Considerations stepper motor for speed

[Litterio Andrea Guainella]

Hi Mark,

as you mention in this post

> In PnP you cannot usually consume that much current, because at high speeds the back-EMF is too high. You'd need low-inductance steppers or higher voltages. Be sure to use the full 36V!

Your article here is most clarity about important motor voltage but I think that argument should be discussed more thoroughly because me and who like me is not an expert is still confused.

Maybe use 60VDC with right driver can be a best match?

For example this motor with this driver could improve speed and smooth motion? (compared for example with liteplacer stock version)

What do you suggest alternatively?

LAG

[mark maker]

Hi LAG,

I guess you're on the right track. But I have no practical experience, I just know for a fact that the voltage is the limiting factor, and that few people realize that, so I wanted to say that, but I cannot provide a proven solution.

As for the Liteplacer and similar DIY designs, another question is the stiffness and balance of the machine. There is probably a point where increasing the motor power (voltage) is too much. Like here 😁

https://youtu.be/FUh36RUHzdU

On my Liteplacer, even with conventional 24V steppers, I already have very long camera settle times, because the head is so badly balanced, and the frame so weak, it vibrates like hell.

Going for 36V (that the Rapid board supports) sounds like a reasonable compromise.

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/abb15ce8-463e-4219-bbc7-6d52eff16716n%40googlegroups.com.

[Ian Arkver]

3-phase steppers are better at high speed than 2-phase ones since the torque curve doesn't roll off at high step rates as much. Also resonance problems at high step rates are lower. This what I've read anyway - I've not tried 2-phase to compare.

Also - only 3 high-current wires needed.

I went for 3-phase steppers and matching controllers from Lichuan (I think) with a 36V supply and I'm happy with the machine speed. As Mark says controlling vibration gets to be more important the faster you go.

Regards,
Ian

[Litterio Andrea Guainella]

Yeah I know odrive but I don't think anyone actually used it (maybe because liteplace software run in single thread as mention here).

As far as the rigidity of the system is concerned, I think that sooner or later I will design my own version of pnp :-)

In the meantime I will try as much as possible to improve the performance of my liteplacer.

@Ian Thank you for your suggestion. Do you have a video out of curiosity?

> As Mark says controlling vibration gets to be more important the faster you go.

Yes indeed professional solutions have gigantic machines probably because the ballast is important to overcome the obstacle of vibrations

LAG

[Arthur Wolf]

On Mon, Jul 4, 2022 at 8:32 PM Ian Arkver <ian.ark...@gmail.com> wrote:

3-phase steppers are better at high speed than 2-phase ones since the torque curve doesn't roll off at high step rates as much. Also resonance problems at high step rates are lower. This what I've read anyway - I've not tried 2-phase to compare.

I've compared both: leadshine setup, same current (3A), 2-phase first, then moved to 3-phase, and it was night and day, allowed higher speeds without stalls, and much less resonance (very similar use case to PnP, glue dispensing, so high-speed/low-mass)

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ddc56625-eeea-46bc-9c4b-66b35b570086n%40googlegroups.com.

--

勇気とユーモア

[Ian Arkver]

On Monday, July 4, 2022 at 8:18:51 PM UTC+1 Arthur Wolf wrote:

On Mon, Jul 4, 2022 at 8:32 PM Ian Arkver wrote:

3-phase steppers are better at high speed than 2-phase ones since the torque curve doesn't roll off at high step rates as much. Also resonance problems at high step rates are lower. This what I've read anyway - I've not tried 2-phase to compare.

I've compared both: leadshine setup, same current (3A), 2-phase first, then moved to 3-phase, and it was night and day, allowed higher speeds without stalls, and much less resonance (very similar use case to PnP, glue dispensing, so high-speed/low-mass)

Thanks Arthur, that's good to know. The price for 3-phase kit was only about 10% more and most of the cost was in the shipping, so it was an easy choice.

One more benefit - 1.2 degree full-step angle.

@LAG - Sorry, not shot a video - no time for PnP  :-(

 

Also - only 3 high-current wires needed.

I went for 3-phase steppers and matching controllers from Lichuan (I think) with a 36V supply and I'm happy with the machine speed. As Mark says controlling vibration gets to be more important the faster you go.

Regards,
Ian

[Litterio Andrea Guainella]

Sounds good for 3-phase motor.

What do you think of this couple?

Leadshine HBS507 + 573HBM20

So for recap:

  - low inductance (only 1.85mH)

  - 2.0 N.m (as well as sufficient)

  - 48VDC as power supply motor

  - closed loop for best result (no lose step)

  - 3-phase for improvement performance and smooth movement

Maybe sound like be better

LAG

[Ian Arkver]

That's a nice set. I suspect you'd need a really solid frame to get anywhere near 2Nm torque used without shaking it all over the place.

I think 2Nm is fine for the Y axis, but you might want a lighter 1Nm one for the X axis since that motor's mass will be significant in Y axis moves. (Assuming the regular axis naming - I'm not familiar with Liteplacer).

The Lichuan ones I got were similar I think but didn't have the D9 connector on the controller, so changing the encoder wiring was a little easier.

https://www.aliexpress.com/item/32452247259.html

Closed loop for the win.

Regards,
Ian

[Saito Ryosuke]

Hi.

I have applied the Lichuan LCDA357H+LC57H380 combination to OpenPnP.
This controller accumulates positional errors during operation.
The effect is serious: the coordinates are shifted (to undesirable positions) during placement.

I also like closed loop servos.
However, I would recommend choosing an open loop stepper motor over using this controller.

Ryosuke

2022年7月5日(火) 17:08 Ian Arkver <ian.ark...@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a5443ae0-2355-419d-b656-f1fb533ccf79n%40googlegroups.com.

[Litterio Andrea Guainella]

Hi Ryosuke,

damn! A closed loop driver is usually chosen for its reliability compared to a traditional stepper.

Maybe it was just faulty?

I don't know how to interpret openpnp graphics but it doesn't look good.

Maybe someone knows how to interpret them?

@Ian  you who have used them have you noticed this defect?

LAG

[Ian Arkver]

No, I haven't noticed this, but I haven't done enough long term positionally accurate tests to see an accumulation. I'm also not familiar with those OpenPnP graphs or how to interpret them. I just ran the backlash compensation from I&S and it came up with a number (happily a small one). I didn't inspect the results.

It's worrying though, and makes me want to do a bunch more testing. Someday I'll find the time for that I hope.

These controllers have a 200kHz max step rate on the opto inputs, iirc. Could you be exceeding that Ryosuke? Are you 100% sure that your controller is delivering the right number of pulses?

Regards,
Ian

[mark maker]

Hi Ryosuke,

Haven't we traced that back to faulty Units per Pixel or Steps per Millimeter config? Or was this another user?

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/eed04ba2-488e-4a77-892d-5e4f45ed72aen%40googlegroups.com.

[dc42]

FWIW we now have a Duet3D CAN-connected controller board for stepper motors with built-in encoders. See https://docs.duet3d.com/Duet3D_hardware/Duet_3_family/Duet_3_Expansion_1HCL. It can be used with 24V to 48V power and up to 6A peak motor current.

The reproducibility of position achievable with this setup is limited by the resolution of the encoder. A typical encoder has 1000 pulses per resolution, which is 4000 transitions per resolution. So in theory we can resolve position to 1/20 of a full step using a 1.8deg stepper motor. In practice the 4 counts per pulse may not be spaced uniformly, so it is better to assume resolution to about 1/10 full step. Using typical belt driven XY axes with 5 full steps/mm, this gives a theoretical resolution of 0.02mm.

[Jarosław Karwik]

Well, if you had machine with line encoders for X/Y - yes. Actually 3 line encoders  if you have gate setup ( for both sides of the gate) 

Most machines do have encoders only on the motor shaft, so the final accuracy is not that good. On the top of that you need hysteresis, otherwise the system would be unstable.

Unless you have steel frame ( or granite/polymer concrete ) you would have to multiply you theoretical resolution by 10x to be around final accuracy. 

It would still not be bad however.

[bert shivaan]

Hey dc42, is it possible to get firmware for the regular CAN tool board to be more useful in the openPNP world? To do this it would be nice to be able to control 3 motors instead of 1. Plus at least 6 outputs for micro valves 24vdc(low side switching is fine).Some folks would like to have 2 vacuum sensors, so maybe 2 analog inputs. 

Now if needed things are not built in thats not an issue, like aof course there is only 1 stepper driver. But if we could wire an expansion board to it that would give 2 more drivers. Fet boards can be made no problem if the I/O are there.

Anyway, is this something that could maybe happen? It would be so super great for a PNP machine.

Thanks,

Bert

Dllpdf.com

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/4f1a868e-4096-4c56-9649-b365a5e04e01n%40googlegroups.com.

[dc42]

@cncmachineguy it would certainly be possible to make such a board. The easiest way to prototype it would be to use a SAMMYC21 dev board (see https://docs.duet3d.com/en/Duet3D_hardware/Duet_3_family/Using_the_Sammy-C21_development_board_with_Duet_3) and connect it to three TMC2209 stepsticks and the mosfets on some stripboard or perfboard. We have a firmware build configuration for the SAMMYC21 on Github. You would only need to change a few tables and constants to support the additional drivers and mosfets.

Another way would be to hack a regular tool board by connecting two TMS2209 stepsticks and the additional mosfets. You would need to replace a few of the 0402 input protection resistors to allow some of the inputs on the tool board to be used as outputs instead.

[bert shivaan]

Thanks dc42, My hope was to be able to use the off the shelf tool board with the expansion board that fits on the mini5. that would give us the 3 motors.

I have no idea how to write the firmware for the tool board. I think there are folks here that could use this, but not enough to warrant a new/special board. Fet boards are also a plenty in the arduino community.

Is this something Duet would be interested in helping with (the new firmware)?

Bert

Dllpdf.com

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e695be5c-1631-4c28-9a71-09a5214af4f8n%40googlegroups.com.

[bert shivaan]

Hi James, They sound interesting for sure, where can I find more information?

[dc42]

@cncmachineguy I hadn't considered using the 2-driver expansion board for the Duet 3 Mini in conjunction with the tool board.

To use this arrangement we would need to find 4 spare MCU output pins for the additional step and direction signals (assuming that you don't need stall detection) and another 3 to drive the 3 additional mosfets; so 7 output pins in total. There are a number of MCU pins on the tool board that could be repurposed:

- 2 pins are unused and broken out to test pads

- 1 pin is available as a general purpose output, typically used to drive a BLTouch or other Z probe in 3D printer applications

- The input pin for one of the two buttons on the board could be repurposed, if we also made a minor change to the bootloader

- The two temperature sensor inputs could be repurposed as outputs, if some components on the tool board were removed and the values of others changed

- The two fan tacho inputs could be repurposed as outputs, if some components on the tool board were removed and the values of others changed

- Three of the four inputs could be repurposed as outputs, assuming they are not needed to support endstop switches or Z probes, if some components on the tool board were removed and the values of others changed

So there are enough MCU pins and it would be possible to hack a tool board to do this. The shared signals for driving stepper drivers are broken out to test pads, so already available.

If you or anyone else wants to try this, I can help with the firmware modifications.

[bert shivaan]

I would love to try this. I think this would be a GREAT asset for the openPNP community. I have plenty of tool boards in stock, so happy to try anything you ask hardware wise.

Will the tool board reprogram from the Duet over CAN?

Stall detection would be nice, but we can only use it for Z, so that will be on the tool board driver. The expansion board drivers will drive the 2 C axis.

If we use stall detection for Z homing, no end stops are needed for homing.

so bare minimum I think is this:

4 output for step/dir 2 axis

2 output for vacuum valve -required

2 output for blow valve - not required but VERY nice to have

1 output for camera light control - required

2 analog inputs for vacuum sensor - not required but people like them

The removal of components is not too bad to do, not sure about the changing of others?

Oh, maybe we need a new topic? we are hijacking this one pretty good.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/eff463bf-7844-478d-af40-5570c52785e7n%40googlegroups.com.

[bert shivaan]

also 1 more question, as a reseller of Duet, aam I allowed to make these changes and sell them this way? I know the factory warranty will be void, but I don't want to violate my agreement. Not sure it would but just checking. I can always just provide the "process" and let the buyer make the changes. I am sure most everyone here is able to do this, afterall we are building PNP machines :)

[bert shivaan]

Of course we can't forget the outputs we already have there. the 2 fans and the heater.

[dc42]

> Will the tool board reprogram from the Duet over CAN?

Yes.

>> Stall detection would be nice, but we can only use it for Z, so that will be on the tool board driver. The expansion board drivers will drive the 2 C axis.

If we use stall detection for Z homing, no end stops are needed for homing.

so bare minimum I think is this:

4 output for step/dir 2 axis

2 output for vacuum valve -required

2 output for blow valve - not required but VERY nice to have

1 output for camera light control - required

<<

So that's 5 x 24V outputs, of which the tool board already has 3. Are you sure you don't want an X axis endstop on the tool head?

I assume that you might want to drive the two C axes at the same time. If not then I think we could use a shared DIR pin.

>>

2 analog inputs for vacuum sensor - not required but people like them

<<

I suggest we repurpose the temperature sensor inputs for those.

>>

The removal of components is not too bad to do, not sure about the changing of others?

<<

The values that would need changing are resistors whose values need to be reduced. So the easiest option is probably to solder another resistor on top of the existing one. These components are mostly 0402 so you need a fine-tipped soldering iron.

>>

also 1 more question, as a reseller of Duet, aam I allowed to make these changes and sell them this way? I know the factory warranty will be void, but I don't want to violate my agreement. Not sure it would but just checking. I can always just provide the "process" and let the buyer make the changes. I am sure most everyone here is able to do this, afterall we are building PNP machines :)

<<

I don't see any problem, as long as the buyer is aware that the factory warranty is voided.

Feel free to start a new topic to discuss this further.

[bert shivaan]

new thread:

https://groups.google.com/g/openpnp/c/t0eAAJYMZzI

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/3401a4f2-c902-4a05-b849-6c556da32aaan%40googlegroups.com.