Here's a good start: https://synthiam.com/Community/Forum/Search?query=Stepper&page=1

To see what many others have done - should prevent much trial and error. I don't have any experience, but all those people do. Should find something useful in there :)

Thank you. Looks like nobody has really solved this neatly yet. I am interested in steppers because I heard that they remain stiff after the power is shut off, unlike servo robots whose arms come crashing down to break everything the moment you switch off your robot. Do stepper motors really have this default locking attribute? Remote control skateboard motors auto-brake somehow as well. Also, the shafts are centred on steppers, meaning no more centralising with an artificial shaft :):):):) On the other hand, the manufacturers provide a shiny bald shaft by default that requires you to clumsily squeeze something around it lest everything slip off. Also, steppers don't need alignment during installation. They're in their starting position no matter what. Maybe I need to convert my servos to continuous rotation for this benefit. The other option is fast linear actuators (rare, slow, expensive, oversized and generally incompatible with servos) - but this would only work for tilt, not pan. For braking I've conducted hundreds of experiments with solenoids, springs and micro servos to create holding brakes but my apartment isn't exactly a lab so I've put that on 'hold' for the moment. Thoughts?

Quote:

I am interested in steppers because I heard that they remain stiff after the power is shut off,

The power to the stepper needs to be maintained to hold position. I believe the current required to hold position it determined by the rotational load being applied. The more load applied, the more current needed to maintain position. As Richard R said, the gear drive system helps determine if the position is held ( worm gear type reduction is automatically locking). I never heard them to hold position without power applied. Ron

One of my old braking ideas with a servo was to create a sparsely threaded worm rather than a threaded worm with many spirals per centimetre. I mean, think of the peg winding mechanism on the head of a guitar - the worm drives there have around 5 rotations per centimetre. But what if we brought that number down to one or less? I was never able to figure out the maths of the shaft gear diameter and number of teeth to worm spiral ratio, though. The worm drive can only be turned from the pinion gear- the worm itself is locked if the pinion is not moving. So if the worm cylinder is bearing the weight of something, it is braked at all times by default. A steep or obtuse spiral thread would bring the speed of a linear actuator up much higher if this were created. Linear actuators are often around 14 times slower (14:1 ratio) than the pinion (shaft gear of a servo). If anyone is good at maths, can you please make everyone a fast linear actuator from a standard hobby servo so we can all have the brakes that servo manufacturers can't be bothered installing by default? I'm willing to sacrifice some speed for a holding brake. Its not ideal, but its free, and it will be faster than any other linear actuator on the market.

Then add another spur gear, but this time centre the shaft. This will result in the original direction of rotation and we will have created both a holding brake (a basic one that costs speed but improves torque) and a centred shaft. Please put the final cog on the wide side of the servo motor. Then encase the whole thing in a cylinder. If you've done all that you might as well keep the wiring on the inside and just have a PWM port. You should cut off the mounting tabs first, then print the indentation in the design to secure it. Standard hobby servos are around 20x38x41mm, from memory. Cutting off the tabs will result in the lowest possible diameter.

These are the only gears I know of that fit onto Hitec splines directly https://www.servocity.com/html/servo_mount_gears.html. They may have too many teeth for a steep spiral thread but they will help you get started. If you know anyone who custom makes metal gears cheaply, please let me know too.

I don't know of a motor that "locks"'position. The only reason one would appear to "lock" when power is removed is due to extremely high gearing. Some real cheap servos will have a greater "locking" appearance because they have uneffecient plastic gears without ball bearings - but that would reduce the lifting strength and use a great deal more energy to move. Also the "locking" wouldn't be dependable.

I thunk what you're looking for is some kind of locking mechanism that can be attached to the gear, like a winch. For every click of the gear, a spring can push a tooth into a gear. But that would only work one way. Unless you devised an electromagnetic coil to hold it.

Either way, you are inventing something new. It can be done - it'd start by getting a 3D printer :)

I will take one of the "servos" apart that I use and send a photo. The "Servo" has no controller in it so you would have to control it from some other device. It will give you a good idea of how to incorporate the gearing to allow the "lock" to happen. These hold position with very little current and are extremely silent. Give me a couple of minutes to get a photo out.

I believe this is similar to what you had described above. These are the ones Toymaker also uses. Its not a great form factor unless you are designing your own robot, but if you are, they work quite well.

What control device will tell the motor stepper driver to move? This controller should also be able to operate an electromagnetic brake. A solenoid commonly used on paper printers for feeding, could be used to release a mechanical gear brake. These solenoids can develop a reasonable amount of power. Using a pawl against a spur gear on the motor shaft, held by a spring, could allow quite a bit of holding power.

I know this makes things a bit more complex but if you don't want to use a worm gear drive (keeping it simple) this may be a way to accomplish what you want.

Just another thought.

Ron

CochranRobotics, what kind of motor is that? It appears to have a lock :)

Andy Roid, I already know (after a lot of effort) two ways to activate an off-brake at exactly the same time an individual servo moves in such a way that a servo's non-movement reapplies the brake. Linear actuators that I've found nearly always have a shallow spiral and are therefore extremely slow.

The 2 methods I have figured out to build an electronic holding brake (rather than a spiral worm holding brake) are described below.

Let me first say that I spent weeks messing around with soldering these RC switches from pololu (not recommended) https://www.pololu.com/product/2803 and some RC switches like these (this kind is fine for switching LEDs and solenoids ON without any fade capabilities - they just stay turned off until rotation reaches something like 90 degrees clockwise rotation - which is fine if you want to switch a non-PWM device such as an LED on using a PWM input) http://www.ebay.com.au/itm/like/261684396842?lpid=107&chn=ps. I spent ages on it.

I racked my brain to figure out how to simply make the solenoid (a little pushing pin in a magnetic coil with a spring like this http://www.ebay.com/itm/2pcs-DC3-12V-Push-Pull-Type-Solenoid-Electromagnet-DC-Micro-Solenoid-ACM-/361406950245) pull away from the wall the very instant the servo was given a pulse using RC switches. Solenoids theoretically would be the correct braking mechanism, but they are extremely weak when they are small. The magnetic strength depends on the number of coils of insulated copper and the current that pulls the pin against the spring. I even bought rolls of insulated copper wire to try to make them myself. That was a nightmare.

I originally used teeth in a rotation wall around the servo that the solenoid pin could lock into, but it takes more power to pull past a tooth than I thought. Then I decided that the the solenoid should friction-press with a rubber stopper like a drum brake instead. The problems were twofold: The magnet couldn't be strong at hobby servo scale and so the spring couldn't be strong. But... to save all of hobbykind from the torment of dealing with RC switches that only activate after 90 degrees rotation etc, I will tell you how to INSTANTLY activate a device the EXACT moment the servo receives a pulse. This means I also solved the LED fading issue as well: Ready? This took me ages. But its bloody obvious after you hear it. This is how to activate an off-brake/holding-brake the moment the servo receives a pulse:

Attach the device to the wires of the DC motor inside the servo!

After all, the motor only receives a signal when the electronics in the servo tell it to move. So there's that massive problem solved. I advise anyone wanting to build robotic switches to contemplate hacking servos before you search google for 7 years reduced lifespan worth of stress. (That's how I measure the ease of a google search - by the length of time it will take off my life to find the solution).

Bear in mind that the brake will only release as gradually as the speed of the servo motor if you solder the off-brake solenoid to the DC motor inside the servo, but at least its both instantaneous and compact. A servo manufacturer could tweak this idea into a single servo with attachable brake as a single product. I was going to, but I'm broke.

So I've mentioned the mechanical gearing solution using a standard hobby servo with a low thread count worm drive as a holding brake system. I've mentioned solenoid 'electropin and spring' holding brakes activated from the servo's motor. I've mentioned the powered auto-brakes used in remote control skateboards (more research required). Now I'm going to tell you something similar to the solenoid but more controllable.

A servo-spring holding brake. This is not an automatic brake and takes up quite a lot of space, requires brake-release programming, and will probably double the size of your robot. But its the only reliable thing I have figured out to reliably work. Simply allocate a nano servo (if its strong enough) to every servo in your robot and set up a spring that forces a rubber stopper into the housing wall of your tilt or pan segment. Then program the nano servo to release at the same time as your movement servo wants to move. You need to program the nano to push back to brake, even though the spring will do that automatically when power is cut. This will reduce jitter in all robots, allow them to hold their poses using minimal power, and stay in position when power is cut. I wasn't going to tell anybody because I wanted to build it first then show people, but I think a lot of you guys have built a lot of advanced stuff and will actually help me out by refining a brake that I could use. Anyway, there's the main answer to hundreds of hours of unpaid research. ;)

I hope one of you builds a one-size-fits-all holding brake solution for standard sized hobby servos that is upscalable for stronger holding power. The whole effort is in making it compact. Also, I highly recommend friction grip rather than teeth for several reasons.

look at automotive motors used for climate control door flaps. These are Bosch ones for Audi's, but other manufacturers may also work for you.

I have tried to find a video to see how fast these blend door actuators move and how loud they are, and what wiring is required to move them in either direction. I assume I can connect one to the motor terminals inside a servo and control them with the EZ-Robot program if the voltage is the same. Its a pity they're so large - too big for most of my projects, but exciting to know the engineering to brake a motor has already been devised. Do you have a video that shows maximum speed and noise? If its only as fast as a standard linear actuator then there's unfortunately no benefit to using them. They look slow because the spiral thread count on the worm gear is high.

Edit: I just looked at your first comment and read that you mentioned toy actuators. Here is a toy that is very small and still works fast:

https://www.youtube.com/watch?v=VWRgi4uuxEU

https://www.youtube.com/watch?v=xoLxWgtntjE

Arms up and down, and neck left and right are using these motors.

Fantastic design, David. Did you use a 3D printer for the housing? Also, why didn't you choose servos - was it because you wanted brakes or were they less noisy? How small are your motors? Where did you get them from? What is causing the head to wobble?

https://plus.google.com/u/0/communities/115080096051481641499 has many of the answers you are looking for.

I wanted positional information back from these motors. This is important in the design as the arms need to be in a certain position to pocket themselves. The motors are very strong and durable. I also wanted motors that draw very low current and are silent. The "locking" type design was also desirable.

The motors are about 2"x2"x1" or about 5cmx5cmx2.5cm without the mounting holes in the 3 corners or the housing around the plug in.

Ebay has these motors. You can get after market ones but they normally have plastic gears for the worm gear which I don't like. They strip out in time.

http://www.ebay.com/itm/AUDI-A6-HEATER-AIR-VENT-FLAP-MOTOR-REGULATOR-0-132-801-129-4B1-820-511-C-/272163433606?hash=item3f5e346c86:g:WnYAAOSwwpdW4Oos&vxp=mtr is an example of one of the motors on ebay. These run at 12v to 24v. I run mine at 12v.

On the head wobble, the video shows the first design. This was using one of these motors mentioned for head up/down movement. There was too much weight in the front of the head in this design for this motor. It worked but needed springs to make it wobble less. I have since redesigned this to use an actuator motor for the up and down movement of the head, which eliminated the wobble.

The movement on this one is very loud (robot moving forward and such) due to me using a ball caster. This has been removed with an omni-direction wheel to make it much quieter.

Great Rafiki design! I like the idea of these motors. Do they plug into a PWM slot? Do you know the torque? Speed?

They have to be controlled by something else. There is no onboard controller, just the motor, gears and pot. I use arduino mini's to control mine. The mini takes serial commands and moves the motors to the desired location, but also reports back the location of the motor. I can query the arduino for the current location of the motor also. The Arduino drives something (sabertooth in my case) but could drive an LBridge or some other type of driving unit. LBridge would be the cheapest route to go and the arduino does all of the heavy lifting logic anyway so a sabertooth is really overkill for the purpose of driving these motors.

Thank you. And how much torque and speed, do you know?

I don't have specs. it works great for my needs. How much weight are you needing to lift? Speed, how fast do you need? Get one, try it out and decide if you want to use it. It is much stronger than any servo I have seen.