I am an artist and I have no electronics skills, at all. I have a specific project that has brought me here, on my knees, looking for help. I need to power a person along a track, on 4x 90mm skateboard wheels. I started with eskate world (BLDC/VESC) but could not get the low RPM (<600rpm) with any precision control. So I moved to 2 x stepper motors (geared down 2:1) with arduino UNO control. The steppers are each rated at 3nM, so I figured I should have enough torque - except I can stop the wheel with my fingers. We've tried every stetting on the driver, and the power source (2 x dewalt 20Vmax 9Ah in series (@44V)) seemed capable of delivering enough charge. I tried an alternative (48V /40C no difference)
I measured the amps at failure and it peaked at under 2amps before the motor stops. I had help with the code and the setup but the person who did that is out of answers, and I have no idea at all.
I guess I'm asking - is this setup capable of doing the job (moving a person on a mostly flat track) and where is the weak point? Would a servo motor be better? the driver? arduino code? I'm out of ideas and money and desperate to get this 2 year project operational. I humbly throw myself at your mercy! I will post whatever detail I can. I've attached the wiring diagram.
That stepper motor driver should be capable of driving your motor - I assume it can provide 4.2 amps continuously.
I think what's missing from your analysis is a measure of the force that is required to move your truck. It will vary by a large amount depending on the smoothness of the surface, whether the surface is level and the quality of the wheel bearings. The only way is to measure the force. One way is to connect a string between the truck and a spring balance and measure the force needed to pull the truck.
After that you need to convert that force to the torque needed at the stepper motor shaft. That is essentially a matter of gearing, including the diameter of the driven ground wheels.
When you have done all that I suggest you get a motor with at least twice the calculated torque.
I would expect that a DC motor with suitable gearing would be a better choice than a stepper motor. Stepper motors are appropriate where very precise positioning is required. You say "could not get the low RPM (<600rpm) with any precision control" but you have not told us what sort of control you need. With suitable gearing a DC motor can drive a shaft at any slow speed.
Stepper motors are about the worst choice possible for traction, they are very inefficient and pretty slow (with torque decreasing dramatically with speed).
For low speed control of DC or BLDC motors you need a servo loop and an encoder - that will give you
full torque at any speed, even stationary.
Thank you mark and Robin - you guys are legends and I'm flattered.
The track is mostly flat and hard, a single rail. I looked to escooter and eskate for a guide on the torque needed (esp as I only want to go 2-4mph, slow and steady) and the median is around 1-3nM to get an adult up a hill and to around 15mph. So I used two of them, and geared down (2:1)to get a total of 12nM, which I figured was enough headroom. Further, the holding torque aspect of the stepper seemed attractive from a safety point of view.
As for control - ultimately I want to take the rider around the track through an art installation, so to be able to slow/stop at various points (with triggers) would be ideal. I know the stepper is great for precision but not necessarily position ( I had tried to avoid looking at servo for costs!)
I looked on ebay and they steer me toward 'hybrid' stepper/servo. like
but I notice it says 1nM (1/3 the torque at 3x the price) and can this rig then be positionally controlled via the arduino? Obviously I'll need new code - I'm pretty wiped for funding after the bldc and now stepper motor fails. Are there decent servo motor options at the less expensive end you might recommend?
Thanks again fellas.
ps I'm really good at art installations and experiences!! This part not so much.
The symbol for the newton is 'N', the symbol for metre is 'm', so 'Nm' is the unit of torque.
You have a BLDC, you have a driver for it? Then you can build a servo loop around it and make
your own servomotor. Its encoder -> PID loop -> motor driver.
Do you have details of the BLDC and its driver, and the mechanics of the transmission setup,
ie gearing, wheel diameter, wanted thrust, torque, power etc etc, all your calculations, for a
sanity check.
in the BLDC dept I have several motors from eskate world - a couple of them lower KV (90-140 KV) hub motor types, and a couple of VESC drivers. The main issue was getting them to run enough torque at low revs - I'm limited in the amount I can gear them down (mechanical configuration) and the holding constant speed and braking was an issue - As Paul wrote above, this ride could get hairy quickly - they need to be slow and safe and reliable.
The steppers seemed to offer good torque at low revs, and more precise control. I have the motors installed and operational - its just hard to accept they dont have enough power when they are rated at 3Nm each?
If they are rated to 4.2 amps, any suggestions as to why I can only get to under half that before they fail?
Is there possibly a servo setup that can do what I want (inc Arduino control, positional control, solid holding torque and high torgue at low RPM) that isn't too expensive?
I'm in dire straits and could really use some help.
REYX:
If they are rated to 4.2 amps, any suggestions as to why I can only get to under half that before they fail?
I don't know what you mean by "fail". Perhaps you just mean missing a step?
I suspect the problem is that you don't have suitable measuring equipment. A multimeter will not respond quickly enough. The stepper motor driver will check and adjust the current many times (maybe hundreds of times) per step.
Have you taken steps to measure the force needed to move your truck? Without that it is hard to see how there can be progress.
A diagram that illustrates the space within which the motor must be located and the limitations on the use of gearing would be a great help.
A link to the datasheet for the most promising of your DC motors would also help.
I think measuring the amount of force might be beyond me at this point. I think it is fair to say it lis LESS than an electric skateboard or escooter, which come in around 1-3Nm. Anything more precise than that at this point of the project seems extraneous.
When i say fail, I mean the motor stops spinning and makes a terrible grinding sound and needs to be manually spun to start spinning again. I've seen it in many stepper tests.
The bldc setup struggled with torque holding, and that was a safety issue for which I'm yet to find a solution. I'm interested in what the VESC can do with regen braking, but never made a breakthrough.
The gearing issue is 40 teeth is biggest gear before it interferes with the rail, so I figure max is 4:1
Does anyone have any more information on the servo loop option? Clearpath seems to have a lot of advantages, except for cost.
I need strong two matching motors and drivers - slow, steady and responsive to arduino triggers.
Robin do you have a dc motor setup in mind? I'm all ears!
You are confusing force and torque. You can measure the force required to pull a person along a track, riding on wheels, simply by using a luggage scale. Be aware that starting up from rest requires much larger force than moving along at constant speed.
That force can be converted to the torque that must be applied to the wheel axles by a motor; simply multiply the force by the wheel radius. Review of force and torque, applied to robotics, here: How to Build a Robot Tutorials - Society of Robots
Be sure to use appropriate units. 1 kg force = 9.8 Newtons
I mean the motor stops spinning and makes a terrible grinding sound and needs to be manually spun to start spinning again
That sounds like a terrible mechanical failure. Nothing to do with the driver or power supply.
The toothed belt and gear arrangement substantially reduces the motor shaft torque.
I suggest to drop the stepping motor idea and consider something like this gearmotor. I have one of those for operating a heavy duty window vent and it is extremely well made, with a dual ball bearing stainless steel output gear and shaft.
REYX:
in the BLDC dept I have several motors from eskate world - a couple of them lower KV (90-140 KV) hub motor types, and a couple of VESC drivers. The main issue was getting them to run enough torque at low revs - I'm limited in the amount I can gear them down (mechanical configuration) and the holding constant speed and braking was an issue - As Paul wrote above, this ride could get hairy quickly - they need to be slow and safe and reliable.
The steppers seemed to offer good torque at low revs, and more precise control. I have the motors installed and operational - its just hard to accept they dont have enough power when they are rated at 3Nm each?
If they are rated to 4.2 amps, any suggestions as to why I can only get to under half that before they fail?
Is there possibly a servo setup that can do what I want (inc Arduino control, positional control, solid holding torque and high torgue at low RPM) that isn't too expensive?
I'm in dire straits and could really use some help.
Thanks again all.
Rey
REYX:
(and here's a photo of the truck with gearing limitations and other heartbreak)
That belt appears to be slack. A toothed belt needs to be rather tight to prevent the belt from jumping when torque it applied.
Paul_KD7HB:
That belt appears to be slack. A toothed belt needs to be rather tight to prevent the belt from jumping when torque it applied.
Paul
Thats a problem I hope to encounter!
jremington: The toothed belt and gear arrangement substantially reduces the motor shaft torque.
I suggest to drop the stepping motor idea and consider something like this gearmotor. I have one of those for operating a heavy duty window vent and it is extremely well made, with a dual ball bearing stainless steel output gear and shaft.
jremington: The toothed belt and gear arrangement substantially reduces the motor shaft torque.
I like the idea of a geared motor, but it says the torque rating is .3Nm ( and even that is the biggest 24V model) and given the steppers I already have are rated at 3Nm, I don't think it will be enough.
I'm still at a loss as to why the dual 3Nm with 2:1 gearing are so weak? Shouldn't the Nm value be the final
If there was a cast iron alternative I would be right on it.
Thanks again for all the thoughts - I really appreciate it. I hope I dont come across as ungrateful for the input, I really am, but I just spent the last 1k dollars on the stepper motors and coding and yet I'm back at square 1.
You must be looking at the 470 RPM motor (0.39 Nm), which would be much faster than your stepper can rotate under load and a poor choice for your application. There is a trade off between RPM and torque for given power input.
You still have not made the basic measurement suggested in reply #9. Your project will probably not be successful until you do.
Take time to learn some basic mechanics, determine the required torque and RPM by direct experiment and forget the stepper. The Society of Robots site I linked earlier has other relevant tutorials.
Stepper motors, like reciprocating steam engines, exert maximum torque when stationary. But, use a toothed belt, or any flexible drive, and it loses that advantage. It will start sort of 'cogging' on heavy load, A basic explanation being that as it starts moving, the torque diminishes, just as the drive belt begins to stretch, and it springs back to where it started, driven by the elasticity in the belt. Gecko were one of the first companies to manufacture low cost drivers, here's a link to some basics
I DID A TEST! It was not pretty, and def not scientific, but I got some data.
From a standing start, 15 kgs of force was sufficient to accelerate 2m/s (90kg load, on 4 x 83mm 80A, 8mm ballbearing wheels) up a 10% grade. This meets the requirements.
when rolling back (negative force?) it was possible (with great effort) to exert 34 kgs of force to accelerate back up the grade to 2m/s. This exceeds the requirements, but always good to know.
If 15kgs is sufficient, 25kgs would be a number I'd be comfortable working with.
25kg x 42mm = 105kg/cm =1kg/m = 10Nm (I could have a decimal place out. Pls excuse)
That is divided over two motors, and geared down 2:1, which gives us 2.5 Nm per motor.
Pretty close to the 3Nm is should have with these infernal steppers, but I think I have to let them go eh?
Does anyone know anything about Clearpath servo's - they cost a bit but seem to perform above par.
Thanks again all concerned. This will be a magical art experience once complete - I've built the track and this is the last hurdle!
