Stepper Motor Application

[Troikos]

Electric Reverse

So I've been tossing this idea around for a few weeks now about using a stepper motor in the drive system. Its primary purpose would be for electric reverse as well as electric brakes and I was wondering what the groups thoughts on this would be. Power systems aside (already have them planned and not really relevant), How practical would would an application like this be?

Motor Specifications:
Frame Size: NEMA42
Step Angle: 1.8 degree
Voltage: 5.86VDC
Current: 8.0 A/phase
Resistance: 0.67Ohm/phase
Inductance: 12mH/phase
Holding torque: 300Kg-cm 4120oz-in
Rotor inertia: 1620 g-cm2
Number of wire leads: 4
Weight: 11.7kg
Length: 201mm

I especially like the NEMA42 motors because of their extremely high torque and 3/4" keyed shafts. Even though the motors are only rated for 3-5 volts they are typically driven with 24-80 volts dc with current limiters.

Here is a graph of a typical stepper motor torque curve. I couldn't find one for a comparable motor but the curve is similar on all of them. Basically, the torque goes down as speed goes up which is why I think it would only be practical for an electric reverse.

Thinking I could use it as a brake because my vehicles are remote controlled so it would be easier to just energize the coils in the motor to stop the vehicle than rig up some sort of mechanical system.

Thanks for the read, here is an obligatory video feature stepper motors playing music with their resonance.

http://www.youtube.com/watch?v=Xk_XaJ7gE4Q

 

[itsid]

What?

You want to drive a kart with a stepper motor?
WHY????

Do you need to move the kart an exact distance at a time at very slow speeds?
See? the same answer applies to the question if using a stepper motor as a driving motor for a kart is a good idea

Again, if I understand you correctly you want it to brake the kart and have reverse.. that's fine, and possible with any AC or DC motor, no need for Steppers or Servos here.
Plus, steppers do heat up very quickly and die because of the heat even quicker if not used correctly.

Also don't mess up holding torque with rotational torques,
300 kg/cm is great for keeping the object stationary (like the spindle for the heavy duty mill you run)
It is not applicable as a brake system, since the motor is already turning...
you'll have more like 20 kg-cm which is close to nothing (200 grams)

The same goes for acceleration, as soon as the motor is turning it'll loose most of it's torque, because that's how a stepper motor works.
it's about precision, it's more powerfull when turning slowly to keep the precision..
you don't need any precision, you need power on the other end of the rpm band.

just forget about it.
Sorry, but that's not a good idea at all.
It'll be easier with a modified servo motor of the same size... still useless (you'd remove all the 'servo-ness' to make it work properly and end up with an expensive DC motor)
but at least a Servo is a DC Motor which you would want for the given task.. a Stepper isn't even that really (yes, it's a motor and it's DC ..but still it's no DC motor in the end)

'sid

 

[Troikos]

Hah, Thanks Sid. I didn't think the torque drop off would be so much. Personally, I've never used them before so I'm not familiar with their limitations. It was just an muse I have been throwing around. Thanks for the info and back to the drawing board.

The problem I'm having is finding a DC motor with a large enough shaft to make it practical. There seems to be no middle ground between scooter motors that have a tiny shaft suitable for a #25 roller chain or the motor is simply over sized in the case of most golf cart motors.

Idealy, I'm looking for a 12-24v motor with at least a 1/2" shaft. Back to the search for me.

 

[Oxymoron]

Look here.

 

[Troikos]

Thinking that a golf cart starter-generator is the perfect fit for this project. Can be run in both directions and when its not being used as a driver, it can charge the battery. Runs on 12vdc, has built in mounting brackets and a good sized keyed shaft. Plus they are relatively cheap to purchase a new one.

 

[Jim-L-L]

I used to rebuild Aircraft & Helicopter Starter Gen's as a Apprentice, they ran as a DC motor until the turbine engine started and then as a gen after up to 350 Amp for the big ones, they were designed to run all the time at 3500 rpm so had high tolerance bearings the actually start torque was not that high as they were geared low out of the accessory gearbox and used all the turbine internals to maintain inertia, the problem we always found that if a DC motor is pushed to hard the armature will over heat and melt the solder and then the centrifical force will throw the solder out causing failure of the armature another problem was arcing on the copper commutator and carbon brushes which required major overhaul. In short try and find a motor that is made for continous duty

 

[Troikos]

I'm actually not planning on using it for continuous duty. At most it will run for 5-10 seconds at a time as an electric reverse and I can gear it down because its not needed for forward motion. Really, its more of a convenience thing than anything else. The vehicle will be running a raspberry pi as the main controller and I will be outfiting all of the motors with temperature probes and could use that as a thermal cut-off if needed.

One thing I did want to ask is if it would be a good idea to put in a large cap to handle current spikes. Is it even needed for a motor this small? Is it a good idea anyway?

Thanks for the info!

 

[itsid]

Ah it's just for the push back..
then it should be fine.
(maybe a car starter will do too.. should be cheaper and easier to find)

The raspberry Pi is great as a brain but it's a sh1tty controller, it cannot handle anything close to the currents you'll need.
So you need a motor driver to run it and IIRC that's a rather expensive task.
It's cheaper to find a decent small motor controller and just add a throttle (such controllers cost about the same as the raspberry, since you don't need nothing fancy nor extremely powerfull at all)

But I like the raspberry idea alot, enough room to play with different electronics at the same time,
temps rpm readings, speed, tilt, and about every sensor you can find (acceleration, GPS, light,sound etc) and you could even fire up a remote control with your smartphone if you feel like it.

Don't go too big on the cap, or you'll have more coasting than intended;
a decent controller shouldn't produce any spikes noteworthy so you can just leave the cap,
if you build the motor driver yourself for the raspberry pi and are unsure about spikes, get a midsized cap perhaps, just to be on the safe side if you wish.
I'd still consider it unneccessary though.
you can run a starter motor by just connecting a battery via croc clamps, the spikes involved wear the brushes a bit, but that's about it...

'sid

 

[Troikos]

I may not be well versed in kart building but micro controllers I am familiar with. I'm well aware of the limitations of the Pi and have all of the support systems planned. i2c servo driver(the pi can't control servos directly), ESCs for the drive and if needed, I have a few arduino nanos kicking around for controlling sub systems. The Pi is mostly for data logging and display using an onboard 5" LCD screen as well as pushing data to a tablet via wifi. It also needs to check all of the inputs and ensure that if control is broken, it can recover and prevent a crash autonomously (kinect for obstacle detection).

The electric motor would be driven using a PWM ESC so there shouldn't be any coasting if the cap is hooked up parallel with the battery. I just don't want anything browning out due to voltage drops which would be extremely unlikely anyway considering everything but the motor and steering servo will be using voltages stepped down to 5v.

You had mentioned that I should just find a small controller and add a throttle. Not sure if you saw my intro thread but this is actually remote controlled. Here is the video of the prototype that I'm running right now. All of this is actually for version 2.

http://www.youtube.com/watch?v=Xc-jv1anm6I

I2C Servo Driver: Adafruit 16-Channel 12-bit PWM/Servo Driver - I2C interface - PCA9685



Thanks for the input sid, you are well versed in everything it seems.

 

[itsid]

Ahhh good to know,

Some "non programmable" motor controllers do provide a very basic com port as far as I know, so maybe you'll be able to tap into that, so you don't have to custom build a motor driver module (I don't think there is one for the raspberry pi already *shrugs*)

In any way you could emulate throttle and brake switch for all controllers with the Pi,
they're mostly hall effect for the throttle and a simple to ground switch for the brake.
all that's left for an input there is the batteries, so you're good.
(the MACH1 scooter for example comes with a very basic controller that can be had for less than $40 on amazon capable of 1000W @ 36V which should be enough for most starters)

And, now that I know that you know about controllers, I'm sure I will make use of that sooner or later

But one thing..
Servo controlled steering?
How big a kart are we even talking here?
And will it be radio controlled?
Please say it's going to be autonomic

Pictures, drawings.. and moore infooo please

'sid

PS Well I'm curious about almost everything, that's about all there is to it

 

[Troikos]

Here are the pictures from my intro thread.



Its built using the frame from a mobility scooter right now so its a little smaller than a standard kart. I have a Honda GC160 on it right now. The video showed it using an electric actuator with a pot for feedback control but my new steering servo should be in this week. The servo alone will be one of the more expensive parts on the vehicle.

The car is controlled using a 2.4ghz RC controller right now. The current version needed an arduino to read in the RC values from the receiver and voltage from the pot to drive an h-bridge hooked to the actuator. Suffice to say, I'm relieved that I will have a proper servo in place. The servo itself is run directly off 12v with a standard RC input. I'm also using an RC ESC to drive the electric motor. You can get those with large power outputs for cheap too. There is also a servo hooked to the throttle of the engine.

While I like the idea of it being autonomous... I don't like the idea of an autonomous car weighing in at over 200 pounds driving at speeds in excess of 40MPH making its own decisions on where to go. The autonomous functionality would only be there to lock the breaks and do collision avoidance in the event of a lost control signal.

The next 1-4 weeks will be like Christmas will all of the new parts coming in and you can definitely expect a build thread. Tomorrow I heat the garage and take measurements of the existing frame and components to use as a reference for the drawings for version two. While its still -40C and colder outside, the frame build wont start until spring so I have some time to get the drawings and schematics done up and the electrical system designed and programmed.

I hope you wont mind taking a look at the drawings and correct me where needed. This will be the first big welding/steel project I have ever done.

This car will bring a whole new meaning to the term over-engineered.

Picture of the servo for reference.

[EDIT] Just to expand on your reference to controlling a motor with the pi, because the pi runs a high level OS, its terrible for doing simple things quickly. It cant do PWM with any sort of effectiveness because the processor spends so much time running around completing background tasks like a child suffering from ADD that it could never keep the timing consistent or even correct. Best to pass off the simple repetitive tasks like PWM to a sub system like that servo driver I linked.

 

[itsid]

Wow, that is a massive servo

Oh btw I think there is a RTOS for the Pi.. I think you can even run a bare Unix or Linux both are if I'm not mistaking real time operating systems *scratching the head again* uhm...
Sorry, since I don't have no Raspberry Pi I'm not exactly sure, but I DO THINK I've seen a RTOS for the Pi somewhere.
So it should be very well capable of running PWM signals, just as good as an Arduino for example or a vintage Atari portfolio running DOS and since it's heart beats at a faster rate even more precise than those two

But I guess, that's where you'll loose some of your other software that your little friend wants to run.

Anyways.. it's really nice (I even remember it from your welcome thread.. I just forgot to connect your name to it.. my apologies )

Oh and about the autonomy of it.. with enough sensors and proper fallback codes I think it'd be great.
Of course you would ant a remote killswitch and a homing signal,
but from all that I know so far the raspberry is practically capable of all of that.
It's just all the low level software that's making this hard to achieve...

But with all the autonomous mowers everywhere I think it shouldn't be too hard to find some code examples.

Someone must have written or at least ported code for the Pi already to achieve that..
I'll take a look around tomorrow.

Man, that servo IS huge!!
That'd be a nice knee joint for a bipedal half scale robot.
How fast is it? and what's it's torque?

'sid

 

[Troikos]

The come in a variety of configurations. I wouldn't want to be buying many of them though.

http://www.invenscience.com/index_files/torxis_rotary_servo.htm

The strongest (and slowest) one is capable of lifting and holding 110lbs.

http://www.youtube.com/watch?v=zWJ3icH9YLE

Here is a quad to RC conversion using this servo for steering. Mine will be styled after a car though and not a beast like theirs (which is awesome by the way). Planning on decking mine out with all kinds of tech so going off-road isn't really in the plans.

http://www.youtube.com/watch?v=BSJcWFKSOKI

I'm definitely looking forward to the autonomous part. I'll have to find a large parking lot and make an obstacle course out of barrels or something.

 

[itsid]

Sooo

autonomous Pi.. too much hits on google to list or even read the first page
That'll be fun once you get there
Many things have been done already and as many parts of the code are available ...

Okay, that servo is a beast...
it's way too slow and still too weak for a knee joint though (Yeah I knew that already, nevermind)
But that is an impressive video for sure.

Their 'thing' looks like it's made off a minibike frame for the main chassis.
Not exactly a bad idea for a roll cage and since they treat it rough they may have use for it

'sid