Driving a differential synchro

[dr pepper]

I recently got a azimuth rotator instrument I want to drive, it has a differential synchro in it, 24vdc, both the stator and the rotor have 3 phase windings, usually this device indicates the diference between 2 3 phase signals.

I'm thinking if I feed the device 3 phases with 120degree shift between each which works with 3 wire dc 'syns, then wiring the rotor to the stator with 2 wires crossed to generate a 120 degree shift.

Will this work?

 

[Warpspeed]

Just wiring the rotor to the stator would cause the synchro to turn to one fixed position and just sit there.

It really needs two independent three phase drive signals, and its the phase difference between these two independent drive signals that determines the position of the pointer.

 

[dr pepper]

Thanks for that.
Yes after I'd posted I thought its just going to go to an arbitrary 120 degrees position, and when I tried it that what it did.

I did get the synchro to turn, but you wont like what I did, I put dc across 2 of the rotor connections to emulate a permanent magnet as in a dc synchro.

It now rotates however current is high in the rotor and I'm probably going to get some remanence effects but if I magnetize the core it'll still do the job. I have a spare driver in the H bridge so I'll use that to limit the current through the rotor.

 

[Warpspeed]

What is it you are trying to indicate, what is the signal source for this synchro going to be ?

 

[dr pepper]

This particular synchro is part of an aircraft instrument, its a azimuth rotator for a radar dish, I modify old instruments into other things.

https://www.youtube.com/watch?v=Pt933ZWahNc

The trash on the breadboard isnt part of the circuit, just the L293 on the right and the 'duino are driving the device.

 

[Warpspeed]

This sounds like the beginnings of an interesting project.

 

[BradtheRad]

Since you're familiar with this type of rotator then you may be familiar with ordinary tv antenna rotators like the one we had in our house. It had a dial which I turned to the compass direction of the broadcast. An internal motor turned the dial, while outdoors a motor turned the antenna. Everything stopped when the internal motor lined up in the same direction as the dial which I initially turned.

I believe the dial mechanism controlled the outdoor antenna. Anyway you might get some tips by looking at circuits which turn the dial.

 

[dr pepper]

That ones fairly easy, this is a similar system, a differential synchro can be used in a positioning system like this.

What I'm doing is use the device not as intended, but as a display device where a register in the arduino becomes a position of the shaft.

Biasing the rotor with dc works, there is some notchiness when the synchro tunrs, much like a stepper motor microstepping without current control, however this isnt a problem as the device feeds a gearbox.

I dont quite understand how I'd properly drive this device, would I need a 6 phase source?, that wouldnt be hard, I'm using a sine wave lookup table and 2 offsets, 3 extra offsets and 3 extra pwm o/p's is easy enough.
Or would I need a steady 3 phases and the other 3 offset by a varying number of degrees to produce a shaft position.
I'm thinking the latter.

 

[FvM]

Two 3-phase voltage systems and one phase offset between both.


u11 = sin(wt), u12 = sin(wt+120°), u13 = sin(wt+240°),
u21 = sin(wt+ofs), u22 = sin(wt+120°+ofs), u23 = sin(wt+240°+ofs)

 

[dr pepper]

The guru speaks.
That is just what I have programmed.
I left the code as is, just added 3 more phases UVW, each one adds a phase shift offset to its counterpart RST, I then need to alter the bit of code that keeps track of position so it looks at UVW instead of RST.
Off to work now but will try it when the dragons have gone home.

 

[dr pepper]

OK so I tried it.
Kinda works, the rotor turns 1/2 a trun then switches back to the start and repeats.
My code doesnt wrap around the 360 entry sinewave lookup table, so its obviously falling off the end after 1/2 a turn.
All a bit of a waste of time though, upon closer inspection this device is for input only, the gearing to it from the display is geared down massively, the synchro does a full turn only when the display has gone from 0 to 360 full range, if I were to try and use the synchro to drive the display it wouldnt have enough torque to overcome the gearing.
So to use this indicator I'd have to re-hash the gearing and use a geared motor or something.
Heres some pics for those that are interested, the big holes in the chassis is where the synchro mounts.

 

[Warpspeed]

Something you may like to consider is driving your synchro at a lower voltage and a lower frequency.

These synchros were originally designed for airborne applications which typically have 28v 400 Hz three phase power distribution systems.
But you can reduce both voltage and frequency (roughly in proportion) to achieve similar currents in the windings.

Its exactly the same principle as with variable frequency dives and induction motors.
A 230 volt 50 Hz motor can be run with 23v and 5 Hz and deliver full torque, with full rated motor winding current at one tenth the speed.

There may need to be a correction to allow for the resistance of the windings, but that should not be a problem.
It may be more convenient to run these synchros off 5v or 12v or 15v, especially if its only driving a pointer.

 

[dr pepper]

I'm not using fancy gear, er my fingers, if I reduce the freq I get to a point where torque doesnt increase and current does, I assume the core is saturating at this point due to the V/S rule.

I modded the code so it does wrap around the lookup table and the synchro now rotates well all the way around, seems to work ok at 7v where its pulling 300ma (actually with a H bridge that would be 14v), at its rated 12v it pulls 1a and gets warm, but then its sposed to be a sine wave not a pwm representation of a sine wave, my code is very similar to a vfd, I maintain a few of those here at work.

I cant use this device as is, however its been a good exercise learning how to drive it, maybe I'll put it in another project, its a little i/o hungry needing 6 pwm i/o's.

 

[Warpspeed]

I'm not using fancy gear, er my fingers, if I reduce the freq I get to a point where torque doesnt increase and current does, I assume the core is saturating at this point due to the V/S rule.

It won't saturate unless the current is higher than its supposed to be.

 

[dr pepper]

Yes agreed, I've designed ferrites for smps before, silicon steel cores isnt something I've played with though, if I was that bothered I could 'scope across a current sensing resistor and have a look.

Well it works fine now.

Interesting that a diffrerential synchro is sensitive to phase rather than amplitude and phase like a ac synchro.

 

[Warpspeed]

Never played with one myself, they are becoming rather rare items these days.
Its all done with incremental shaft encoders an stepper motors now.

I would imagine that amplitude only effects the torque produced. If the mechanical loading is very low or zero, the positioning should still be pretty good even with minimal drive power.
The excitation frequency may not be that critical either, and that might be a convenient way to set the operating current.

All pretty interesting stuff.

 

[dr pepper]

You might think so, but when you look at the 3 phases, the average voltage across them is the same, one might be near the supply and the others low, or they could be all part way, either way average is fairly equal, so the torque is constant, thinking about it I psoe it has to be or it wouldnt be stable.

Yes drive freq isnt critical, though it does produce more torque near its rated freq.

I have some more isntruments with these in, a couple with a whole bunch of them.

 

[dr pepper]

Synchro's and resolvers are not dead yet, the latest toyota engine with variable valve timing uses a resolver.