steppers with lots of Amp... + M906 command

A Former User · 14 Sept 2018, 16:04

Hi,
I am currently comparing datasheets and calculating lot´s of stuff for the right steppers for a custom buildup. Since the printer will be a big one (ca. 700x700mm printbed -> 1000x1000 size) I have quite some weight to move and combined with the right ratio and size of steppers I have currently for the X-/Y-Axis steppers (nidec KA60) in mind that are rated with 3.3A and that I would like to run with ca. 2.8A-3.0A by the trinamic chip...

Question: Is the trinamic chip stable over hours if not days with so much amp? Would I Need a heatsink + active-cooling on the chips?
Question: Would the command simply be "M906 X:3000, Y:3000, ..."? Or is there any safety-feature that I have to disable to draw more then the 2.4A from the driverchips?
Best, Lucas

wilriker · 14 Sept 2018, 17:34

@lb You cannot run the drivers higher than 2.4A. The firmware just won't accept higher values.
But at around 2A you definitely need active cooling. A fan cooling the back side of the Duet will be sufficient. No additional heat sinks required if the fan blows cool (room temperature) air.

dc42 · 14 Sept 2018, 20:22

I use 2.8A NEMA 23 motors for testing the Duets at high current. According to the data sheet, 2.4A is about the maximum safe standstill current. If we could guarantee to reduce the current at standstill, we could go higher. Unfortunately the TMC2660 doesn't do standstill current reduction in hardware. Doing it in firmware in a way that is responsive enough when a motor starts moving and guarantees to reduce the current within a few tens of milliseconds of standstill no matter what else is going on is difficult.

A Former User · 17 Sept 2018, 11:03

Thanks so much for your fast and good answers! Will dig into a motor that is around 70-90% with 2.4A then...
Planning with heatsinks (those watterott sells for their trinamic-driver although I know they are put on the "underside" of the chip -> can heatsinks be put on the bottomside or should I put them on top of the chip?) on the chips and a fan from below (found that is already explained very nicely on the setup-page shame on me )

-> Ah, that KA60 with 3.3A would have run so fast and energy efficient from my calculations, now search for which "stepper got talent" starts again

wilriker · 17 Sept 2018, 11:10

@lb What kinematics do you build? CoreXY where the bed is moved "just" up and down? Or Cartesian with the bed moving in Y? You said that there is a lot of weight to move but I think you underestimate what a high torque NEMA17 can do.

@deckingman is moving a bed of about 8kg total weight up and down with a single NEMA17 with 59Ncm. Given, its top speed is rather low at 7.5mm/s but in Z you don't need high speeds anyway.

Also I just recently bought Wantai 42BYGHW811. They "only" have 48Ncm and I also run them at just 52% but one of them will throw my about 1kg bed around at 250mm/s if it was made from Styrofoam instead.

deckingman · 17 Sept 2018, 11:21

@wilriker Actually it's acceleration that is limited by the available motor torque. So for a really long Z move, I could attain a higher speed. It doesn't take much force to keep a mass moving at a constant speed, but it takes a hell of a lot more force to accelerate that mass.

A Former User · 17 Sept 2018, 11:21

@dc42
O.K. didn´t know that.
The webstie (https://www.trinamic.com/products/integrated-circuits/details/tmc2660-pa/) claims it features some sort of auto-adjust of the current ("coolStep
load dependent current control for energy savings
up to 75%")... looking at page 12 of the PDF (https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2660_datasheet.pdf) it is like you said, the more you want "torque reserve" in a high-responsive environment, the less you can benefit from "cool-step"...:

-> Seems like you already played with it. Could it be made accessible on the duet (In some sort of "developer"-mode)?

-> Another question it says you can drive up to 2.8A (with ext. mosfets up to 4A)? If one could tune the cool-step on the chip could we "unlock" those 2.8A with a proper cooling-fan & heatsinks (those biggies here e.g. https://www.watterott.com/de/SilentStepStick-Kuehlkoerper-14x15x13) on the drivers?

wilriker · 17 Sept 2018, 11:29

@lb coolStep is already available via M915's Tnnn parameter. But it fully depends on a fine-tuned stallGuard (that will be configured with the same command).

A Former User · 17 Sept 2018, 11:30

@wilriker
Hi, thanks for the fast reply. Will be similar to a ultimaker split axis design - only with supported outer rails - (x and y independent with their own motor and belt-synchronisation of the 2 sides): so X & Y are equal only flipped and rotated 90°, the moving mass of each (X or Y) will be ca. 5KG (3.xKG the linear-elements, belts, pulley-wheels, axis, bushings, etc/ 1.x the direct-drive printhead with hotend, fans, etc.)

-> only 1 motor to optimize for 2 axis

Yeah, guess I will focus on ac-/de-celeration and not on the high-end-speed after that -> isn´t really needed on a printer since the "position" moves are not so many like on a cnc-mill or similar

For Z (10mm aluminiumplate ca. 10KG + 5KG rest) there is a big 2.8A 2.xNM 0.9° oriental stepper on 2*TR16x2 (belt-synchronised) moving it up and down -> that one for sure does not need any speed as far as I plan in the moment. If there should be any problems on the z-axis I could still play with ratio of the pulleywheels...

By the way this is not for me (would not have the money for that) but for another person...

A Former User · 17 Sept 2018, 11:32

@wilriker
Wow thanks!

wilriker · 17 Sept 2018, 11:34

@lb said in steppers with lots of Amp... + M906 command:

Yeah, guess I will focus on ac-/de-celeration and not on the high-end-speed after that

In this case this might be handy.

A Former User · 17 Sept 2018, 11:34

@deckingman
Thanks so much!
Yes right. Will focus on more torque in the slow-zone then, won´t need the high abs speed anyway

A Former User · 17 Sept 2018, 11:48

@wilriker
Thanks, only found the download calc-sheet, but haven´t seen the online calculator.

What it misses for my setup to be of a value:

-> Reduction ratio that I seem to have to use on this setup (since reduction is going with "/ ratio^2" for inertia but only "/ratio" in speed very important in my understanding)
https://pdfs.semanticscholar.org/f1d3/f1eee2a1a3380eb5cdbc3a605a21c6058bfd.pdf
& with a formula for belt-driven-systems:
https://www.motioncontroltips.com/faq-how-do-i-calculate-the-inertia-of-a-servo-driven-system/

-> Inertia of motor-shaft-clamped pulley-wheel or whatsoever on motor should be added-able to the motor-inertia

-> max speed vs back-emf
https://duet3d.dozuki.com/Wiki/Choosing_and_connecting_stepper_motors#Section_Back_emf_due_to_rotation

-> positional error vs. rotation speed based on current-settling (determined by (m)H)
https://www.orientalmotor.com/stepper-motors/technology/stepper-motor-basics.html

-> efficient over lifetime by matching drive-supply-Vdc for given motor-H
https://softsolder.com/2011/05/05/thing-o-matic-mbi-stepper-motor-analysis/

But maybe I shouldn´t be so worried about all that

wilriker · 17 Sept 2018, 11:57

@lb This calculator is solely about acceleration.
For speed vs. back-EMF you can use the EMF calculator at https://reprapfirmware.org.

You could approximate the reduction ratio by reducing the axis mass that you enter into the calculator. But I will think about adding gear ratio to the calculator also.

Can you give me a formula how to calculate the inertia of a drive pulley? I know how to do it for the rotor inertia based on the value given in the datasheet. If this is based on the pulley's diameter this could even be added easily without another input field.

A Former User · 17 Sept 2018, 11:58

@wilriker
edited the "mess" above, please have a look at the "old" comment again

wilriker · 17 Sept 2018, 11:59

@lb You somehow managed to convert all links (except the first) to be images of the text of the link destination. I would not even know how to do that if I wanted to.

A Former User · 17 Sept 2018, 12:10

@wilriker
Oops - hopefully fixed now - please check again! (Learned: Do not copy what is in field of openoffice-calc by just clicking on the field and (on windows) press ctrl+c but rather select what is in that field first and then press ctrl+c...)

wilriker · 17 Sept 2018, 12:13

@lb said in steppers with lots of Amp... + M906 command:

What it misses for my setup to be of a value:

-> Reduction ratio that I seem to have to use on this setup (since reduction is going with "/ ratio^2" for inertia but only "/ratio" in speed very important in my understanding)
https://pdfs.semanticscholar.org/f1d3/f1eee2a1a3380eb5cdbc3a605a21c6058bfd.pdf
& with a formula for belt-driven-systems:
https://www.motioncontroltips.com/faq-how-do-i-calculate-the-inertia-of-a-servo-driven-system/

-> Inertia of motor-shaft-clamped pulley-wheel or whatsoever on motor should be added-able to the motor-inertia

I just looked at the second links with the formulas. I don't think that this matters too much. Because usually the mass of belt and pulley are negligible compared to all other masses in play. I don't think it makes sense to add 10-50g to even 500g axis (although it lowers the acceleration a bit). It won't have a large effect but makes the number of input fields in the calculator a lot more.

But maybe I shouldn´t be so worried about all that

I have a feeling that you are dead right with this assumption.

P.S.: I think I will add gear ratio to the Advanced Settings.

A Former User · 17 Sept 2018, 12:26

@wilriker
cool! Looking forward to it

Can anybody help me for my "open-office-calc"-sheet for the motors how to convert the given g cm^2 from the motor to put into perspective J for comparison... or do I get it right when J = g*cm^2 ?

wilriker · 17 Sept 2018, 12:35

@lb OK, gear ratio is now added under Advanced Settings.

About your question on how to convert g cm² into J. In the online calculator this is done the following way:

rotorMass = (motorCount * rotorInertia) / (pulleyRadius * pulleyRadius) / 1000.0;

Is that what you asked for?