PDA

View Full Version : Stepper rotor/shaft re-mounting?



Rufe0
01-12-2015, 10:59 PM
Hello

This may sound a little odd but I had a bit of a brain wave...

Has anyone ever tried or heard of drilling out the shaft of a stepper motor and then fixing in a new shaft?

Neale
01-12-2015, 11:46 PM
Can I ask why you might want to do this?

Rufe0
02-12-2015, 12:16 AM
Well I can imagine someone would need to do it to repair a motor or to fix it to/drive something that would be difficult without a special shaft of some kind.

My crazy idea is to gang multiple motors together, to create a 6/8/10/12 or more phase motor with corresponding resolution. I guess you could do it with couplers and unmodified motors but it wouldn't be as good.

Neale
02-12-2015, 09:16 AM
As with many interesting ideas, I think people have come up with better ways to do it...
Modern servo drives use a motor to provide turning power, and a separate encoder to measure position. A typical encoder might have 2500 identifiable positions per revolution, giving much higher resolution than steppers with the benefit of full motor power available at all positions, unlike steppers using micro-steps. And in general, stepper motor shafts are capable of taking any torque that the motor could generate - sufficient to stall the motor, at any rate.
I'm not even sure if the price of steppers these days makes them worth repairing, if you've managed to abuse one to the point that the shaft is damaged. There's probably collateral damage internally!

uli12us
02-12-2015, 11:24 AM
Afaik, you won't repair a stepper motor, if you disassemble him it is destroyed. The motor runs, but have lost the torque.

JAZZCNC
02-12-2015, 12:24 PM
There's resonance and the Drives to think about here.!! Resonance from one motor will travel to the other sending the drives Bonkers. Each motor Drive will struggle to give stable performance because the Resonance compensation will be all over the place. Even the best Digital drives will struggle and Cheap drives with fixed Compensation will run horrible. Cheaper drives struggle with resonance from one motor so no chance with multiple motors and resonance cripples motor performance.

None starter to me.!

Rufe0
02-12-2015, 02:33 PM
RE: Neale
I disagree about there being a better way already. DC/BLDC/AC servos or any none stepper based servos will all have the same fundamental flaw compared to a stepper. Because they have very low 'steps' or none at all, they have to microstep, and microstepping is fundamentally bad for torque and it's not great for positional accuracy either. Effectively it is running the motor backwards and forwards at high frequency to achieve an in-between position, it's not hard to see the flaw there. This means even the best controllers in the ideal scenarios are going to be at best +-1step, more likely +-10steps and much worse under a changeable load. Any change in load even 0.000001nm is enough to move the motor slightly out of step.

Add in a gearing system with backlash and it will be all over the place under a changeable load or direction changes.

If you have a really high count encoder the problems are minimised but it's really only bandaging the problem.

However if you had a stepper with many phases you could use full steps and still achieve high resolution. You get none of the problems of microstepping, no oscillation problems and additionally you get a better torque curve than any single motor because the phases are running slower than the motor. For instance at 100rpm motor speed you get the torque that you would have got at 50rpm with 2x the phases. Consider 5 motors, 1000steps, 100rpm torque at 500rpm. You get dead on accuracy no +- nothing and it would be held at that position no matter the load(assuming you don't have too big a load). Clearly this would be advantageous.

Don't take my word for it lookup 5phase stepper motors, the benefits are clear and I believe it would only get better with more phases.

RE: JAZZCNC
This brings me to JAZZCNC point. The only real downside to this (other than building the thing in the first place) is needing a custom motor driver. Which is beyond most people yes but it's not impossible. Even a lowly £10 Arduino has the hardware capability to control many motors, more than you would need and the code has already been written, at least for 5phase anyway, it's pretty much copy paste to add more phases.

uli12us
02-12-2015, 02:54 PM
Thats not correct, you can hold a BLDC or Synchronservo with 2 different tensions between 2 fases in an angle that is somewhere between that you have with only one of the coils with tension.

With steppers you have full torque only at very low speed, and the speed is all but smooth.
With those speeds you cant really work, today with good tools you need a feedrate of more than 60ipm. Rapid feeds of 2500ipm are not seldom, no stepper can move so fast, at least, if you want a
resolution of 0,04thou, like in many industrial machines.

Rufe0
02-12-2015, 03:19 PM
Thats not correct, you can hold a BLDC or Synchronservo with 2 different tensions between 2 fases in an angle that is somewhere between that you have with only one of the coils with tension.Which is exactly what I said...


With steppers you have full torque only at very low speed, and the speed is all but smooth.
With those speeds you cant really work, today with good tools you need a feedrate of more than 60ipm. Rapid feeds of 2500ipm are not seldom, no stepper can move so fast, at least, if you want a
resolution of 0,04thou, like in many industrial machines.That is the whole point of a many phase stepper, it would be improved in all the areas you mentioned.

Neale
02-12-2015, 04:57 PM
I don't know the answer to this one, which is why I am asking. Rufe0 says that servos with encoders need to microstep to achieve high positional resolution. I assumed that the reason for things like 2500 line encoders was to achieve (approximately) the same level of resolution as, say, 12 cascaded 200-step steppers with rather less rotational inertia. Do servo systems microstep? I've never worked with servos so have no experience of how they and their drivers are set up. I do assume that there will be a small amount of rotational uncertainty as the servo hunts between two adjacent encoder lines, but I can't believe that this is any worse than the rotational uncertainty of a stepper under varying load as it only generates a restoring torque when displaced by some (very small) distance from the exact "step" position.

JAZZCNC
02-12-2015, 05:38 PM
RE: JAZZCNC
This brings me to JAZZCNC point. The only real downside to this (other than building the thing in the first place) is needing a custom motor driver. Which is beyond most people yes but it's not impossible. Even a lowly £10 Arduino has the hardware capability to control many motors, more than you would need and the code has already been written, at least for 5phase anyway, it's pretty much copy paste to add more phases.

Get on with it then and let us know how it works.!! . . . . Think you'll fine there's big difference between one drive running multiple individual motors separately than running multiple motors trying to be one single entity.!
Then there's also difference between being able to do it and doing it well.!! . . . . I wasn't saying wasn't possible just won't work well enough so none starter.!

Personally I would rather invest the time in something worth while and spend the money on decent Servo's if I needed more than a Stepper can offer.

Also your Stepping analogy of a Servo is kind of wrong. Yes while it's true that a Servo is NEVER truely in absolute Position and always searching for perfect Position there's no stepping going off.
The actual Positioning error of correctly tuned Servo drive/motor setup with resonable encoders is still FAR superior to what a Micro Step based system can provide with far superior torque/speed performance.

That doesn't mean steppers can't match a Servo on accurecy but only upto a point then Servo wins every time.

The Pulse requirements are much higher but still they are electrical pulses not physical detent steps like in stepper so don't suffer from resolution restriction like steppers have. There's only so many Micro steps one full Step can be broken down into before can't resolve anymore, or resolving more is pointless because resolution can't be achieved by the Stepping motor.
Provided the Encoder resolution is high and your Controller can provide the Pulses servo motors don't have this problem so resolution is much higher and even thou they never actually reach there intended destination they get very much closer than any Micro stepped motor can.!

Steppers do a great job and can be very accurate but only to a point then you need Servos and High count feedback for ultimate accurecy.
That's why High end systems will use Encoders and/or Scales for positional accurecy.

Rufe0
20-12-2015, 12:20 PM
Very helpful Jazz, I was going to write a scathing reply at the time but I've been too busy for the last few weeks. Hopefully I'll have chance to get a controller built over Christmas.

JAZZCNC
20-12-2015, 07:18 PM
Very helpful Jazz, I was going to write a scathing reply at the time but I've been too busy for the last few weeks. Hopefully I'll have chance to get a controller built over Christmas.

Don't know why you'd want to write Scathing reply because I wasn't having a go at you in anyway I honestly meant "get on with it and lets know" nothing more.
The fact we disagree shouldn't warrant a Scathing reply IMO.! If we all agreed on every thing Life would be very boring don't you think...:rolleyes: