PDA

View Full Version : Micro steps



erniehatt
31-05-2021, 05:25 AM
Hi All,
Is the an ultimate setting for the micro steps of the motor.

A_Camera
31-05-2021, 07:10 AM
Depends on many things. There is no "one size fits all" answer. The less is the microstep the higher the torque. The higher the microstep the smoother the rotation.

erniehatt
31-05-2021, 08:04 AM
Depends on many things. There is no "one size fits all" answer. The less is the microstep the higher the torque. The higher the microstep the smoother the rotation.

Thanks for that, so best idea would be to try a few different settings to see which works best.

Neale
31-05-2021, 08:08 AM
At a purely mechanical level, as A_Camera says, micro-stepping can give smoother movement. No micro-stepping and a stepper sounds like marbles in a cement mixer. However, because MS uses the balancing of currents between sets of coils in the motor rather than using full current in one coil (full-step positions), the torque available is reduced. In practice, some MS is used and the general recommendation is somewhere around x8-x10, giving 1600-1800 positions per revolution. Note that this is done for smoother movement and does not necessarily give you a corresponding improvement in position accuracy as the MS positions aren't quite as well-defined as full step positions.

There's another factor, though, and that's your motion control system. Say you are aiming at a max stepper speed of 600rpm (that's a bit slow generally but makes the sums easier!) So, 600rpm is 10 revs/sec. Full stepping means you need 180x10=1800 pulses per sec, well within the capabilities of even a parallel-port system. At x8 MS, that becomes 8x180x10pps, 14400pps, which is starting to push things a bit. At a realistic 1000rpm, that's 24Kpps, which is probably beyond a parallel-port system. That's one reason why people use external motion controllers these days - can deliver higher pulse rates to match realistic machine speeds and the more accurate pulse timing also helps smoother motion.

erniehatt
31-05-2021, 10:03 AM
At a purely mechanical level, as A_Camera says, micro-stepping can give smoother movement. No micro-stepping and a stepper sounds like marbles in a cement mixer. However, because MS uses the balancing of currents between sets of coils in the motor rather than using full current in one coil (full-step positions), the torque available is reduced. In practice, some MS is used and the general recommendation is somewhere around x8-x10, giving 1600-1800 positions per revolution. Note that this is done for smoother movement and does not necessarily give you a corresponding improvement in position accuracy as the MS positions aren't quite as well-defined as full step positions.

There's another factor, though, and that's your motion control system. Say you are aiming at a max stepper speed of 600rpm (that's a bit slow generally but makes the sums easier!) So, 600rpm is 10 revs/sec. Full stepping means you need 180x10=1800 pulses per sec, well within the capabilities of even a parallel-port system. At x8 MS, that becomes 8x180x10pps, 14400pps, which is starting to push things a bit. At a realistic 1000rpm, that's 24Kpps, which is probably beyond a parallel-port system. That's one reason why people use external motion controllers these days - can deliver higher pulse rates to match realistic machine speeds and the more accurate pulse timing also helps smoother motion.

Thank you, By external motion controllers, do you mean external drivers like DM 542 's

Neale
31-05-2021, 10:50 AM
Three basic bits go together. The software, usually on a PC, something like UCCNC or Mach3, provides a user interface so you can control things, and reads the gcode file that effectively says something like "go from this XY position to that XY position." Then there's a motion control stage that converts those coordinate moves into something your machine understands - it turns XY moves into the right number of pulses to drive your machine the right way, interleaving X and Y pulses ("steps") to get smooth movement. This bit used to be done by Mach3, for example, but nowadays is usually done by the external motion controller - something like a UC400 or AXBB device. That produces step pulses and direction signals that go to the stepper driver which, in effect, amplifies them to high-power signals that go to the motors to drive them. The DM542 is a stepper driver - the last bit in the chain before the motor itself.

All a bit of a simplification but that's the gist of it. Takes a little while to get you head round first time through the process.

erniehatt
31-05-2021, 11:42 PM
Three basic bits go together. The software, usually on a PC, something like UCCNC or Mach3, provides a user interface so you can control things, and reads the gcode file that effectively says something like "go from this XY position to that XY position." Then there's a motion control stage that converts those coordinate moves into something your machine understands - it turns XY moves into the right number of pulses to drive your machine the right way, interleaving X and Y pulses ("steps") to get smooth movement. This bit used to be done by Mach3, for example, but nowadays is usually done by the external motion controller - something like a UC400 or AXBB device. That produces step pulses and direction signals that go to the stepper driver which, in effect, amplifies them to high-power signals that go to the motors to drive them. The DM542 is a stepper driver - the last bit in the chain before the motor itself.

All a bit of a simplification but that's the gist of it. Takes a little while to get you head round first time through the process.

Thanks Neale, I do know how it works, just not sure of the terminations, What you called the motion controller I call the motor controller.
I do have all the parts need, NVEM V2 6axis controller, Nema23 4nm motors and DM 542 drivers, just hoping that these will do the trick for me.
I am not to bad on the mechanical side of things, electronically a little sus.

Neale
01-06-2021, 01:31 PM
Tricky one, answering questions - you never know what the other person knows already. But given that these forums are also read by people looking to learn, I tend to err on the side of giving a bit too much rather than too little info. So, nothing personal intended!

"motor controller" is definitely wrong, though, as this implies something that directly connects to a motor which the motion controller does not. Might also be interpreted as something like a DC spindle motor controller which is a very different animal. Good luck with the project - lots of fun!

dazp1976
01-06-2021, 06:48 PM
Work out what kind of resolution you want your machine to have.
For instance if you want it at 1000th of a mm.....(I use metric)
So that's 0.001, so 1000 microsteps is 1mm,
so 1000 is your rev per unit,
if it's 5mm pitch screws then multiply that 1000 by 5,
driver steps per rev is 5000.
This would then need at least 42khz to be able to reach 5000mm/min rapids.


P.S. I hope you are one of the lucky few that gets a MVEM working. There are a lot of bad experiences out there.

JAZZCNC
01-06-2021, 07:42 PM
Work out what kind of resolution you want your machine to have.
For instance if you want it at 1000th of a mm.....(I use metric)
So that's 0.001, so 1000 microsteps is 1mm,
so 1000 is your rev per unit,
if it's 5mm pitch screws then multiply that 1000 by 5,
driver steps per rev is 5000.
This would then need at least 42khz to be able to reach 5000mm/min rapids.


P.S. I hope you are one of the lucky few that gets a MVEM working. There are a lot of bad experiences out there.

That's not a good way to think of it and poor advice.! . ... Micro steps shouldn't be thought of or used to gain resolution, that's the purpose of the pitch of lead or ball -screws or whatever method you use for linear motion and any ratios which are applied.
Most steppers can't resolve resolution much past 2000ms so anything after that is wasted in terms of resolution gains and then all you are doing is losing torque and stressing the motion controller, yes you gain a little smoother motion from the steppers but still not worth it and the high pitch noise is horrible.

No stepper really needs to be run past 2000ms and if small motors I'd recommend you don't go past 800 to keep all the torque you have, my preference is 1600ms for 3-4nm nema 23/4 motors but this varies depending on motor and drive combo's and how well they handle resonances, etc.

erniehatt
01-06-2021, 11:38 PM
Tricky one, answering questions - you never know what the other person knows already. But given that these forums are also read by people looking to learn, I tend to err on the side of giving a bit too much rather than too little info. So, nothing personal intended!

"motor controller" is definitely wrong, though, as this implies something that directly connects to a motor which the motion controller does not. Might also be interpreted as something like a DC spindle motor controller which is a very different animal. Good luck with the project - lots of fun!

no problem Neale, better to much info than not enough, thaks for the imput.

erniehatt
01-06-2021, 11:44 PM
That's not a good way to think of it and poor advice.! . ... Micro steps shouldn't be thought of or used to gain resolution, that's the purpose of the pitch of lead or ball -screws or whatever method you use for linear motion and any ratios which are applied.
Most steppers can't resolve resolution much past 2000ms so anything after that is wasted in terms of resolution gains and then all you are doing is losing torque and stressing the motion controller, yes you gain a little smoother motion from the steppers but still not worth it and the high pitch noise is horrible.

No stepper really needs to be run past 2000ms and if small motors I'd recommend you don't go past 800 to keep all the torque you have, my preference is 1600ms for 3-4nm nema 23/4 motors but this varies depending on motor and drive combo's and how well they handle resonances, etc.

Thanks Jazzcnc, your imput is much appreciated I give it a try at 1600, when it is all completed that is.:subdued:

erniehatt
01-06-2021, 11:49 PM
P.S. I hope you are one of the lucky few that gets a MVEM working. There are a lot of bad experiences out there.

I have heard that some have problems with it, but also some successes, but I have also heard of failures with other controllers, so maybe I will get lucky.:smile:

dazp1976
02-06-2021, 12:40 AM
That's not a good way to think of it and poor advice.! . ... Micro steps shouldn't be thought of or used to gain resolution, that's the purpose of the pitch of lead or ball -screws or whatever method you use for linear motion and any ratios which are applied.
Most steppers can't resolve resolution much past 2000ms so anything after that is wasted in terms of resolution gains and then all you are doing is losing torque and stressing the motion controller, yes you gain a little smoother motion from the steppers but still not worth it and the high pitch noise is horrible.

No stepper really needs to be run past 2000ms and if small motors I'd recommend you don't go past 800 to keep all the torque you have, my preference is 1600ms for 3-4nm nema 23/4 motors but this varies depending on motor and drive combo's and how well they handle resonances, etc.

I was only kind of talking the theory behind it.
In terms of mills with steppers I've never gone above 3000 / 600. With steppers it's a bit pointless tbf.
Atm mine is at 2500 / 500 / 5mm. (I like to keep it ll in5's lol).

When it comes to proper servo's however.