Step Motor / Driver Performance

Alphawookie

Hey y'all.
I've been fiddling around with the BOE and a "little step" (for a couple months, my first gizmo , and have run into a few snags.
Every NEMA type motor I've hooked up to the little step fails to deliver anything close to the motor manufacturer's specs.
Different manufacturers, quality and price, some are six wire, some are eight, and all hooked up and running fine to a low speed. I'm testing them on a rig with regulated power, monitored scurrent, and known loads.

When I try to drive a NEMA 17, it starts to jitter aqbout 1100 steps/sec, and stops spinning at 1225 steps/sec.
The NEMA 23 jitters at 620, stops at 650 ST/sec
The NEMA 34 jitters at 180, stops at 200 St/sec
The smaller the motor, the faster it runs.

Torque is similar, but the smaller ones are closer to the rated torque.

The current drawn by the setup is similar as well, the smaller motors draw closer to the rated current.

They all run a little faster with LONG ramp times, but not much.

They all run a little faster, and deliver less torque with wave steps.

At half wave they don't do much in the way of torque at all.

For now, the only code I'm using is a one line serial command to tell steps, speed, and step mode.


I'm looking to use a stepper as a fractional hp motor for a gizmo I'm working on. Steppers seem (by spec) the way to go but I'm stuck. They just won't live up to the specs.

Is there another driver out there somewhere that is worth trying? I need·to get some more power from these motors.

Specs :
Motor voltage 5V
Unipolar, 2.5A / phase
Serial control for use with the BOE.

I saw on posted on another thread, http://www.allegromicro.com/sf/3967/·But it looks more like a chip. The new USB thing by parallax is the form I'm looking for, nice screw terminals and such. I'm getting better at soldering, but need to work on that. Lead solder rules, who knew?. Overall I'm having a great time fiddling with this, but the reuslt isn't what I'd hoped for.

Any leads would be heartily appreciated.

Thanks
Tom






▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Engineering is the art of doing·something well with one dollar, which any bungler can do with two after a fashion.

T Chap

You are going to get more torque the way around with bipolars. The key is higher voltages. If it is rated at 5 hit it with 20 or around there, but you have to limit the current with a chopper circuit to the rated specs. At 5 volts you are are going to get nothing out of any stepper, especially a 23,34, or even 17. But at >24v you can get incredible torque, although at high speeds the torque starts to diminish greatly. I have nema 23 2 amp uni's that I run at over 10 revs per second, at 800 steps per rev, half stepping at 24 volts. Above that the torque is no good. Bi polars outperform by far. There are some off the shelf drivers like the L298 that is just ok, with the 297 and a 298 you could get your feet wet with pwm, current sense, and chopping concepts.

Although functionally the chip linked above is great, you wont get any performance at 750ma. I wish you could get a self contained driver at $2.75 that really did something. If they offer something similar that will let you add your own mosfets then it would be great, maybe they do.

This one is 2.5 amps, 35v rated, so you could try that one with 35volts on the 17, and 23, the 34 could stand more voltage.

http://www.allegromicro.com/sf/3979/

Post Edited (originator99) : 11/5/2006 8:40:09 AM GMT

Alphawookie

Yo!

Thanks for the info. From a complete stranger it's a welcome surprise.

Higher voltages make the magnetic field develop that much faster? I'm just starting with the basic V=IA and didn't figure the voltage mattered. Bucking up the current just to chop it back sounds rather inefficient.·I had hoped to use steppers because of the efficiencies, and torque characteristics they offer.

I'm limited to a 5V, 6A supply by competition rules. Please correct me if I'm wrong or offer any opinion to the following.

If I get you right, I'd have to buck the current up to, say, 24V. Probably lose 15% off the top, 1 amp.

Chopping it up and directing it properly would·eat up·at least·an amp in the electronics, another 1 amp. At least, that's what the Little step does. The total drop across is huge. I know it·matters at 5V, but at 24V, would·it be a non-issue???

The efficiencies gained in the motor seem to be·lost by the supporing electronics. And it would all be very special stuff indeed, by my standards.

I'm a newbie. While I could probably get these steppers·to run off my PC with off the shelf stuff I've found and 110 AC, I'm not likely to do so at 5V with a Stamp. I can barely solder.

PWM control of DC motors is a better solution for my gizmo and limited experience. All the R/C stuff is pretty plug and play.

I'll post some pics of my gizmo once we get closer to the competition. It's fun, fun, fun.

Thanks,

Tom





▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Engineering is the art of doing·something well with one dollar, which any bungler can do with two after a fashion.

T Chap

Bucking up the current to chop it is highly efficient actually. There is no wasted power other than the losses related to the motor's inherant issues. Using other methods to limit current such as power resisters is highly inefficient, more power goes up in heat than gets used. Chopping means only the actual power required to maintain max current is used. Higher voltage means much faster rise time. YOu cna buy off the shelf drivers all over the place depending on your needs(current/whole.half,micro step, number of axis', reads encoder etc). The higher the current the higher the price as usually when you get over 2 amps you need drivers with outboard power mosfets, then you are talking about 8 mosfets per bipolar, 4 for uni to run one motor. There are dozens of sites that sell anything you want in any configuration and price.

Alphawookie

http://www.hsi-inc.com/DCM8028_DCM8055.php
http://www.princetel.com/fc_smd.asp
http://www.oceancontrols.com.au/motor_controller/stepper_motor_controllers.htm
An hour or so with Google and a dial-uo connection produced a couple likely candidated.

The current bucking solutions I came across all look like IC's. That may not be a problem for the gurus, but it's got me stymied. I'll not be able to make a board for this.

Any buck converters that are plug-and-play? Or, a controller with an internal buck/chop option would be great, if it exists and takes 5V as an input.
I'm feeling a bit over my head. I'll digest this for a bit and see what comes out.

It beggs the question, is the effort worth it? Will steppers be twice as efficient as DC gearmotors? I'm having pretty good luck with automotive motors. 35% efficiency or so.

Thanks
Tom

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Engineering is the art of doing·something well with one dollar, which any bungler can do with two after a fashion.

T Chap

Most of the drivers will accept 3-5 volts as logic inputs(step and dir). Designing a board from scratch to manage higher current that 2 amps is a major undertaking. There are simple and cheap solutions for 2 amps, like already mentioned you can try the L297 and L298 combo. Keep in mind you need two L298's to run a bipolar( one for each winding). This cannot be breadboarded, their pin spacings are staggered .05 centers and a pain for testing. You could thoug solder leads onto them for testing. Personally I found them prestty wimpy and ruled them out for any serious torque application. There are many drivers off the shelf that are plug and play, just add step, dir and hook up your motors and run.. You should be reading through CNCzone.com for anything like what youa re looking for as afar as larger current/voltate drivers. Gecko makes some that are widely used, but there are many others. Google cnc controller and see what you come up with.

Rearding efficiency, that is based on the application which if unkown. Bipolos stepppers(and uni's too) with the right driver have incredible low speed torque. If you are just mobing an object with no regard for precision then any large dumb dc motor will provide lots of power. Just depends on what you need to do. Most people wouldn't g to the trouble and expense of adding bipolar drivers just to move an object, as they are limited by speed and also torque at high speeds.

It was really overwhelming to to me when I first started designing drivers. There are many issues to sort out, then it gets easier as you get a fundamental understanding of all the priciples involved. I am very dense but for some that curve can be fast. Just read everything you can on googel about the various methods of chopping, current sense(comparator and high watt resistor on low side of each winding basically), PWM generation, how to implement the chop, whether using AND gates in line between your processor and the driver. The comparator turns on at a set threshold voltage across the sense resistor. For example .1 sense R is around .1 volt per amp. So for 2 amps you set a reference voltage on the comparator at .2 volts, and when the voltage across sense R = .2, the comparator turns on, enabling the PWM to start chopping your driver. There are various techniques to do this, again, based on your needs, preferences and budget.

Alphawookie

Google.
This project would still be a concept were it not for that magnificent piece of work. If I only had a nickle for every google search I've done (in stead of the other way 'round)....... I'm kind of following your mindtrack, but it's still a bit like quantum physics. I get it a little, but the folks who actually make stuff happen with it know WAY more than I can learn in the time I have.

I found some DC converters that look pretty simple, but what I'm looking for is brute force. Cheap, relibable, and most of all EASY! I was hoping I'd be out of the learning curve with steppers, but it's still a long way off. Too long for this rig, but maybe on that big telescope.......

It's fun to play. I'll test some more DC motors to see if I find what I'm looking for. One important thing about design is to eliminate the false avenues before you get too far along, and over budget.

Thanks, You've made up my mind. Sorry steppers, there's other ways to go I know pretty well. Cheaper too.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Engineering is the art of doing·something well with one dollar, which any bungler can do with two after a fashion.