PWM/HPWM Stepper Motor Controller

[DavyJones]

Mark thanks.... I was looking at power supplies online last night. Do switching type power supplies present any problems for the stepper motor? I see quite a few cheap ones online and some even have 24v/12v/5v output which would be perfect for my project without have to have any type of step down converters. 24v for stepper, 12v for solenoids and 5v for pic and accessories. I've probably pitched to many old pc's with perfectly good power supplies in them over the years whats wrong with me.... I just re-read your post I've got boxes of old laptop power supplies around here I just found an old dell one that is 19v 3 amp

Thanks
David

[DavyJones]

I know we are getting a little off topic with regards to PIC's and Picbasic I hope that's ok my project overall is dependent on the PIC and the programming needed.

Sorry about the noise in the background my wife was watching TV.


I'm thinking of something like this to power the overall project which needs 3 different voltages.
Then the step down for the other 2 both are relatively inexpensive.

[DavyJones]

I was thinking ahead a little bit and for now my testing has been to test the solenoid activation and stepper motor each by themselves. Now they will need to work in conjunction with each other wherein the stepper motor will need to continue to move (and keep the conveyer belt moving) at the same time the solenoids are cycled on and off. There will be gaps in between the items I will be spraying. I am sure I can complete this using 2 picks, one to control the solenoids and another to control the stepper motor. I would add an interupt test to the stepper pic so after the last row has passed I can stop the stepper anywhere in it's loop. I don't see anyway around this since everything is single threaded (at least thats what I've always thought) and branching from the stepper routine would add timing to the stepper loop which I think would through everything out of whack. using 2 pics is no problem and getting them to talk to each other not going to be a problem either because I think I'd just need to use interupts either with an INT function or just checking the state of a pin at some point.

I wanted to know though if anyone has ever used one of these and/or if there is any support in Picbasic for this dsPIC33CH. It's funny the example they choose to show what you can do with this is pretty close to what I am needing this for.

[HenrikOlsson]

David,
You've done a great job so far!
As you might know the torque of step motor falls off as the speed increases. Increasing the drive voltage counteracts this drop in torque. Doubling the voltage pretty much doubles the speed at which the motor stalls due to "overload". This is within reason of course, too much will only result in heat in the motor and if you do get what you need out of it with whatever voltage there's no real gain in increasing it even further.

One rule of thumb from a drive manufacturer (Geckodrive) is that supply voltage to the drive could be as high as 32*(sqrt)L where L is the inductance in mH.

Lets say your motor has an inductance of 2.8mH, then the drive voltage can be around 50V. Obviously the drive has to be rated for the voltage as well.
I don't know any details on your particular motor but I'd certainly try a 24V supply or even a 48V if you still need more speed.

Slow speed torque is highly dependent on the current setting of the drive.

Finally, it's hard for a motor to start "instantly". You'll be able to reach much hiher speeds by ramping the speed up/down.

If you generate the step pulses in hardware, like with the CCP module (as with HPWM ) then you have all the time in the world to figure out when to fire the solenoid(s).

Nope, PBP is for 8-bit PICs only, no 16 or 32bit devices.

/Henrik.

[DavyJones]

Henrik, Thank you and thanks for your explanation. I've not done any work with the stepper motors so I'm still learning. I'm actually not trying to get much speed out of the stepper. I will be driving a conveyer and it will be moving small parts under my spray nozzle so it's not going to go fast at all so the parts can get evenly coated.

I'm going to start to look into generating the pulse with HPWM I've never done that before

Thanks for letting me know about that ds device good to know.

David

[mark_s]

If you generate the step pulses in hardware, like with the CCP module (as with HPWM ) then you have all the time in the world to figure out when to fire the solenoid(s).

/Henrik.

Henrik, I always wanted to try using the CCP module for this purpose, but not sure how to keep track of the number of pulses. Maybe feed the pwm output into a timer to keep count?

Mark

[DavyJones]

Mark maybe we can work this out together with Henrik's assistance. I went to the micro chip site and used their processor selector to find one that had the PWM built in. It looks like the 12F752 has what I need to try to get the HPWM to work. The datasheet says

• Capture, Compare, PWM (CCP) module:
- 16-bit Capture, max. resolution = 12.5 ns
- Compare, max. resolution = 200 ns
- 10-bit PWM, max. frequency = 20 kHz

Thats what I am looking for right? I've not read the rest of the datasheet. I picked this one out because I think I have one here already to start to work with.
Thanks
David

[HenrikOlsson]

Mark,
Yeah, for precise positioning feeding the pulses back into the PIC and count them OR integrate generated velocity over time to calculate position. Possible but not exactly easy to get right, I'd say.
But for the application described here I don't think precise position control is needed, just precise velocity which the CCP module will do without any software overhead - except during acceleration.

Depending on the needed frequency and acceptable jitter perhaps the NCO available in some devices would be even better for this.

David,
At your desired conveyor belt speed, what approximate frequency are you generating? Is it 100Hz, 1000Hz, 5000Hz?
For the CCP module you could use the HPWM command but if you want to implement acceleration then I would ditch HPWM and set the registers manually.

/Henrik.