo pumps is power dissipation in the motor windings. However,
I don't need the pump to run up to speed in just 60 seconds (like I
understand it would with the recommended supply), and think I might get away
with a fixed frequency 1250Hz 3 phase square wave supply - who knows, but I
don't want to invest too much effort in this, given that the pump may turn
out to be beyond use anyway.
Any pointers to great integrated solutions to this would be much
appreciated.
Scrim