Stepper Motors in Parallel

[Brian GParticipant @briang]

Back in April, Neil, describing his Prusa 13 Build **LINK** said "Something that surprised me, was the effect of having two steppers wired in parallel for the z-axis. You turn one, and the other moves in unison!". Does anybody know if this is normal behaviour for stepper motors or if a particular combination of motor and driver is required?

(If practical and reliable, it would provide a way for a model railway controller to be operated from more than one panel, solving a problem at our club.)

Brian

[Brian GParticipant @briang]

[Nick HulmeParticipant @nickhulme30114]

Steppers wired in parallel do this in the absence of any other equipment.

You do get lost steps though, so maintenance of exactly matched indexing is not a guaranteed function.

[Joseph Noci 1Participant @josephnoci1]

Not sure Nick's answer is answering the question though…

Two stand-alone steppers of similar spec and size, with no stepper driver, power supplies, wire phase to phase will behave as Nick says – turn the shaft of the one motor and the shaft of the other will/may turn, simply due to the induced voltage in the turned rotor inducing a current in the windings of the second motor, making its rotor turn. However, this does not work reliably, especially at low rpm – the induced voltage is too low to generate sufficient magnetic field in the second motor. Making the driving stepper much larger helps this, but no real practical value to this. The second motor will regularly loose sync with the driving motor.

Since Brian asks

'Does anybody know if this is normal behaviour for stepper motors or if a particular combination of motor and driver is required?'

Since this implies a stepper driver is involved, I suppose the answer would be – you can connect two (or more) identical stepper in parallel and connect them to a standard stepper driver, at twice ( or more) the required current for one motor, and they will step together and remain in sync. If they are mechanically in parallel as well, any loads will be shared, and so will loss of steps. This is often done in long axis drives on NC carriages – one stepper at each end with a common leadscrew or toothed belt drive.

Brian, I presume you want to fit a stepper in each control panel, connected to whatever they must drive, and drive them from a common stepper driver module? The issues are :

Is the load seen by each stepper the same? Stepper sizing here might be important to ensure the one does not lose steps, etc – since they would not be mechanically coupled and loss of sync is possible

How far apart are the two panels? The phase drive cables between stepper and driver should not be to long – a few peak amps easily flow so the longer the cable, the thicker the wires would need to be. Also important, since the motor drive current is pulse width modulated, ie, high-ish currents switched in square wave edges, the EMI generated is horrendous. And the longer the cable, the better the antenna, so you have a very broad band high powered electrical noise generator, connected to a nice antenna..You may need to use shielded cable..

Joe

[Brian GParticipant @briang]

Thanks

It seemed too good to be true as unless there was some kind of feedback in the driver, the only connection between the first rotor and the second would be from induced current – but I had to ask

The application was to allow a single "cab" to be controlled from multiple locations by turning knobs, a slip clutch could protect the potentiometer, so all the knobs could rotate continuously and catch up any lost steps. I really don't want to go down the rotary encoder and software route as I would be the only member who understood it, which is never a good idea. It looks like a conventionally motorised pot with +/- buttons at the remote location will be the simplest solution, and the easiest to repair.

Brian

[Neil WyattModerator @neilwyatt]

As has been said, two steppers wired together will, more or less, move in synch, just as two identical brushed motors will.

I think adding a load, as in your example, will not be reliable.

There is a simpler solution, and that is fitting a rotary pulse encoder at each station and wiring them all (two signals + earth) into a simple microcontroller that controls the output.

If steppers are driven by a constant current driver the current gets split between them, so to get full benefit you need to double the driving current, unless the intent is just to apply a more even or balanced drive (as in a prusa-style 3D printer's z-axis).

It seems that as synchronous rotation is what steppers tend towards it is very hard for two driven parallel steppers to get out of step. If you think about it, if one is moving and the other not this will tend to increase the drive current to the stationary one and vice versa. Long,long ago I was told that two brushed motors in parallel will synch speeds for the same reason.

Neil

[Brian GParticipant @briang]

Posted by Neil Wyatt on 08/10/2017 09:01:44:

…

There is a simpler solution, and that is fitting a rotary pulse encoder at each station and wiring them all (two signals + earth) into a simple microcontroller that controls the output…

I'm trying to avoid anything that couldn't be replicated without me Neil, and sadly even burning a PIC falls into that category (only one member is familiar with Arduinos and he is probably leaving soon). Might have to give it a try though, I could always leave a programmed "module in a box" as a spare.

Brian

[Enough!Participant @enough]

Posted by Neil Wyatt on 08/10/2017 09:01:44:

Long,long ago I was told that two brushed motors in parallel will synch speeds for the same reason.

Long, long ago in a land far, far away (from you) I was project engineer on extendible boom mechanisms for Gamma-Ray and Mass Spectrometer experiments on the Apollo missions.

The boom was powered by an (initially single) DC brush motor. Under certain conditions though, we ran out of "steam" with the single motor and decided to fit two motors connected in parallel both mechanically and electrically. I don't remember the nitty-gritty at this distance but we ran into all kinds of stability problems.

To fix the problem we needed to fit a number of diodes in the electrical connections to the motors to prevent the windings from "talking" to each other.

[Carl Wilson 4Participant @carlwilson4]

The problem you have using two stepper motors like this is that the rotor is a permanent magnet.

If the rotor were wound and could be excited with an Ac supply in both motors then you would have a device called a “synchro”. Used to be used a lot on aircraft systems.

[Mike PooleParticipant @mikepoole82104]

Have a read up on selsyn, resolvers and synchros these all follow each other or produce an angular error, probably more complex than the solution you are looking for but very useful on some industrial equipment and also used on various military equipment like gun turrets.

Mike

[Nick HulmeParticipant @nickhulme30114]

Posted by Joseph Noci 1 on 08/10/2017 07:04:52:

Not sure Nick's answer is answering the question though…

Yeah, I kinda read the question and having built exactly the same printer and noted exactly the same effect…………………..

[RobinParticipant @robin]

If you wire steppers in parallel, what happens when you switch on and they leap to the nearest matching position. Will the always leap the same way? What does your driver do on power up? Does it go to a set home sequence or does it remember the sequence at the preceding switch off?

[Nick HulmeParticipant @nickhulme30114]

I've kept an eye on the 3D printer and it seems to maintain sync very well unless I do something stupid.

[Ian S CParticipant @iansc]

I've got a couple of stepper motors from an old printer, and they can be wired together, and if you turn the shaft of one, the other motor follows it.

I remember years ago, Dad trying to explain the use of Selsyn motors as used in war time radar equipment, I also seem to remember them being used for remotely tuning aircraft radio transmitters/receivers. Early radio gear we used in our Cessna 180/185 aircraft were the American ARC 5 sets that had been used in RNZAF fighters and Bombers during the war (P-40 and Corsair, Hudson and Ventura, and probably the C-47). These small sets were quite large, and fitted at the back of the cabin, with the remote controls up on the instrument pannle, there was also a trailing aerial that had to be reeled in before landing. Arn't transistors good!

Ian S C

[Neil WyattModerator @neilwyatt]

Ok really simple solution, no knob just 'up' and 'down' buttons at each station working a geared motor to turn the speed control.

[Andy CarruthersParticipant @andycarruthers33275]

Rapier FSB1 used synchros, they were a pain to get to grips with but I recall they also used Gray coded wheels to give accurate positional information

I recently bought a stepper with similar positional coded output – putting two in parallel, pulsing then waiting before measuring position should give either identical results or a difference to correct, assuming one is designated master and the other slave

[Brian GParticipant @briang]

I hadn't thought of synchros, possibly because as a child I never really understood my (dockyard electrical fitter) dad's description of them, and partly because it was for driving a train, not training a gun

In the end, the most easily understood/repairable option is probably to use an ALPS motorised + friction clutch pot for the master control and push buttons for the others. Then again, they are dual-gang, and it would be fun to read the resistance of the current master one with a PIC, convert it to a digital code and drive another to the same position… I would just have to burn a few spares for after I am gone.

Brian

[MeunierParticipant @meunier]

Posted by Andy Carruthers on 09/10/2017 13:56:10:

Rapier FSB1 used synchros, they were a pain to get to grips with but I recall they also used Gray coded wheels to give accurate positional information

Crikey, haven't seen Gray's wheel mentioned since C&G Computers 'A' in '65 !
DaveD

[Brian GParticipant @briang]

Gray isn't dead yet – one of the programming exercises that came with my PicKit 2 was to create a lookup table to convert the (binary) result from a PIC's A-D converter to Gray code.

Brian

[Martin KyteParticipant @martinkyte99762]

Grey is a single bit change code so there can be no ambiguity when chnging from one position to the next. Not sure why you would want to convert an A to D binary output to grey though?

regards Martin

[Brian GParticipant @briang]

I assume Microchip to chose this exercise because there aren't many ways to demonstrate a lookup table using a push button, a potentiometer and 4 LEDs, the only inputs and outputs available on their demo board.

Turning the pot varies the voltage from 0V to Vref and the analog to digital converter returns V/Vref x 1023 in binary, which is the subject of an earlier lesson in which the high 4 bits are displayed on the LEDs. The lesson on lookup tables uses the conversion to Gray code simply as a way of demonstrating the table's function with these limited resources, the LEDs lighting in Gray code sequence as the pot is turned.

Brian

[Author]

Posts