We think of motors typically as pretty dumb devices. Depending on the kind, you send them some current or some pulses, and they turn. No problem. Even an RC servo, which has some smarts on board, doesn’t have a lot of capability. However, there is a new generation of smart motors out that combine the mechanical motor mechanism with a built-in controller. [Bunnie] looks at one that isn’t even called a motor. It is the IQ position module.
Despite the name, these devices are just a brushless DC motor (BLDC) with a controller and an API. There’s no gearing, so backdriving the motor is permissible and it can even double as a motion sensor. The video below shows [Bunnie] making one module track the other using just a little bit of code.
To show an even more impressive example, [Bunnie] put together a 2-axis robot arm using these modules and some cardboard (see the second video, below). Not only was it built in about an hour, but it is programmed by an operator moving it as desired.
The nice thing about the smart motor concept is you can tell it a position and speed and it handles all the drive considerations. This can get complex with a regular motor. This isn’t the only smart motor out there, of course. And BLDCs are common enough — you can even build your own if you like.
Hey,
To first video, I have seen something like that before :) refer to: https://en.wikipedia.org/wiki/Synchro
Problem is that a small synchro (I like the name Selsyn better :D) needs 400Hz-ish AC to function at all and they don’t have a lot of torque unless they’re huge and heavy…Raising the frequency will not help much, as it also increases eddy current losses.
A BLDC motor can give a lot more torque.
Thanks for mentioning. Was not aware of that technology, so I googled a little (using the word Selsyn) and I must say that it’s a very interesting technology! Sure it aged… but in many cases it is very interesting to know how they solved these problems in a time where electronics were non existent. Cool concept.
Regarding the IQ position module… fun concept but considering that that are a lot of variables you need to keep in mind I don’t expect this taking the maker-world by storm. Because in many cases you could solve this need using a stepper motor. These are widely available, easy to control and very very very cheap.
but if you’re making a waldo interface for a puppet/animatronic or training a robot by manually moving a simple proxy – then it could be an excellent way to do it.
Oddly I was just thinking of doing a retro piece on synchro/selsyn.
I had pleasure to work with selsyn in my technical school… This experience :)
You might want to look at this as well: https://en.wikipedia.org/wiki/Rotary_variable_differential_transformer
Stronger versions of these motors are what we need to put in hard robots! Many hard robots cannot get any negative feedback because they use highly geared motors.
It would be suitable for a mouse scrollwheel as mechanical encoders have a short life and optical often doesn’t line up with detents. then you could have an intermittent setting like windscreen wipers that scrolls indefinitely.
So make sure the optical slits line up with the detents? Or just go for a magnetic sensor.
For some reason I find this really interesting. A few questions, though:
– What’s the holding torque like? Given it’s a brushless DC motor, would it be limited by the heat that the coils in the motor can take?
– What kind of resolution is necessary for an optical encoder in a situation like that in order to prevent “slippage” between the two motors? E.g.: If they hold one, and twist the other, then let go, how accurately can the second motor snap back? To the nearest degree or two?
– How is the circuitry designed so that spinning a motor by hand doesn’t damage the circuit connected to it? Some sort of zener diode to prevent current from flowing backward?
I’m totally new to this kind of stuff, but quite interested in learning more.
Since motors often use the rotating parts and attached fan blades to move cooling air trough themselves, holding position with a locked, energized rotor is -never- a good thing.
Usually results in meltdown and a repair bill.
Cooling is going to have to come from another source when the primary motor heats up under low speed & high torque demands.
I’ve been in an electric motor rebuild shop. Locked, overheated rotors were the main repair tickets.
Since you’re more of an expert than I am, can you help explain to a newb why locking/overheating of motors can’t be something that’s avoided in software? Is there no feedback that can be utilized?
To some extent this is handled in hardware by reducing the current flowing through when it’s not moving, using a chopper of some sort. But fundamentally, if you want the motor to hold position, you have to provide enough current to oppose whatever force is trying to move it, and the result is that it heats up. You probably need to provide a fan that runs at the same speed regardless of the motor speed, or add heatsinks.
Sorry to say that my time was so far back that the Engineers were more skilled at picking a pre-heater for the breaker boxes.
The idea being that you kept the circuit breakers at just below the thermal tripping point of the motors expected current draw under a stall out. The idea being that a stall would only need a few seconds to trip the breakers, since they were already hot.
It still did a little bit of the cumulative damage that eventually melted down the motor.
This was industrial stuff though, augers, grinders and other non fine degreed parking things.
Position holding was a choice of electro-mechanical actuated braking or worm drives if any liability was involved.
DC motors had large brushes still.
If you had called something a brushless DC motor, You’d have gotten a funny look and told you were referring to a poly-phase induction motor and asked why the hell you wanted to fool with the extra “un-needed” hardware steps of power rectification.
In motor controls a lot of stuff still relied on zero crossing and/or current reversal to reliably switch conduction states.
Some stuff might have had some sort of feedback by thermal or current sensing, but this era of folks still viewed “electronic” control units with a bit of disdain and as not quiet trustworthy yet.
Another shop built a motorized bed that simply ran a non-stop pattern to facilitate the layup of an extruded product.
Think shaft screws like what we now would call a printer bed.
This “newfangled” “table” had to be run continuously for a minimum of three months (paper form. signed off on by engineer and certain other staffers) before the customer would accept it into their process line.
Damned long winded way to just say that I’m a bit out of date in this stuff, huh!
Sorry that meds for insomnia & other health issues, seem to have fogged out most of ability to retain what I once learned from the jobs.
Honestly Wish I still had more left in my cranium, to share around here.
Ohh and look into Induction heaters.
This is sort of what happens to a stalled motor if current is still fed into it.
I saw enough melted aluminum & magnesium in the repair shop to demonstrate the analogy!
Given a little gearing these are exactly what I’ve always thought should be used in 3d printers. Should give smooth 3axis movement with no microstepping or other hacks. Also with encoding no real need for limit switches and probably no missed-step problems. Also probably a heck of a lot quieter than than steppers in most jobs.
The Nema 17 has 200 steps per revolution (effectively 200 windings), while these motors have around 12. They are microstepping like crazy to get any kind of smoothness to the position, so their accuracy is likely unusable for anything like a 3D printer and I would also assume their torque is very uneven depending on their position.
Wow where to start…..
Big Boy CNCs use brushless servo motors NOT steppers. Im not saying that these motors are 100% the same quality but your assertion that they are microstepping or that their torque would be uneven reflects a serious lack of knowledge as to the nature of motion control and motors in general.
Stepper motors are driven with choppy square waves. Their 200 steps does not make them smoother….it makes them STEP….and steps are NOT SMOOTH. Lets not even get into torque stability….steppers just dont have it.
A brushless servo is driven with a sine wave….
Where a stepper jumps from one pole to another as they are switched on and off…..
A brushless servo juggles between two poles playing a precise game of magnetic tug of war…..
There is a reason camera gimbles are using brushless not stepper motors….SMOOTH.
Finally…..lets say you tell your motor to move from position A to position B at a given speed……
Your stepper motor starts its bumpy choppy slide…..and there is some binding on your rail….
Your stepper motor being ignorant to the world around it just keeps stepping even if it doesnt move at all….
Your stepper motor loses step after step until it frees and keeps going believing that it not only is and has been moving at the same speed the whole time but also that its going to arrive at its TRUE destination…..
and it could not be more wrong,
You give the same command to your brushless servo……
The stage responds with a perfect acceleration curve,
smooth buttery movement….
and it hits that same binding in the rail…..
And finally you are correct….the torque becomes uneven….
because the motor can sense that it is slowing and it adds more power and torque to overcome the resistance, maintain its speed, and course to its desination……
If it should bind significantly and it cannot maintain….you can have it pause and alert an operator….or you can have it keep trying until it makes it to its goal position……
But unlike your epileptic stepper lost on the wrong side of the bed confused in a haze of malfunction…..
The servo knows where it is….where it needs to go….and will work its best to make it happen…
SMOOTHLY and PRECISELY.
Steppers are cheap, dirty, and easy……thats all they have going for them over a brushless servo.
Mechaduino can do some things the IQ can, and costs less – https://hackaday.io/project/11224-mechaduino
He made a robot arm in a cave. With a box of scraps!!
check out coolmuscle. Ive been using their smart servos in process automation for years. You can usually pick them up on ebay for a small fraction of their list.
“The nice thing about the smart motor concept is you can tell it a position and speed and it handles all the drive considerations. ”
I could see the IC/Motor tailored to each other at the factory. Making for a more even experience with each motor sold.