Brushless motor control 1

[RadioGold]

I produce a 4 motor rotocasting machine (as it rotates 2 aluminum frames, one within the other, the load on the motors is constantly changing). Taking the example for one axis I have 2 sensorless bl motors (spinning in opposite directions) connected to gearboxes (250:1) spinning a frame suspended between them. The mechanical connection between the motors is sloppy compared to locking their shafts together. I use a pic microcontroller to supply the signal to hobby store ESC’s on each motor. I want to improve this situation for 2 reasons – The ESC’s, though cheap, automatically shutdown when additional torque is required? and more importantly I need to know the rpm’s to continually adjust the motors to the current speed setting. The various speeds required throughout the casting process are downloaded from a PC.
The motors are hobby store variety (750 rpm/volt 150w) running on a 9v supply.
These are the options as far as I know:
I can purchase motor control chips to go between my pic and 6 driving transistors and get feedback of the actual speed. Which chip can do this I do not know at the moment?
Or
I can mount my own reflective encoder (simple as 3 spokes 3 spaces) on one or (hope not) both of the motors and run the transistors direct?
Or
I can simply look at the 8 spinning spokes on the ends of the motors (reflective sensor with digital or ADC to the pic) and lookup/calculate the motor position based on the readings?

I don’t have a problem doing any kind of programming on the pic; I’ve used them for years in commercial applications.
I have to use the existing 9v switching power supply that fits my enclosure.

Any ideas or circuit recommendations would be appreciated.

 

[zekeman]

I would like to help ,but you are not too clear; a sketch or a clearer description is needed.

I don't understand the need for 2 motors to drive each frame.

Are you using a differential?

 

[RadioGold]

I have a site:

http://www.coldgold.rotoquad.com

The machine is under the "Roto Casting Machine" link.
The machine is torqueless.
The drive is better distributed.
I use gearing so I can use small motors but more important I need to monitor the speed (hard to do on a frame doing 10 rpm).

 

[MikeHalloran]

If you are going to go into the motor drive circuit to get a speed signal, you only need to monitor one phase, not all three, since they're locked together anyway. So as to avoid messing it up, I'd put an optoisolator's source in parallel with a motor winding, with some current limiting resistors and such of course.
I'd be more inclined to go with a retroreflective or Hall effect sensor looking at the rotating assembly, and not mess with circuits supplied by someone else.
I'm a little fuzzy about what your control system looks like now.



Mike Halloran
Pembroke Pines, FL, USA

 

[RadioGold]

Right now I have 2 pic's and esc's on each motor. 1 pic runs the 2 main frame motors via their esc's. It also hooks to a PC via USB to get program instructions. It transmits the speeds and cycle times to the 2nd pic running the inner frame (the frames run independently).
There is no feedback from the motors at the moment and I'd like to get rid of the esc's.

 

[MikeHalloran]

This is the part where you reveal part numbers for the motors and escs and pics if you want more help.



Mike Halloran
Pembroke Pines, FL, USA

 

[RadioGold]

Your getting too personal we hardly know each other!
The motors are 9v 150w max. A pic is a pic is a pic and the esc's are gone.

For anyboby else out there what's the easiest why to drive these motors (min. parts) and the programming is not an issue. I don't mine puting a pic on each motor if need be. Without a driver chip I was thinking 6 darlington transistors on each pic (got watts to burn). What's the min. in feedback that I'd need to run a motor (or better a pair) smoothly.

 

[zekeman]

"The machine is torqueless."

What is "torqueless".

If you need 2 motors to drive one shaft , wouldn't you be better served by using a phase locked loop control system so both motors will lock onto the same signal and when in lock you don't need much speed control circuitry; you can only use one encoder for the feedback signal.
Distributing the load on the motors will be tricky due to backlash, but can be done.

 

[RadioGold]

Torqueless - As being without torque.

Other machines only drive one side of a frame (one design uses a single motor and fixed gearing between the main frame and the inner frame) and use a simple bearing on the other side. It wouldn't matter if the molds that get mounted on the machine were perfectly balanced and centered but that's not the case. Also within the molds you have liquid plastic that gets thicker as it cures to a solid.

I've looked at circuits that use a couple of resistors and a transistor times 3 to feed back the signal from the coil that's not being activated at that particular time. Looking for the zero crossing voltage I beleive?
The circuit also had 6 fets with 6 firing transistors all hooked up to the pic.
I think I'll just have to experiment:
1 reflective sensor looking at the 8 spinning spokes on the end of the motor with a pic and 6 darlington transitors?

 

[VEBill]

If I were designing such a two axis frame, I'd instinctively use gears and drive shafts (*) to bring the inner frame drive shaft out through a hollow axle on one end of the outer moving frame. So the outer frame drive shaft would be exit on one side (trivial), and the inner frame drive shaft would exit on the other side (slightly non-trivial). Obviously the outer frame drive would also drive the inner, so you'd need to account for that. The end result would be that both motors (and their associated position sensors) are mounted to the fixed frame.

At that point the sensors become trivial. Infinite options (e.g. shaft encoders). Which is why this approach, although slightly more complicated at the outset, makes things much easier when you get to this point.

(* Another option is a speedometer cable snaking from driving the inner frame out through the hollow axle of the outer frame. Might be easier than solid shafts and gears to turn the corners.)


Q: Why is it for sale with a "sophisticated PC based control system" if you've still not yet sorted out the sensor technology? Is there a temporal discontinuity? Are you from the future?

 

[Skogsgurra]

150 W at 9 V? Sounds highly impractical. And so do many other details. Try to get in touch with a down-to earth person that knows about drives and physical laws. Pouring Pics all over the place won't help.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[zekeman]

"Torqueless - As being without torque.

Other machines only drive one side of a frame (one design uses a single motor and fixed gearing between the main frame and the inner frame) and use a simple bearing on the other side. It wouldn't matter if the molds that get mounted on the machine were perfectly balanced and centered but that's not the case. Also within the molds you have liquid plastic that gets thicker as it cures to a solid."

Sounds more like a sales pitch than reality. You have not made a compelling case for 2 motors and moreover, overcoming imbalance would not be enhanced. 2 motor control to me appears unnecessarily more complicated. Also I have a hard time understanding why one would need precise speed control for this process and pay for the system that would try to achieve it. Plus there is no way you can get it.

Speed control of slow speed unbalanced loading to me is a nightmare and I will bet it is not solved by your control methods. The 2 axis dynamics problem is the problem including gyroscopic effects.

Before I would even begin talking about speed control, I would attack the dynamics, then get a realistic speed control. I think simpler is better.

As somebody has said, you need to hire someone skilled in dynamics and control methods, perhaps 2 consultants.

 

[RadioGold]

Thank you all.