Hello, I am finding it surprisingly hard to source motors and I 'm wondering if I 'm missing something obvious. The application I have in mind is motor control of a camera lens by spinning the rings. The numbers I have seen for turning a ring by putting a pencil against a scale and pushing are approximately 27 grams at a distance of roughly 20mm from the lens central axis. If I have done the math right, this is some 0.0052 Nm. So, I am looking at a fairly modest (read, small, low power, high internal resistance and cheap) BLDC motor to drive with a SimpleFOCMini, shipping from the EU.
So, this is where things get weird, I can’t find that many small, low power, high internal resistance BLDC motors shipping from the EU. If I google for gimbal motors (the closest thing I can think of to filter out the obvious motors meant for driving propellers), I generally get big and chunky ones or results shipping from China. I could, of course, order from China but I 'd rather avoid the delay. I could also try using some small, low resistance motor meant for spinning props (there are trillions of these to choose from) but I thought that perhaps I should take it easy and not start burning SimpleFOCMinis on a first date. I suspect, however, that I am searching wrong and there’s other search terms I could use so, does anyone have any idea where and what I should be looking for and where to find it?
Sourcing motors can be difficult 
Generally speaking my observation is that some motors are “standard” products, intended for use in RC models, quadcopters, gimbals and similar applications. Some are designed specifically as robot actuators (expensive!).
These items you can order from many shops, including Amazon and AliExpress, and I’m sure you’ve seen many of the available models.
Other motors are embedded in products, and they are often made to specifically fit the product. You’ll find the parts and samples for them on AliBaba and AliExpress, and the specialized companies that make them. They may send you a few samples, but generally they’re not interested in you unless you want to buy 1000 pieces or more. Using their products is much more difficult because the expectation is that parts of the solution will be designed into the product on your side.
There’s also standard motors for industry, which you’ll find if you search in the professional space. Companies like Maxon, Faulhaber or Portescap will probably have a motor for you - at $200 or more for small quantities 
About small Chinese motors, you might want to have a look at: Miniature BLDC motor for laser scanning head - #25 by quentin
How about small stepper motors?
I’ve savaged old DVD/Blueray Players which are cheap on ebay.
Their head drive is pretty accurate.
Just found a small and crappy video in which I demonstrate how to build a small Delta printer with them.
(It not a printer, but a coffee-stirrer gimmick)
Thank you all for the suggestions, I will look into them all.
In answer to why not using stepper motors, the main reason is too silly, the only thing currently at the shop is the SimpleFOCmini so that’s what I can get my hands on. Having said that, the more philosophical answer is that control of BLDC motors in this manner is more exciting and new and I am here mostly to goof around with a new exciting control toy which, I hope, will replace servos in many applications in the near future. Steppers are an obvious but old-school choice for such an application. I 'd much rather fail while trying something novel than succeed in a manner I could have anticipated five years ago.
Along those lines, I am very curious to see if making use of cheap small motors meant for propulsion (whoop and such) is likely to become a servo/stepper killer in the next few years. Currently it’s still very early days but I am wondering if we will eventually see the stepper-with-integrated-controller become obsolete because of the foc-controller-you-can-screw-to-the-stator-of-a-bldc.
What is your use case? I’m a little confused about the end goal here.
Cheers,
Valentine
Well, @Valentine , if you put it like that, the obvious answer is finding a largely acceptable excuse for spending a modest amount of money to expose myself to a DIYers implementation of a technology that is currently becoming increasingly mainstream.
Regarding the excuse itself, well, the use case I have set myself, simply because it happened to be what I was recently playing with, is turning a conventional camera lens motorised by somehow interfacing the rings. I tried servos and it’s doable but awful (the necessary gearing makes the result either too large or too complicated). I have set myself the fluffy targets of “not too expensive”, “easy to source materials” and “low complexity” for any experimentation. Not sure what will come of it but I 'm interested in seeing where the ride takes me.
My question is, what are the specs of that nebulously fluffy mysteriously tiny motor?
- Dimensions
- Voltage
- Current / resistance
- Torque dynamic/stall
- Sensoring
- Mounting
- Cooling
- Price
- Does it purr when petted?
Cheers,
Valentine
Well, now we are getting in nebulous territory indeed, so let’s start from the use case of motor control of the lens rings from my first post. The rough requirements are:
- Move the ring(s).
- Don’t burn or break if you hit the movement limit.
- Provide enough precision to be usable.
- Don’t be huge and/or expensive.
Let’s flesh these out a bit more.
We need to turn the rings of the lens, which requires a torque of 0.0052 Nm if I 've done the math right in my first post. However, this is not torque of the motor, it’s torque on the ring. The connection can be geared and the ratio will have to do with how close is placed to the ring and how it is coupled (gear, belt etc). The speed of rotation is not terribly important, the use case requires fine control and modest speeds so it is conceivable that a tiny gear and many revolutions on a somewhat slow motor are perfectly fine. My thinking is that I will likely start with very slow speeds and see about speeding up.
The size is kind of important. I can fit three typical 9g servos, which are around 22mm x 12mm around the lens. It would be great if I can be in that ballpark, especially since the smaller the motor is, the closer I can get it to the lens ring and the smaller I can make the gear on its shaft, if I decide to go for gears meshing directly (which is the simplest approach).
Price. You can get a cheap and cheerful small servo for single digit dollars / euros. Obviously, this won’t be the case here but I would prefer not to go too crazy. An upper limit limit of around 30 euros per motor seems like a reasonable value.
Now, given the above, I have been looking around the threads suggested above and in this thread, @runger mentions the EMAX RS1106.
Now, if this is actually usable, it would be beyond wonderful. It is tiny, easy to source and so crazy cheap it can actually compete with a servo (well, price of the simpleFOC notwithstanding). However, I have no idea of its phase resistance and I 'm fairly certain it will be super low, in all kV versions. Now, I am planning on using the SimpleFOCMini V1.1 and its product page says in scary letters that:
SimpleFOCMini has been designed for gimbals motors with internal resistance around 10 Ohms. Using this board with high-power drone motors may burn the DRV8313 chip integrated on the board. The chip can handle up to 3Amps of current for short periods of time, but in for long-term operation (without the heat-sink) the current should stay under 2 Amps.
It is possible to use this board with low resistance motors but it requires a bit of experience and it is not recommended for beginners. Ask community about it!
So, if using the EMAX RS1106, I will almost certainly be operating with a wildly lower resistance motor than the SimpleFOCMini expects. I am not terribly sure how much experience would be required to “use this board with low resistance motors” as the note says but I would be willing to give it a shot, given the other characteristics (size, availability and price) of the EMAX RS1106 motor.
How do you like the look of these?
Or these:
Both of these look amazing, especially the ht1105 which includes an encoder for a very reasonable price. However, there is the drawback that they ship from China.
So, right now, it looks like this:
- EMAX RS1106 or similar. Very cheap. Very easy to buy from EU. Bad phase resistance (too low).
- GB1105 or similar from same manufacturer. Not too cheap but OK. Ships from China. Good phase resistance (12.5 Ohm).
- ht1105. Not too cheap but includes encoder so price actually quite good for combo. Ships from China. Good phase resistance (8 Ohm).
Based on price and availability, I would lean towards low resistance EMAX but am not too sure if this is a good idea and how easy it is to make it work. To begin with, my power demands are pretty low so I could perhaps get away with something as silly as sticking some resistors to the leads but I have no idea what is involved in making a low resistance motor work, i.e. the bit in the product page that says:
It is possible to use this board with low resistance motors but it requires a bit of experience and it is not recommended for beginners. Ask community about it!
So, does someone in the community have any input on this?
I think it’s doable, but as a FOC-newbie I’d start with a regular gimbal motor setup to wet your feet.
If you are an experienced programmer you will get the hang of it pretty quick.
To tame such a beast you need to slowly ramp up speed.
That can be done in the loop and/or by setting lame PID values.
I’m not sure you understood my original question. Using a drone motor for highly accurate motion control such as lens adjustments is not going to work. The reason you need high ohm is not because you need low current. The low current and high ohm resistance is just a byproduct. You need a zero-cogging motor with high pole count, high bemf and slow angular speed. Here, check this out, dirt cheap, four for the price of a big-mac meal, this is a 6mm diameter motor, I believe 6pp:
https://www.amazon.com/Happymodel-KV19000-Diameter-Brushless-Quadcopter/dp/B07BSHRND4
The person on the picture definitely needs some high quality mani-pedi work done because his thumb-nail looks disgusting but that’s tangential.
That tiny thing develops max power at 100k rpm. You run this at low RPM, you will burn it immediately. Can’t squeeze blood out of a turnip.
That tiny drone motor won’t work, period, don’t waste your time. Your only option is a custom built motor using a swiss cnc mill, at least 15pp or more, diagonal magnets or poles to remove cogging, and manual winding with 0.05 mm copper. You will need a microscope to wind it.
All that said, nothing stops you from experimenting with drone motors, that’s how you gain experience.
Cheers,
Valentine
@Valentine , OK, I see your point about the dirt on the guy’s thumbnail but let me point out that, beyond that, it is reasonably well trimmed and symmetrical. I have to shamefully admit that my own fingernails have often been in much worse shape on all counts.
Now, I think I am slowly learning things on this thread so please bear with me. I am kind of guess-realising that there is a physical limit to how easy it is to wind small things. So, when making small, cheap motors, it is easier to make them low resistance and high kV. On top of this, there is the point that there is a lot more demand for small multirotors than there is for small gimbals and that small multirotors typically require high RPM anyway so these motors suit them. Therefore, it is unlikely we will see low RPM / high RPM motors competing on price and size with their propulsion cousins. Have I got that right?
Now, a major reason I am dragging my feet and posting on the forum is that the wonderful motors @runger mentions (shipping from China) is that they give a delivery time of a month and a half. The EMAX (and a bucket load of similar) can be on my doorstep in a day.
However, I did find this GB2208 shipping from Germany. It’s a bit too chunky, stronger than required and on the pricey side but within the envelope of “not great, not terrible” on all counts. I think I will try a couple of those since they look quite promising.
Mine too but I don’t go posting them for all to see, right?
There is no market for these motors and also when needed they are custom and purpose designed for very specific cases by Kodak, Cannon, Nikon, etc. other lens manufacturers, or watch manufacturers. Yes hybrid watches have microscopic BLDC motors inside, mine has like six of those (Citizen Exceed Satellite GPS) pancake motors squeezed into a tiny space.
No, you won’t see those tiny ones because there is no market. You my have some OEM for cameras and lenses but who fixes cameras these days? If I drop my lens, I thrash it and buy a new one since the alignment is gone anyway, the motors inside are last concern.
That said I’m not saying you cannot find ones but from time to time this same question pops up here, and so far I haven’t seen anyone coming back with a good source.
Good luck nd please come to report your findings.
Cheers,
Valentine
I think we can differentiate between “maker projects” and real products, although certainly sometimes there is crossover.
So I think for one’s personal project, certainly it can sometimes be about pushing the envelope, but sometimes it can just be for other reasons like learning or just because.
So I don’t really think you need a custom BLDC motor with diagonal magnets (even if it would be better). I think you could take a crappy servo, change it over to continuous rotation, add some more gears and see how it goes… going for a BLDC is anyways going to make it a cool project 
Real motorized camera lenses anyways use different motors like ultrasonic piezo motors to achieve their super fast movements in the narrow space around the lens.
I do think, like Valentine, that the 19000KV drone motors won’t really work for your application, and will be hard to control with FOC due to their high speeds.
But I think any kind of gimbal motor will work from the electric and control point of view, and it’s just a question if it will work mechanically in terms of friction/torque.
I will report my findings as soon as I have some. I ordered my simpleFOC minis at the end of last month but still haven’t received any notification and the shop now shows them as out of stock, oh dear.
I have some old motors lying around like the 1000 kV version of a SunnySky X2814 which should at least get me started, then I 'll see what my next steps will be.
Only the very high end prosumer lenses use USM, the consumer grade use either micro-dc motors (old-fashioned brushed) or micro bldc motors.
The tiny bldc may work with gearing but that’s again a rabbit hole, since this gearing must be high precision.
I’m very curios what comes out of this. I’ve never seen a hobby auto focus lens design.
@dabbler please come back and show us what you got, this is interesting. I would imagine a PTZ with autofocus hobby effort a very good learning example.
Cheers,
Valentine
Gentlemen I promise I will let you know when I have something. For the time being, all I can do is wait for my simpleFOC minis so I am spending time in idle chat. I can give you some background if you want. The camera is the Raspberry Pi High Quality camera and the lens is the zoom lens for it.
It’s a neat combo because you can stick it on a tripod, power it up in range of your wifi and have a totally controllable and programmable camera (closer in concept to photographer camera rather security camera) without having to be physically next to it. For night photography, not having to freeze next to a tripod is considered to be a Good Thing in some circles.
So, if you can muck about with the camera settings from your computer, the next obvious next step is not to have to muck about with the tripod and lens. Effectively, making it into PTZ but not just Z but also A(pperture) F(ocus), so PTZAF.
I thought about getting a gimbal but where’s the fun in that and also I didn’t want to spend some stupid amount of time figuring out what gimbal would give me what I want, especially since I have no idea what exactly it is I want.
So, to basics. Servos are dime a dozen, I have a bunch lying around and they are so easy to control that even the RPi can control them without much fuss. The PT part is doable with servos. We are not talking about a gimbal meant for counteracting the vibrations of walking or a vehicle but simply pointing a cam on a tripod. This task can be given to servos. The lens rings however, are trickier. A typical servo gives something over 90 degrees of travel, say 100 degrees or so. The rings of this lens want somewhere in the range of above 100 but under 180, say 140 or so. It is easy to 3D print a toothy ring (or groovy ring for using an O-ring as a belt or whatever), slip it along the lens to the desired ring and secure it with the ring’s screw.
The hard part is interfacing it. In order to make 100 degrees become 100something degrees you need a gear on the servo of a somewhat larger radius than the lens. This is awful. The further away the servos go, the bulkier the whole thing becomes but, most importantly, it becomes jigglier. It’s a lot easier to make it stiff with the servos as snug as possible against the lens but the further they go, teeth skipping becomes a real PITA. It would be possible to do some fancy gearing by introducing more gears but then the design becomes nightmarish, especially since these gears will need attachments for their axles etc.
Well duh, is the answer, use continuous rotation servos. Well, yes and no. With normal servos, you can hide feedback under the rug by making the servo travel range coincide with the ring travel range, which is why I really wanted to be able to use them. With continuous rotation you need to be able to somehow provide and then parse and manage feedback and the RPi is not great at that. So, introduce a small Arduino or RPi Pico to handle that maybe? OK, if we introduce another board, we might as well go for continuous rotation with feedback such as the Parallax Feedback 360°.
But now we have strayed from the original vision of simple, cheap and easy to find components (standard servos and 3D printed parts). OK, if we have mangled the “simple” and “easy to find”, let’s try not to stray too far from the “cheap”. Sure, you can build something if you are willing to spend more money than the cost of the camera but that is kind of idiotic. Enter simpleFOC. The cost of a continuous rotation servo with built in encoder is in the same ballpark as some simpleFOC controller plus something or other to use as an actuator (sourcing of which started the original message). However, a rough calculation shows that the FOC coolness factor is roughly two orders of magnitude higher and the possibility of electroboom style shenanigans adds some colour to life in a manner that a continuous rotation servo plus encoder could never aspire to.
Which brings me to today. I will likely build the PT part with servos and hope that my simpleFOCminis will hopefully materialise at some point. If they don’t, I might try my luck with the Parallax servos instead.