high precision drive for solar telescope

[chal]

Hello,

I want to use dc motors with planetary gears to drive two axes of a solar telescope as uniform as possible. One of them makes 0,5 turns within 12 hours, the other makes 0,25 turns within 6 hours and then returns to the starting point within the next 6 hours. At the other hand I need a much higher speed for referencing the system (1rpm would be nice). Looking for the motors and the gears I was told that the minimum speed of the motors should be 100 to 150rpm to get uniform motion. Could I go down to 30rpm or lower and still have good control of the speed? What resolution should the encoders have? Thanks for your help!

 

[OperaHouse]

I think I would use a steper motor with a microstepper drive. That would give you low speed, accuracy, holding power, and higher speed motion with a minimum of hardware.

 

[chal]

Hi OperaHouse,
thanks for the tip! I already thought about using stepper motors but we experienced with another telescope control that the speed resolution wasn't good enough and we had to replace it by a dc motor.

 

[cbarn24050]

Hi, sounds like to trying to do this the hard way. You can use a fixed speed clock motor to drive 1 axis, and a small motor/gearbox for the other.

 

[jbartos]

Suggestion: Low speed induction motors with vector drives might be the solution.

 

[OperaHouse]

A microstep driver can increase the number of steps per revolution by more than a factor of a thousand. You can change the number of steps on the fly to get positioning speed and then switch back to high resolution. Buy a couple of $200 drivers and connect them to a laptop and you have a system. You get to use any old stepper motor you have around. Are you going to take time lapse photos? You will need a lot of resolution.

I remember back in early 70's. A mentor of mine would build onesies of custom low light level TV cameras for Hughes. He had one connected to his telescope so we could watch in comfort in the dead of winter. He would point it to the moon and turn off the clock drive. It looked like you were flying over the surface of the moon.

 

[wired1]

I think the 100 rpm parameter was given as a minspeed to minimize the effects of cogging or “torque ripple” from the motor. You might ask the source of that info to give his basis for it. Maybe he thinks your app requires a low speed performance level that it really doesn’t. I would think 30rpm is do-able, but what about the high end? Going from .5 turns per 12 hours to 1rpm calculates out to a speed range of 1440. You need two drives – one for tracking and one for referencing. My previous company designed a planetary for their product that had two input gears – high speed and low speed. Can your source for the planetary do that?

On the other hand, a single stepper motor could probably handle the speed range if you can clarify the resolution requirements.

 

[chal]

I just talked to a supplier of steppermotors and servodrives. He recommended to use servomotors because it might be a problem to get uniform motion below 30rpm with a stepper motor. (one axis had to change speed from +30rpm to -30rpm within 2 hours, ratio of the planetary 50000:1, resolution <2.5rpm)

to cbarn24050: I'd prefer the easy way, but our requirements for accuracy are pretty tough...

to jbartos: are there small versions of induction motors? We need about 120Nm after the planetary and the motors I found therefore had been extremely oversized.

to Operahouse: Are you talking about resolution of the telescope? We want to get down to +/-0.5 arcseconds. We'll only do solar observation, there are some nice things to see too, e.g. flares, sun spots...

to wired1: the high end isn't as critical, but higher speed would be more comfortable. No, we won't get a special planetary, but maybe we can combine two standard ones. What do you think about the resolution/ motion profile above? Do you think it's do-able with a dc motor, digital encoder (e.g. 1024 counts) and a servocontroller?

Thanks you all for the information!

 

[wired1]

my comments follow:

-resolution: With a stepper or a servo, it will be possible to achieve the resolution but the accuracy will be difficult. Will the mechanical system be able to hold to the .5 arcsecond level?

-digital encoder: How is the system set up/aligned/reset? If a one-time-only setup, it would seem to call for an absolute position encoder system. That would involve multiple absolute position resolvers or encoders geared together.

-Dc motor: A servomotor is basically a dc motor specially built to have low inertia.

-Stepper vs servo: Please clarify the statement about stepper motors at low speed. Stepper motors are excellent at low speed. One thing they don’t do well is accel or decel to a different speed. The servo will handle accel/decel better as well as high speed running. Is the supplier saying that 2.5rpm the minimum speed of the servo?

-Gear ratio: assuming motor resolution of 1024, 1024 x 50000gear ratio = 51200000system resolution. I calc about 20x the resolution you need. Remember, the motion controller’s ability to process counts this high might come into play. A 24bit word is typical – that’s only 16 odd million.

 

[SERVOCAM]

Stepper motors are excellent at low speed. This really depends on the motor driver runing it. Typically you will have resonance problems with stepper motors at speeds < 0.5 rps (30rpm) - depends on motor, each is different. There are drives that have anti-resonance and and ways to remove low-speed resonance.

You should be able to use stepper, brush servo or brushless servo. Just consider system resolution and how good the controller can regulate the velocity. Your budget also may push you to one of the specific technologies.

Maybe your 100 - 150 rpm spec was because the gears are pretty sloppy and chatter at slower speeds? Or maybe your source is not aware of other technologies?

Take a look at my AZ/EL Gimbals for some ideas. We do a lot of applications like this:

http://www.aerotech.com/products/optmnts/optical_mounts.html

Cameron Anderson - Sales & Applications Engineer
Aerotech, Inc. -

http://www.aerotech.com

&quot;Dedicated to the Science of Motion&quot;

 

[OperaHouse]

Just for your edification, you should look at Intelegent Motion Systems at

http://www.imshome.com

and look up the IMS483 microstepper. This can turn a standard stepper motor into one with 51,000 steps per revolution. Of course that depends on load because it creates a proportional magnetic field. This could eliminate you friends objection to steppers. I just like devices which are very adaptable.

 

[jbartos]

Suggestion to chal (Electrical) Sep 9, 2003 marked ///\\
to jbartos: are there small versions of induction motors? We need about 120Nm after the planetary and the motors I found therefore had been extremely oversized.
///Visit

http://www.epanorama.net/links/motorcontrol.html

for motor controls options available.
The 12-pole or higher pole induction motor might be custom made. It will have lower power factor, e.g. 0.5 which may have to be compensated for, e.g. by capacitors, if it poses an inconvenience.\\

 

[jbartos]

Suggestion: Visit

http://www.e-insite.net/ednmag/archives/1994/070794/14df3.htm

for: 12-pole spindle motor

http://www.aosmithmotors.com/html/motorDoctor/Determiningspeed.htm

for:
Less-common pole configurations include 12-pole motors (600 RPM) that are used in applications requiring slow speeds, such as washing machines, and 16-pole motors (450 RPM unloaded), often found in ceiling fans.