Electrical Engineering Stack Exchange is a question and answer site for electronics and electrical engineering professionals, students, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
I am building a stage for tissue cutting and the motion of the stage should be very slow and smooth. The speed range is 0.1 to 1 mm/s. I got a stepper linear motor from PI (M230.10s), I test its speed at 0.1 mm/s and the chatter is 3 times of the setting value.
I hope the motor can motor can move very smooth with chatter around 1%. Is it possible?
One option, of course, would be to add gear/belt/leadscrew drive to the stepper so it can be run at high speed while the mechanism it drives operates at low speed. This way, it'd act very much like a long lever. A large ratio of motor to motion yields MUCH smoother travel and higher torque.
Big vote for the 'obvious' suggestion by Nicolas D to use microstepping which will make the motor move smoothly, 16-microstep driver chips (Allegro A4988 for example) are readily available and there are some drivers which go much further (64 microsteps or more).
However, if you need small, smooth, accurate movements the best thing you can do is gear the system down so that you are not relying on the limits of motor quality for smoothness. Moving 0.1 to 1mm/sec is not a very big range of speeds, you could probably gear the drive from the motor down to a ratio of 100:1 or lower (so 100x better accuracy). If you know the maximum speed the motor can run, choose gearing that will move the device at your maximum speed (1mm/sec) at the maximum speed of the motor. This will give you a far better range of control and increase the achievable accuracy.
You don't say what the gearing of your system is - as you state movement in mm/sec I assume you have some sort of lead-screw style mechanism.
Even quite basic CNC machines / 3D printers can achieve movement accuracy of 0.1mm whilst moving a lot faster than 1mm/sec, using very cheap stepper motors & drives. Most of the inaccuracy ends up coming from the play in the drive system (usually lead screws) or flex in the machine body due to mechanical loads.
Maybe I'm missing the obvious here, but... If you didn't drive your stepper with microsteps, you might want to try that first. This allows the stepper to go smoothly in between "normal" steps.
A stepper motor can keep track of it's position by itself (1) without requiring a position sensor for feedback. That's why steppers are popular. On the other hand steps are not smooth, because... well... because they are steps.
There are other types of motors that don't have steps: brushed DC motors, brushless DC motors. With these types of motors, however, a feedback sensor is needed to know the position of the motor.
I would suggest to begin with a brushed DC motor, because it's easier to drive, and it doesn't appear that application requires high performance characteristics of a brushless motor.
Another approach to your problem is to create a mechanical damper, which would help smooth out vibration.
(1) Unless the load torque is too high, which causes the stepper to skip steps. Fortunately, that's not the concern of this thread.
I recommend a DC brushed gearmotor. Brushless motors tend to have ripple around their commutation points which brushed motors don't have. The gearing in a gear motor allows you to get drive ratios like 1000:1, which is important since it magnifies torque (which increases cutting power, so varying tissue densities aren't a problem) and reduces speed. I also haven't seen brushless gearmotors, so you'd have to get a brushed gearmotor anyway. Just be aware that gearmotors will have a little play (rotation before the gears lock) when switching directions.
More importantly, you want a good control system for it. Ideally, you would use a servo driver (such as a Copley Controls or AMC) with encoder feedback. This would allow you to easily do velocity control on your motor. If that's not an option, then you would need a variable voltage supply with a decent current rating. Maximum motor speed is directly proportional to the voltage across the motor, so a lower voltage means a lower speed.
If none of the above options would work, then you could look into building/buying a viscous damper for the motor. It essentially be a fan blade, encased in an oil, attached to the motor shaft. It creates a viscous drag on the shaft that dampens oscillations and prevents the motor from reaching high speeds.
If I understand datasheet and your question then motor resolution is 50nm corresponding to 1 step. You need drive 100um-1mm/sec, i.e. 2-20kHz, which is high speed at upper motor limit 1.5mm/sec. So IMHO microstepping/gearbox is not solution (it helps for slow speeds but the cost is torque halving as you double micro stepping factor), you need proper ramping to achieve velocity without resonance (i.e. also some kind of chatter). But I think that problem might be derivation. If you calculate/derive at small time interval you see large jitter. I'd look at raw position data.
Note: But I don't understand what is supposed graph waveform, triangle with period depending on cutting travel ? And how the position is measured (as number of microsteps or an external micrometer) ?
For such a slow speed I would go with a standard brushed motor.
Forget about the stepper, you won't get as smooth movement with it as with a brushed motor and anything above 10 uSteps is not accurate so you will end up using an encoder anyway.
Get a high voltage brushed motor (24V at least) so you can drive it very slowly with a large gear ratio and an encoder for feedback. this motor will do the job with minimum damage to your wallet.
