Anomie

I have looked around the web, and can't find much info on this subject. Has anyone here successfully operated a stepper motor controller with an RC receiver? I need to adapt such a controller, but I'd prefer it behave like a servo; i.e. recentering itself and exhibiting similar variation in speed depending upon the amount of stick movement. The stepper controller is fairly straightforward in operation, but I have not the first clue as to adapting it to the output of the receiver. I cannot adapt a servo to this system as they are quite slow in comparison. Worst case, the stepper can be driven by a relay for simple on/off operation. I am hoping this is not necessary.

I know this is a bit of a stretch, but if anyone can point me in the direction of something helpful, I would really appreciate it.

dserratt

My Feedback: (1)

I am looking for the same thing I will let you know if I find anything.

grotto2

My Feedback: (2)

If I read you guys correctly, you're wanting to drive a robot vehicle with the tires directly connected to the shaft of a stepper motor. And you'd implement steering by applying differential or reverse speeds to the tires (or track?).

I'm surprised you have found nothing on this. I've done similar projects reading PWM from channel outputs and driving stepper motors. I've just never put those two elements together, as it turns out. PIC processors like the Basic Stamp (or even better yet the Picaxe) are just perfect for this kind of application.

My experience with steppers was not good. They're slow and weak. If I were to tackle the problem I might find some strong DC motors and modify standard Digital servo amps with some beefier output transistors to drive 'em. Just set the feedback pot to give no motion at center stick. You can get a feel for what will happen by removing the gears from a working servo and watching what will happen when plugged into an active channel.

Hope this helps a little. I'm no expert.

-Ron

Anomie

Hi Ron,

Thanks for the input and info on my post. The purpose of the stepper in my project is to drive a linear actuator for the steering. The stepper is a 1.2A/phase motor and should provide adequate torque through a worm-drive. I still have yet to receive the controller in the mail, and when it arrives there will be need for some experimentation to ensure this will operate properly before fabricating mounts and wiring.

I'm still poking around for a way to make this work. It's likely much simpler than I believe, and some work with the receiver and stepper controller will probably enlighten me.

Reiko

Anomie

Ron, thanks for the idea on the microcontroller. You led me in the right direction. :-)

I have made progress on this project. The stepper controller finally came in the mail, and after studying about the switching requirements of the motor, I now have an idea of a method. The stepper controller handles the switching of the motor coils but requires an additional digital device for telling it when to switch. I am going to get a BASIC Stamp microcontroller and program it for the step direction/rate/limits for the stepper controller. The Stamp is very small, and has plenty of memory to handle the operation of more than one device. I found a web site, [link=http://www.parallax.com]www.parallax.com[/link], and they have some excellent information for the microcontrollers. The Stamp is $14.95 and the manual and software are free to download. It interfaces with a PC via a serial port for programming, and then retains the program in memory. I still have some work to do, but it should come together soon, and I'll post the results of my research here.

~Reiko

grotto2

My Feedback: (2)

Good luck, Reiko. I think you'll find that Stamp is really easy to program to do what you want it to do.

Say, I guess you're going to connect a pot to the steering for position feedback.

-Ron

Anomie

Probably. The limits of the steering can be ascertained while programming, and then factored in, but some data output would be cool. If I ever intend to make this truck autonomous, it would be one of the first steps.

Anomie

So, a little update on my slow progress.

I ran some tests on a pistol-grip radio and receiver today, and the results are serving to make this relationship more complex. The receiver steering output pulse-width changes by ~300usec when the wheel is turned left or right. The change is actually positive when in left steer, and negative when in the right. Even with an inexpensive radio system that I borrowed, the accuracy and resolution of the pulse is extremely clean. I can actually see a variance down to 1 microsecond on the oscilloscope (and it's not even a Futaba!). At least I will have a consistent signal with which to interact.

Unfortunately, the output of the receiver must talk to the Basic Stamp in order to switch the stepper controller properly. As of now, I have no knowledge of whether or not the processor can interpret this type of signal, or if I need to somehow convert that modulation to binary information before sending it to the input of the Stamp. A demodulator? Who knows.... For now, I'll continue to plod along.

Anyway, if someone has experience with something along these lines, I would appreciate some input.

~R

Anomie

I should have ventured beyond high school electronics. After discussing this system with a coworker, I realize that the detection of the rise and subsequent drop in pulse is what needs to be ascertained. The time between makes little difference in that the device must be digital - the output goes high (or 1) at the rise, and goes low (or 0) at the fall. This process must happen quickly enough for the Basic program to decipher the bit rate and perform the necessary switching of the stepper coils. I believe one step for this process would be to investigate the process taking place inside a servo.

More research, and more time.

sfsjkid

Naw, Reiko, they don't teach you this kind of neat stuff in school!

I was wondering do you have to have a stepper? I took a look at your cool truck and you mentioned a lock to lock time of 1.1 second, and I'm going to guess your stepper is around ~50oz.-in. Servos today well exceed these types of numbers. I think Hitec came out with a 400oz job recently. Admittedly it might not be as sexy.

Anyways, its not clear if you meant the servo control pulse in your last post when you said the time between the rise/fall was not important, but if you did I just wanted to say that it is exactly the width of this pulse the servo is looking for, ie the pulse width is what controls the servo position. Grotto mentioned this PWM signal. I'm not sure if it is fast enough, but my first impression is that I see at least an A to D converter in your future?

This might save you a chip, though with more programming, http://www.avr-asm-tutorial.net/avr_en/AVR_ADC500.html.

Good luck!

Anomie

Well, that A to D is one more educational step to bring this craziness to reality. Thank you very much for bringing that to my attention.

To address your points, I would like to run a stepper for two reasons; first and foremost, to see if I can dream up this type of RC steering and actually make it work, and second, to drive this project truck in a completely unconventional direction. I don't know if I would use the term 'sexy', but to me it is intriguing. It's also a pain sometimes.

As to the pulse width as a function, I agree it is absolutely the manner in which the servo is positioned. When I said it makes no difference, what I meant to say is that it makes no difference to the device interpreting that time - just the fact that the edges of the pulse's rise and fall need to be read properly. I didn't phrase that very well

sfsjkid

I understand completely Reiko, thanks for the clarification and I do wish you the best of luck. I'll chime in from time to time.

jdb000001

If the processor has a timer/counter, you can measure the pulse width from the receiver by using it to gate the counter/timer.
You will probably need an A/D input also, so you can use a pot for feedback on the position of the stepper motor.

Anomie

I'm hoping I won't need feedback on the motor's position. I would like the programming to include the number of steps on the motor to the steering limit in each direction. This may not be possible, but so far I have no reason to believe such.

As for the timer/counter, the Stamp can indeed work as both, and I'm working on a small flow chart to outline the commands.

I've decided to initially set up the stepper controller to move the steering via a switch. Some things need to be worked out prior to making a connection between the receiver and Stamp. First, I need to research the step rate limit to be certain that this motor can do the job. The controller documentation indicates a maximum stepper frequency of 40kHz. What I don't know is if this equates to the actual step rate of the motor. If so, that means a top speed of 6000rpm (400 steps/rev, or 0.9 deg/step with this motor), and a maximum lock-to-lock steering time of 100 msec. That's way too fast, but the rate can certainly be slowed.

Second, I need to be sure that the torque required to steer those tires is not beyond this motor.

Obviously, I need to set up a test rig before going nuts with the Stamp. I'll update as soon as I have some further work done.

~R

jdb000001

The stepper motor should have a rating showing the maximum stepping rate for continuous rotation.

You will need to have some type of feedback on the motor position - either switches at the end limits, or a center position indicator, if you don't want a feedback pot. The controller has to have some way of telling the initial position of the motor. There's always the chance the tires will collide with something and knock the steering around, or get in a position where they will not turn.

Anomie

Unfortunately, there is no information on the motor itself. I have a spec sheet, but nothing on maximum rate there either. Judging by other Nema 17 motors, it would seem to be beyond the 12kHz which I am hoping is required.

The steering is constructed such that the wheels cannot force the motor to move. The actuation is linear, involving a lead screw pushing and pulling sections of wire rope that swing the steering by way of a large bellcrank. I have attached a picture of the actuator for reference.



When the motor is attached, it will operate through a reduction gear for a little torque amplification. It may end up a worm-drive, if that much reduction can be had. The actuator will have a tremendous mechanical advantage over the wheels, and because of the lead screw, will not be allowed to move except by the motor. I think that once the motor's position is programmed, it will not change unless told to do so by the stepper controller.

I realize how ridiculous this system may seem, but I am really hoping to operate it reliably in the manner I've designed.

Anomie

Since my last post here, I've advanced the work on this steering system a bit. I know my posts here are far between, but I don't have as much free time as in the past. Progress is relatively slow.

After getting hold of the Basic Stamp and stepper controller, I have designed a PC board to operate as a voltage regulator and also a place to mount and interface all of the steering-related electronics. I have also worked out a preliminary program for interpreting the modulation from the receiver output and sending appropriate information to an L297 control chip. The program is very simple in that it basically reads the incoming pulse width, determines if the intention is to steer right or left, and then performs a few simple calculations before sending a series of 10uS pulses to the L297 clock. The motor will then full-step whatever value is directed by the microcontroller. The mechanical limits of the steering will not be reached.

I will perform a few tests on the oscilloscope before attempting to operate the motor's controller board. Once the bugs are worked out, I can have the PC board manufactured. After that, I can mount the stepper and gearing, and the steering will finally be complete - and after only 18 months.

I have included a shot of the board design, and you can see the mounting locations for both the stepper controller and microcontroller. The 4 regulated outputs will power the ESC, receiver, LEDs, and both the Stamp and stepper boards. The main 12V input will be supplied by 10 NiMh sub-c cells. I may still go with laptop batteries for the mains, but that will involve some additional work for charging.

I want to thank grotto for the idea of using a Basic Stamp. It has very much simplified the development of this crazy project, and I appreciate the help.

dirtybird

My Feedback: (5)

You really should spend some time studying how servos work. For instance find out what open loop and closed loop mean and the difference between the two.

Anomie

The system I am experimenting with is basically an open-loop servo, albeit much more complicated. I don't have any use for an actual RC servo, as the steering lead screw requires much higher motor RPM in order to swing the steering from lock to lock in a reasonable amount of time. This fact is a result of a 27:1 double worm reduction. It is also not important for the controlling program to receive feedback from the motor. It will not be mechanically possible for the motor to be forced by the wheels.

The major downside to this is the fact that parts will break unless there is an engineered weak point or servo saver in the system.

dirtybird

My Feedback: (5)

You should also study gear ratios a bit. What you are trying to do could easily be done driving the lead screw from a servo with the proper gears. Then place a linear pot on the lead screw to close the loop. In fact some of the early servos worked this way. (Kraft KPS-7 was a linear servo, so was the early Bonner servo)
Open loop servos are very wasteful of power and hard to make do what you want them to do

Anomie

What do I need to study about gear ratios? The lead screw pictured above is a .100" pitch, and must travel 9 turns in order to swing the steering from full left to full right. In order for the steering to move with any decent speed (i.e. ~1 second from lock to lock), the motor used must be capable of at least 540rpm for that second. The fastest servos I have found thus far cannot rotate that quickly, even when modified for continuous rotation. The reason for the extra 3:1 gearing is to allow the motor to step at a higher rate. This will keep the motor's voltage up and provide more holding torque when idle.

dirtybird

My Feedback: (5)

Most motors will turn about 6000 rpm with no load. For 540 rpm you would need a reduction of about 10:1. So all you have to do is remove the last three gears from the servo and drive the lead screw from the second gear. That will give you the proper reduction.
Then get a linear trim pot and remove the drive screw and connect the wiper to your lead screw. If the servo drives to one end and stalls reverse the leads to the pot.
You will have a linear lead screw servo. It will work much better than the open loop stepper motor you are trying to make.

Anomie

Well, the intrigue for me is to build something unconventional, and the journey is the draw. I could operate this steering system in any one of a hundred ways, but my challenge has been designing and building this from my own ideas, and not simply making it operate. Making the stepper and microprocessor work seamlessly with an RC receiver is something I have not tried in the past, and it is becoming the focal point of my entire truck. If this is strange, so be it.

It would be nice if you didn't point out how well my steering will NOT work before it is completed.

dirtybird

My Feedback: (5)

OK have at it.

Anomie

OK! Here is something I didn't figure... [&:]

Looking back to when you mentioned the 540 RPM from a servo, I have to admit that it has become enticing to try just that. The reasoning is the time I'm taking to get this truck to do anything on it's own, combined with the cost of all of the supporting electronics. The stepper interface may have to remain a separate project so I can go in a different direction here and hopefully operate the steering without too much more investment. The interface is complex, to say the least, and taking an extraordinary amount of planning and experimentation. It can also be a tad frustrating at times. It's hilarious that I so adamantly pushed in support of this design, and now intend to simplify.

So... thank you for offering up the idea, and I'll post here again when I get some more work completed.