ARC AVI200 RMI re. Stepper Motors

[BushPilotWannabe]

Hi

Has anybody done anything with this instrument?

HAROWE SERVO CONTROL
STEPPER MOTOR
p/n 8MSV-1100-A-2B
4 phase 9.5V

Interfacing the stepper motors for heading and ADF indication seemed easy when the cover was removed. Five wires coming from the two stepper motors (unipolar?) with equal resistance from the white wire to each of the other four wires. However when trying to determine rotation of the four coils, applying voltage between the common lead and any other lead produced a noticeable knocking sound - same when reversing voltage. Running the steppers using half step sequence, the same noise was heard on every second step.

Is this instrument normally this noisy or considering the presence of a 400Hz power source does direct activation require ramped voltage application or split voltage across opposite poles?

Hugh

[BushPilotWannabe]

Some progress. The steppers are running much quieter in full step mode at lower voltage. A headset should mute current clatter . The voltage required is rate of rotation dependent. High rotation lowers the voltage required, 200mV too low stalls the encoder, 200mV too high and the encoder or the gears protest. Will have to work on pulse timing and try activating opposite coils with split voltage. One big problem will be 2400 steps per indicator rotation - or 6.6666... steps per degree.

Hugh

[_alioth_]

One big problem will be 2400 steps per indicator rotation - or 6.6666... steps per degree.
Hugh

it is fine! Which is the big problem? 2400 steps per revolution is great! very good resolution. I am using a 2000 steps/rev stepper for my whiskey compass. With an arduino nano you can easily achive more than one revolution per second with this resolution! You dont need any more.. I think.

Which microcontroller are you using?

And about sound (if not really loud) is part of analogic gauges magic :) It is music :P

Arturo.

[BushPilotWannabe]

it is fine! Which is the big problem? 2400 steps per revolution is great! very good resolution. I am using a 2000 steps/rev stepper for my whiskey compass. With an arduino nano you can easily achive more than one revolution per second with this resolution! You dont need any more.. I think.

Which microcontroller are you using?

And about sound (if not really loud) is part of analogic gauges magic :smile: It is music :P

Arturo.

Thanks for this.

Using a Nano. Controlling the rounding error at 6.66... steps per degree. It will get complicated if direction is changed between multiples of 6 degree heading. I like easy.

This has several gears between stepper and indicator, still about 6 seconds per rotation. There is no way I'm going to spin anything that fast.

Noise may come with an 'as is' purchase. Now I find out that level of the noise, and voltage required, change with rate of rotation. At fastest rotation, the stepper was quiet at 4.2V. 200mV lower and the motor stalled, 200mV higher and the noise got louder. Changing to 25ms between steps - about that of a rate one turn - and the motor stalls if less than one volt higher.

This stepper motor has a 1/8", maybe 10 tooth output shaft. I don't want to screw it up. Amazon has a new (old stock) stepper for around $550. Or you get two used steppers with another RMI. Machining an adapter for an off the shelf, eBay, stepper motor is way beyond my ability.

The RMI does not move the stepper to an absolute position, but moves it left or right as required to normalize the waveform on the rotor of the resolver geared to the indicator. This makes me think that the voltage applied to the encoder coils may not be square DC voltage pulses.

I'll try everything. Hope it is as simple as a sawtooth (inductor - capacitor) circuit to preload the mechanism. Must be noise from the mechanism. There is no noise from the stepper when unloaded.

Next, the Nav indicator is driven by a resolver receiver.

I forgot. Does anybody have any information on the tool required to remove the setscrews retaining the gears? Looks like @ .060" diameter 4 spline internal Torx bit, if they make one.

Hugh

[_alioth_]

T

....to normalize the waveform on the rotor of the resolver geared to the indicator...
...Next, the Nav indicator is driven by a resolver receiver...




Are you going to use the whole electronics? or are you trying to read resolvers and then move steppers?
Are you able to read resolvers with arduino?

About speed, 6 secs por revolution is not bad anyway.

Arturo

[BushPilotWannabe]

I am connecting directly to the encoders. Do not have time, expertise, or money to copy Joe Sim's panel.
https://www.youtube.com/channel/UCsnm5PJ98J_x2Em5FCm8VTQ/videos

As I understand resolvers, any input and output is in the shape and distortion of sinusoidal waveforms. This is rendered in the high density transistor PCB's permanently removed from the RMI. I am trying to control a surplus resolver receiver (two phase rotor 2 phase fields) by applying constant DC voltage to the rotor and changing voltage and polarity on the field windings. So far movement around the eight points of the compass is available but my potentiometer controlled, power op amps voltage follower, circuit is way too crude. Alternately will scrap another RMI and try putting the two synchro receivers (single phase rotor 3 phase fields) back to back and use an encoder to turn the shaft of the loose synchro to move rotate the synchro bolted to the RMI. I wonder how much control is lost using low voltage 60Hz power to the field rather than 400Hz?
https://www.youtube.com/watch?v=N-UxTkTK3p8

Hugh


Are you going to use the whole electronics? or are you trying to read resolvers and then move steppers?
Are you able to read resolvers with arduino?

About speed, 6 secs por revolution is not bad anyway.

Arturo

[_alioth_]

I like reading your ideas and progress.

I have interfaced some HSI, and RMI, servo altimeters, and I choose another path. Replacing original synchro resolvers with 10bit PWM absolute encoders.
10bits is 1024 steps por revolution, 0.35 degrees. Quite good. The best part is you only need 1 arduino digital pin to read the signal.
Absolute encoders are not cheap. About 50$ each, but interface them is quite straightforward.
Reading the encoders, then you drive the dc or stepper motors, and that's all :)

Last weekend I interfaced an attitude indicator. I used an absolute encoder + dc motor for bank.

Here is a vid of it:
https://www.youtube.com/watch?v=F4hwjwpkHTU

Another one, simulating startup sequence with gyro speeding:
https://www.youtube.com/watch?v=xtPEWfXD97E

Arturo.

[BushPilotWannabe]

Nicely done, sir. Space is limited in the front section. I may be able to use the body of the encoders and bring a shaft from an other motor through the rear bearing of the encoder and couple the original shaft and driving shaft between the front and rear bearings of an otherwise empty shell. If the noise is a result of gear lash, the solution must include a rotational damper.

Hugh

[_alioth_]

Sometimes, what I do (just to give ideas), is absolutely replace the synchro with the encoder.
The encoders that I use are small, so I 3d print a chasis to make them exactly the same dimensions as original resolver.
Look. This was to make a vor:

left, the synchro resolver, right the absolute encoder.

http://www.mycockpit.org/forums/attachment.php?attachmentid=11916&d=1480870677




With the 3d print chasis, it is the same resolver dimensions, so I can use the same srews

http://www.mycockpit.org/forums/attachment.php?attachmentid=11917&d=1480870866


Replaced:

http://www.mycockpit.org/forums/attachment.php?attachmentid=11918&d=1480870879

Arturo.

[BushPilotWannabe]

I found enough information about the stepper motors to determine that they are unipolar with a common +9.5V input and the ground connection to the coil is sinusoidal current vs time. I'm not to confident of success but will have a look at the scrapped pcbs.

Hugh