DM542 Stepper driver

[Peter BellParticipant @peterbell11509]

I read the recent posts about TB6600 drivers with interest and why there were two 400pulse/rev settings.

On a machine I’m playing around with I’m using a DM542 driver with a Nema 17 motor and it runs cool, smooth etc.

My question.

The DM542 got all the usual multiple pulse/rev settings ie 800, 1600 etc but in addition it has lots of others, see pic (apologises, couldn’t rotate pic?)

These range from 1000 pulses/rev to 25000/rev—are they aimed for a specific use?

I always thought the basic 400, 800 range etc was possible as it used multiples of the stepper motor windings?

Finally, I’ve got a Nema 34 motor I want to drive can anyone recommend a suitable modern driver?

Thanks Peter

[Peter BellParticipant @peterbell11509]

[Martin ConnellyParticipant @martinconnelly55370]

I think the powers of 2 micro-stepping were used because it is an easy thing to do with off the shelf ICs. Plenty of binary dividers/multipliers to choose from. If you design a system from scratch it is not hard to use the same system as used for a VFD where any frequency can be used from the minimum setting to the maximum setting of the VFD. Just set up suitable values in the driver and the switches to select them.

I think any modern driver with the correct voltage and current capacity will drive your stepper motor.

Martin C

[BazyleParticipant @bazyle]

The high counts are somewhat ephemeral. Stiction in the motor let alone the attached mechanism mean you won't be getting it to move one of those steps at a time, probably not even ten. With a dedicated chip it is a throwaway feature to add and if it suckers in a few extra sales all the better.
Since it is just varying the currents in two windings with a DtoA converter or PWM equivalent the movements are not linear (yeah rotation but you know what I mean) above about 800.

[John HaineParticipant @johnhaine32865]

Microstepping works by adjusting the currents by a sine/cosine weighting to make the rotor take up intermediate positions between the "full step" settings. There's no particular reason why the binary weights have to be used – for some purposes something divisible by 3 would be nice! On a couple of TI driver chips you can feed in analogue voltages from an external D/A to drive the microstepping. But the higher counts are more for smoothness than precision.

Your DM542 should work fine with a NEMA 34 motor if you set the current right.

Edited By John Haine on 16/01/2023 17:20:55

[SillyOldDufferModerator @sillyoldduffer]

Be interesting to measure how accurately a motor steps at 25,000 pulses per revolution. If I understand how they work properly, they might do better than Bazyle's 800 accurate steps per revolution. At least if the load is light.

But my understanding is weak!  I visualise a permanent magnet rotor positioning itself at the average point of magnetism created by two pulse trains in the stator coils, where the average depends on the relative phase of the two pulse trains. As they are generated electronically, their phasing could be miles apart or very close together. My feeling is big and small steps can both be made accurately. But there must be a limit!

What is certain is driving the motor at 25000 pulses per revolution would turn it very slowly. The DM542 datasheet says it has a maximum pulse input of 200kHz, so max speed is 8rpm flat out. At 800 steps per revolution the motor will do 250rpm. Both estimates assume that the input controller can put 200kHz into the DM542, which is a tall order, and that the motor's power supply can respond to high speed high current demands. For this an unregulated supply is better than a fancy regulated one.

Whether 8rpm is good or bad news depends on the application. Stepper motors are popular for driving rotary tables because they accurately step the table through a worm drive giving even more accuracy. 25000 pulses per revolution is bad news for this: As my HV6 clone has a 90:1 worm, an 8rpm motor would take over 11 minutes to turn it once. Life is too short…

Dave

 

 

 

Edited By SillyOldDuffer on 16/01/2023 18:57:17

[Andy_GParticipant @andy_g]

Posted by SillyOldDuffer on 16/01/2023 18:54:27:

The DM542 datasheet says it has a maximum pulse input of 200kHz, so max speed is 8rpm flat out. At 800 steps per revolution the motor will do 250rpm.

SOD's maths are out by a factor of 60

(200kHz at 25k ppr is 480 RPM)

Higher microstepping divisors can help with smoothness and noise reduction, and might have some use in getting smooth rotation at very slow speeds, but I also think that the highest numbers are overkill.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Andy_G on 17/01/2023 07:57:04:

.

SOD's maths are out by a factor of 60

(200kHz at 25k ppr is 480 RPM)

.

__ Would you trust this man to design a precision clock ?

MichaelG.

[Peter BellParticipant @peterbell11509]

Many thanks for all the replies. Thought there was something basic I'd missed.

Driving it from an Arduino Nano and ended up using 2000ppr as its very smooth and working out at the right speed..

Peter

[HowiParticipant @howi]

Posted by Michael Gilligan on 17/01/2023 08:07:29:

Posted by Andy_G on 17/01/2023 07:57:04:

.

SOD's maths are out by a factor of 60

(200kHz at 25k ppr is 480 RPM)

.

__ Would you trust this man to design a precision clock ?

MichaelG.

Only time will tell

[SillyOldDufferModerator @sillyoldduffer]

Posted by Michael Gilligan on 17/01/2023 08:07:29:

Posted by Andy_G on 17/01/2023 07:57:04:

.

SOD's maths are out by a factor of 60

(200kHz at 25k ppr is 480 RPM)

.

__ Would you trust this man to design a precision clock ?

MichaelG.

Judging by clock results so far, NO!!!

Forgetting the difference between seconds and minutes is typical of my love-hate relationship with maths. I know how to do the sum, and then miss a simple step.

At least I show the working. One of the best things about the forum for me and everyone else, is my all too obvious mistakes can be spotted and corrected!

It's why I believe engineering is a game best played by teams! Even so it's embarrassing: sackcloth and ashes AGAIN…

Dave

[Martin ConnellyParticipant @martinconnelly55370]

Dave, do what I do and blame it on autocorrect. You put RPS and the stupid autocorrect changed it to RPM

Martin C

[Michael GilliganParticipant @michaelgilligan61133]

Posted by SillyOldDuffer on 17/01/2023 12:03:24:

Posted by Michael Gilligan on 17/01/2023 08:07:29:

__ Would you trust this man to design a precision clock ?

MichaelG.

Judging by clock results so far, NO!!!

.

A very honourable response, Dave … but please don’t go falling on your sword

You’re doing great work and we’re cheering you on !

MichaelG.

[BazyleParticipant @bazyle]

Anyway this is a mechanical engineering forum (mostly). We should only countenance full stepping generated by a series of cams and microswitches and then make a gearbox to provide the finer resolution

[Author]

Posts