Stepper Motor problem

Thread Starter

Ajandco

Joined Jan 16, 2022
11
Hi. I’m a mechanical engineer needing some electronic engineering input for stepper motor system design. My motors rotate an observatory dome at home.

The motors are controlled by a Python program on a Raspberry Pi connected via DRV8871 drivers powered by a 9V transformer. I have attached a diagram of my setup and listed below what I think are the relevant specs for the electronic components.

The motors all run well in air (unconnected to the dome) and so I think the wiring and programming side is all ok. I’m sure that I’ve selected the correct motor torque, power and number of motors to drive the dome (with a generous overdesign on the number of motors based on the measured breakaway torque to get the dome moving).

My issue is that, although the motors rotate the dome most of the time, the motors frequently and randomly stall ('chatter'), where the motor shafts just stop turning and oscillate clockwise and anticlockwise a fraction until I turn the power off or give the dome some assistance to get it moving again. The number of drive motors installed makes no obvious difference to the stalling issue.

As I'm now sure I have ruled out all the potential mechanical issues that could have been preventing the dome from turning, I'm suspecting it's my lack of knowledge on the selection of the electronic components I’m using.

Spec for Stepper Motors 17HS15-1684S-PG5:

Amps / Phase 1.68A

Spec for the DRV8871:

6.5V to 45V motor power voltage

3.6A peak current

Spec for DC transformer

9V, 22A

I would be grateful if anyone out there who understands stepper motors better than me could cast an eye over my setup and point out any obvious issues / improvements I could make.

Also, any idea what the chattering is a symptom of?

Thanks for reading
Alastair
 

Attachments

John P

Joined Oct 14, 2008
2,025
It sounds as if the chattering is simply the effect of the motors stalling. That's what steppers do.

I am wondering about the whole basic setup here. Exactly how are these 8 motors (8, really?) connected to the load? I'm imagining some kind of gigantic ring gear, with 8 little pinions driving it at different points. Then I wonder what sort of bearing it rides on. A problem that I can see is that if there's any mismatch between the amount of rotation that one motor can make relative to another, you're likely to get the torque from one fighting against the torque from the other. And it's not just "one versus another", there are 8 of them! It wouldn't matter if they were DC motors, which are just torque sources, but with steppers you have to allow them to move to the desired position or very near to it, or you get a stall condition. And you've got a very large moving structure with some inaccuracy likely. Basically what I'm saying is that if you have a rigid connection between multiple steppers, then it's got to be accurate enough to let them move in perfect lockstep, or you'll have problems. But maybe I've got a totally wrong impression of how this thing works, and the trouble is somewhere else.
 

Thread Starter

Ajandco

Joined Jan 16, 2022
11
Many thanks to both for your replies, much appreciated. I have ruled out turning too fast and accelerating too quickly so I think that leaves installation inaccuracies which are of course inevitable. I had not fully considered the difference between steppers and other motors that John indicates and the more I think about it, my problems look to stem from the steppers fighting one another. In theory the motors should work in perfect lockstep, but the potential sources on inaccuracy in the construction are just too many. Based on my requirements for turning the dome CW or CCW to a specified angle, together with limited available installation space, several small steppers had looked like the ideal solution : (

Thanks again.
 

dendad

Joined Feb 20, 2016
4,451
Have you thought of running the motors on a 12V or greater power supply?
My 3D printers are 12V or 24V. 9V may be a bit low as higher voltage lets the current ramp up faster.
Also, if you can fit them, have one motor driver per motor. That in itself will double the motor current, and increase the torque.
2 of the 3 3D printers I have use 2 motors to drive the same mechanism so I think all the motors driving the one gear is ok.
This is probably a silly question, but have you checked they all run the same direction when you tested them unconnected to the dome?
 

Thread Starter

Ajandco

Joined Jan 16, 2022
11
Hi dendadd, thanks for your input. My transformer lets me change the output voltage and I wound it up to its maximum of 14V output but the problem remained. I'm estimating at 14V the transformer output would reduce to about 14A which is still > 8 motors @1.68A/phase (8x1.68=13.4A<14A).

It’s interesting you have 2 motors working on the same mechanism, that should give me hope to try further tweaks to my setup, but in all honesty after 6 months of trying I’ve all but run out of fresh ideas.

I did check the unloaded motor direction and they were all the same and the drives all run perfectly. As for the drivers, my reasoning was that each motor draws 1.68 A/phase, and as the drivers are good for 3.6A then 1 driver should be good for 2 motors (2x1.68 = 3.36A < 3.6A) on full step. My assumption is that each phase is active at a different time which seems logical to me but perhaps the phase loads overlap? I have no way of checking and my knowledge level here is sketchy. Thanks again.
 

shortbus

Joined Sep 30, 2009
10,045
I'm late to the thread but have a question for the TS. How are you driving stepper motors with the DRV8871?
That is a brushed motor driver something that should not work with stepper motors unless you have a new way of using them. https://learn.adafruit.com/adafruit-drv8871-brushed-dc-motor-driver-breakout

Also how heavy is this thing? Is there a reason you chose the smallest normally used stepper motor, the NEMA17? Also most every real stepper motor driver suggests using a 32VDC supply to drive the steppers. The name plate voltage on a stepper is not the one to use it is what is done for some test when they are made. The Amperage value is the one you have to watch when using stepper motors. Using too low a voltage is one cause of the chatter in stepper motion.
 

Thread Starter

Ajandco

Joined Jan 16, 2022
11
Hi shortbus

Thanks for your valid input. The dome is very heavy, geodesic design in wood with fibreglass overlay. It is limited space that forces the selection of multiple small motors. Your point about the DRV8871 looks to be very valid, I hadn't noticed that. I took a vendor's recommendation for their selection and had no reason to question it until you pointed this out, particularly as the motors turn at the rotational speed I expected. Perhaps the motor power output is affected? I would be hugely grateful if you could point me in the right direction for an alternative. Thanks again.
 

Sensacell

Joined Jun 19, 2012
3,432
You may be experiencing torsional resonance.

This can cause stepping motors that run at a constant speed to stall seemingly for no reason.
Some drivers have "Anti-Resonance" features, other solutions involve mechanical dampers or careful choice of slewing speeds to avoid resonant frequencies.

The motor rotor starts to "ring" or oscillate torsionally, excited by the stepping impulses, at a certain point this oscillation consumes enough of the motor's output power that it cannot drive the load and just stops. It remains stopped (jiggling around) because the stepping frequency is well above the motor's minimum "pull-in speed" - it's dead until the system stops.

https://www.automate.org/tech-papers/solutions-to-reduce-stepper-motor-resonance
 
Last edited:

LowQCab

Joined Nov 6, 2012
4,023
Remote-Mounting of a single Stepper-Gear-Motor would seem to be the best arrangement.
Bicycle Cables will allow the remote power-transfer.

The Bicycle-Cabling needs to be purchased in bulk-lengths on a spool.

The Stepper-Gear-Motor will need an Aluminum-Pulley made with the smallest diameter that's practical.
The Stepper-Gear-Motor will preferably have a Worm-Drive-Gear-Ratio of around 100 to 1.

The center of the Inner-Cable should be wrapped around the Pulley 2.5 or 3.5 times with no overlapping.

The start of the Outer-Jackets of the Cables need to be mounted, side-by-side, and
spaced apart by the diameter of the Pulley.
This "Mount" should be movable, and spring-tensioned away from the Pulley, to keep the Cables tight.

The far ends of the Outer-Cable-Sleeves need to be clamped into a single Mount which holds the ends of
the Outer-Cable-Sleeves pointing in opposite directions, ( one-left, one-right ).
This far-end Mount must be rigidly attached to the lower-supporting-structure of the Dome,
and be positioned as close as practical to the bottom perimeter of the Dome.
The Inner-Cables will then be wrapped in opposite directions around the base of the Dome,
and after each Inner-Cable is wrapped 360-degrees around the base of the Dome,
they will meet, and must be clamped in place, together, to the base of the Dome.

This arrangement allows for slightly over 360-degrees of rotation,
in either direction, before requiring reversal,
and has an absolutely absurd amount of fine positioning resolution.

And the Single, High-Power-Stepper-Motor,
can be mounted all the way over in another room if you'd like it to be super-silent.

Bicycle-Brake-Cabling was just an example,
I don't know how big your Dome is,
but similar, larger-sized, Cabling is readily available.
High-End "Morse" style Cables are available that can withstand well over ~1000-pounds.
Bicycle-Cables can usually withstand about ~50-pounds of tension, so it should work just fine in
a situation where You were wanting to use the smallest sized Stepper-Motors available.
.
.
.
 

Irving

Joined Jan 30, 2016
3,843
The DRV8871 can be used to drive a bipolar stepper motor but you need two per stepper, one drving each phase winding. This is what you've arranged to do though its not the easiest or best way to drive stepper motors, and there are much better drivers. You cannot however drive two steppers off the same driver - you have no way to ensure that the phase current is shared equally between the two motors. Doing this will almost certainly, under load, cause one of the steppers to miss-step effectively going into reverse compared to its companion. That is the cause of the 'chatter'.

This is, unfortunately, a common 'noob' problem with home-built CNC machines and cheap low-torque steppers paired (or more) to give the higher torque needed...
 

shortbus

Joined Sep 30, 2009
10,045
The center of the Inner-Cable should be wrapped around the Pulley 2.5 or 3.5 times with no overlapping.
I've noticed you saying that before, but when I was looking to build my portable gas powered windlass, all of the references on windlasses and winches said a minimum of 5 wraps was needed. Even winches say never use if you can't have the 5 rounds/wraps on the drum.

Using the cable though is a good way of doing it. I have a very old snowblower that uses the cable and drum to change the position of the discharge chute, and it still works good.
 

LowQCab

Joined Nov 6, 2012
4,023
I've noticed you saying that before, but when I was looking to build my portable gas powered windlass, all of the references on windlasses and winches said a minimum of 5 wraps was needed. Even winches say never use if you can't have the 5 rounds/wraps on the drum.

Using the cable though is a good way of doing it. I have a very old snowblower that uses the cable and drum to change the position of the discharge chute, and it still works good.
.
The number of "wraps" around a "Windlass" style Pulley is directly related to the
amount of Friction that can be generated between
the Pulley-material, and the material that is wrapped around the Pulley,
and the amount of tension required to generate the required Friction using those particular materials.

In the case of an Aluminum-Pulley, and a small multi-strand-twisted-Steel-Cable,
3.5-wraps, at, lets say, ~10-Pounds of Spring-Tension,
( which would be multiplied by the number of wraps ),
would probably generate enough Friction to break the ~50-pound-rated Cable,
provided that the Motor can provide enough Torque.

In fact, it might even be better to insure that the Pulley will slip before the Cable breaks.
1.5 wraps may be more than adequate.
.
.
.
 

John P

Joined Oct 14, 2008
2,025
The trouble with these cable and pulley systems is that there's no way to keep the drive synchronized--where is position sensing ever going to get done?

For stepper drives, I just bought a few of these:
https://www.pololu.com/product/1182

They're tiny and inexpensive, and seem to work well. Obviously the current is limited, and I don't know how it matches the needs of this project. But really, I don't think steppers are right at all. DC motors would be a lot less trouble--except that as with the cable drives, position sensing has to be dealt with some other way. Perhaps motors with encoders attached would be the way to go.
 

LowQCab

Joined Nov 6, 2012
4,023
"" The trouble with these cable and pulley systems is that there's no way to keep the drive synchronized,
where is position sensing ever going to get done? ""


The same way that it is being done at present, if it is actually being done.

A Rotary-Encoder with a Pulley, with one of the Cables wrapped around it, has zero-slippage.

If that's not "good-enough", an Optical-Slotted-Disc mounted around the
lower perimeter of the Dome would take care of any questions,
but you'd get better resolution from the Encoder and Pulley,
You can also add one to the Stepper-Motor-Shaft,
or use all 4 methods at the same time to verify each others accuracy.

The Cable method has excellent repeatability and stupid-high-resolution.
.
.
.
 

Irving

Joined Jan 30, 2016
3,843
HTD pulley & belt drive would be my choice. Aluminium pulleys and zero backlash kevlar belt. Often used on high-end CNC machines with resolutions < 0.01mm. Also by 'gearing up' a much lower torque motor can give excellent results. A gearing of 20:1 (18/360 pulley set) would potentially allow 1 motor to do the work of the 8 in use, plus increasing the step resolution ie 1 rev motor @ 200step/rev = 1/20 rev of dome = 18° or 1 step = 0.09° , so using a decent controller and stepper motor with microstepping could give ~20arcsec resolution if the mechanics are up to it...

Personally I wouldn't use a stepper for this purpose as the stepping can be a problem. A normal DC motor geared down with a final 20:1 drive and an absolute 10,000ppr encoder on the motor shaft would give ~7arcsec resolution.

The attached paper, though aiming at a more sophisticated solution, nevertheless gives some insight into the design process for both polar and declination axis control systems, including the use of dual drive motors for backlash control.
 

Attachments

BobaMosfet

Joined Jul 1, 2009
2,110
Many thanks to both for your replies, much appreciated. I have ruled out turning too fast and accelerating too quickly so I think that leaves installation inaccuracies which are of course inevitable. I had not fully considered the difference between steppers and other motors that John indicates and the more I think about it, my problems look to stem from the steppers fighting one another. In theory the motors should work in perfect lockstep, but the potential sources on inaccuracy in the construction are just too many. Based on my requirements for turning the dome CW or CCW to a specified angle, together with limited available installation space, several small steppers had looked like the ideal solution : (

Thanks again.
I would recommend a single larger non-stepper motor, with a rotorary encoder.
 

Thread Starter

Ajandco

Joined Jan 16, 2022
11
A final thanks to all who contributed with your input into my problem.

Unfortunately, the design of my dome and its bearing support arrangement just doesn’t lend itself to the alternative external drive arrangements proposed. A pity I hadn’t considered them in my investigations before I started on this journey as I would probably have gone in a different direction, but many thanks for sharing in such detail.

All your responses have given me food for thought and a few ways forward that I hadn’t considered and have also sparked fresh ideas too.

My intention initially is to continue with the stepper motor design rather than the normal DC motor route for now. Steppers just seem such an elegant method of controlling the dome motion, and I’ve invested so much effort into getting to where I am, that I need to try and pursue the potential solutions until I can go no further on that road.

It was great to be able to reach out to you all on your forum when my ideas ran out, so thanks once again to you all for your time and ideas and for sparking my project back into life. I will try and update you at a later date – this could be months. As you know, these things take time.
 

John P

Joined Oct 14, 2008
2,025
Ajandco, I'm sorry that you've done a lot of work which so far hasn't rewarded you with a working system. But one thing ocurred to me that might make a multi-stepper drive functional, and that would be to add some kind of resilience to the individual drives. If each motor delivered its output via a spring (essentially) maybe that would allow some mismatch in the motors' rotation versus the dome's movement, and you'd basically see the dome move based on an average of what all the motors were doing. But then again, I can also imagine that at any given moment, maybe some of the motors would be delivering forward torque and some would be holding the overall travel back, and perhaps that would still be a problem. I can't make up my mind what to expect!
 
Top