The easiest way to run wires through a rotating shaft?

[Paul Nicholson]

I found this closed thread and I'd like to continue it: https://www.physicsforums.com/threads/design-question-wiring-motors.459672/

I am building a rotating ceiling mount (or rotating a floor plate) and am trying to solve the problem of the continuous rotation servo motor getting in the way of the wires that need to pass through. This would be exactly the same issue as placing flashing LEDs on fan blades - how can it be done easily and cheaply with links to the parts I need?

Hollow core pancake slip rings seem like the logical way to accomplish the task of running wires from a rotating base to a stationary robot. Please see my design. But they are very expensive. I am developing a product and need to find a reliable store that sells exactly what I need. This is a minor detail of the robot so I would rather the entire mechanical design already exists and I can simply buy it. Engineers and fabricators cost a lot.

My second idea for how to run wires through a rotating shaft is to put the motor to the side of a hollow artificial shaft and let the wires run freely while using a belt, although this would still result in a twisted wire, and is extremely difficult and time consuming to build with a tensioner spring and all the rest of the measuring efforts to find suitable belts and pulleys etcetera. I really do want the robot to be able to rotate continuously in the same direction as long as it needs to for clothing fashion windows. Rotating 180 degrees is as easy as providing some wire slack, but this would not be desirable.

I am not after concepts (unless they are different from the two I mentioned above), but rather the exact parts I need to build a rotating system to attach to my motor, or a completely intact mechanical system I could easily buy for less than $100. The last thing I want is to spend thousands of dollars on Research and Development for this considerably simple task. A hollow core slip disc that allows for at least 50 Amps and a couple of data lines for less than $100 would be the best, I think.

Thanks in advance for any product links or steering column modification ideas, new concepts etc you can provide.

 

[Paul Nicholson]

I'm not sure if this forum is busy or not, so for my own records I'll post this note about what I discovered today:

Ceiling fans with lights in the middle have achieved what I want already.

After that I found this garage door opener, which does not have continuous rotation nor a potentiometer that I purchased for $210 plus the safety sensors for an extra $90. I'll see how it goes. I hope you guys have some more advice because this still isn't quite right:

 

[Mech_Engineer]

The easiest way to run wires through a rotating assembly is to use a large enough bearing where the inner diameter can fit around a hollow conduit. This is how it is achieved in a ceiling fan for the light.

See here:

 

[Paul Nicholson]

Thank you, Mech_Engineer - great image. I know this is not an electronics forum category, but now that I know about these hollow motors, and I have never seen hollow servos, does anyone know how I could safely wire the guts of a 6v hobby servo into one of these things so I can control it with a potentiometer? I now own the garage door opener (pictured above) which will not rotate continuously because of it's circuitry. I don't mind ripping it to pieces or buying more hollow core motors. I'm not sure if it has a transformer inside yet but compared to the RC motors I've been using, this is mega powerful, and would probably have enough torque to lift a grown man like a hospital bed crane so it might be running on 240V AC. I just need everything to be controlled by my microcontroller to make the robot dance and pose and do martial arts all at human speed. The garage door opener force level (torque) can also be programmed for safety, but this does not stop it moving - it reverses directions (to open the garage door). I could post more circuitry info here if this thread can continue in this mechanical engineering category, otherwise I could create a new post in electronics.

 

[berkeman]

6v hobby servo into one of these things so I can control it with a potentiometer?

Are you familiar yet with how to use PWM (pulse width modulation) of the power to a motor to control its speed? You don't generally use a potentiometer directly to control motor speed.

 

[Baluncore]

Search for 'Slip Ring' on a popular auction website and you will find many Mini Capsule Slip Rings below US$10.

The fascinating anti-twister mechanism is a concept rather than a product.
https://en.wikipedia.org/wiki/Anti-twister_mechanism
http://shashanksharmaportfolio.weebly.com/uploads/2/3/0/0/23007620/1.7.pdf

 

[Paul Nicholson]

They don't sound like they would meet my requirements of 50A minimum.

Are you familiar yet with how to use PWM (pulse width modulation) of the power to a motor to control its speed? You don't generally use a potentiometer directly to control motor speed.

Hobby servos that use Pulse Width Modulation generally come with three wires: Red, Black and White.

For a start, I'm pretty sure that 100% of the power supply's voltage is always going into the motor via the red live cable and remaining steady at all times. You can clearly see this with hobby servos because they have absolutely no holding brakes whatsoever and when power is cut the robot collapses like a ton of expensive bricks. I will discuss braking systems in a separate thread later..

Secondly, the black ground wire completes the circuit.

Thirdly, the potentiometer (a series of tiny lines on a cylinder underneath the active shaft of the motor box [connected to the actual motor via gears] that are detected by a sensor that counts the number of shadows that have passed a small light) tells the motor's CPU the current rotation position so the servo motor's circuit boatd knows how far to rotate to get to the position instructed by the pulses it receives from a controlling device, even correcting the motor to wind backwards if it goes beyond it's target cylinder notch count because it has moved too fast or it's load is too heavy (often resulting in overheating and jitter because there is ridiculously no braking system at all except when thick, heavy steel gears are used).

And fourthly. the white data pulse wire transmits a low voltage pulse every 20 milliseconds, and varies the length of this pulse as the signal sent from the controlling device (ie. microcontroller, remote joystick, myo, epoch mind reading headset, kinect sensor via PC etc) tells the circuit board connected to the motor how many shadows of the potentiometer cylinder need to be traversed to get to the target rotation point. The potentiometer's movement is activated by the motor, so the motor rotates to achieve this.

Unfortunately, time consuming modifications need to be made to most servo motors for the potentiometer position to be sent from the motor circuit back to the controlling device because servos are often made as cheaply as possible. With this motor to controller signal, motors can simply be posed into position manually for extremely fast choreography (frame posing to create a sequence) that is generally MUCH faster and more reliable than any other existing manual methods (rather than automatic sensor-based or artificially Intelligent methods).

Also, servos need to be cylindrical rather than rectangular with a centered shaft on one side and a stabilizing shaft on the other, include a slip disc pass through via a hollow core, and include a fourth motor-to-controlling-device wire for easy positioning of robotic tilt and pan segments. They should also include continuous rotation every time but with a mechanical holding brake system. They must be as small as possible, robust, light, extremely quiet and include a cooling system or not overheat. In general, servo motors are made very badly for nearly all applications at this point in history. If I modify this motor I have just bought, I want it to do all the things I mentioned, most of which are vital.

But getting it made quickly is even more vital at the moment so there is really no time for research and development, and all of these solutions already exist cheaply in one form or another. The objective is to find the best links on the internet for places to purchase the cheapest quality components that meet specifications, or find links to the most user friendly instructional videos on how to assemble these ideas for free easily.

 

[Baluncore]

Why do you need 50A ?
What is the voltage ? AC or DC ?
By increasing the voltage you can reduce the current. That is what transformers and high efficiency switching DC to DC converters are for.

If the 50A was peak current then to provide that peak, you could use a local battery on maintenance charge through lower current slip rings. Slip rings can be wired in parallel for higher currents. There are also rotary transformers capable of transmitting AC power.

Once you have a DC supply reaching the rotating part your problem is solved because you can run a local controller on that deck. There is no need to run signals through slip rings, a bluetooth connection is more than capable of communicating all the control instuctions.

You might consider a "lazy susan" rotation bearing. With the servo motor offset from the bearing you might use a stepped belt or bicycle chain and chain wheel to drive the rotation. One advantage is that the gear ratio of the drive increases motor RPM, so it reduces the torque and current requirements.

If you had bearings supporting both top and bottom you could supply power through both those end connections.

 

[Baluncore]

Hobby servos that use Pulse Width Modulation generally come with three wires: Red, Black and White.
For a start, I'm pretty sure that 100% of the power supply's voltage is always going into the motor via the red live cable and remaining steady at all times.

I thought the PWM RC servos low pass filtered the PWM signal to produce a voltage. That voltage would be subtracted from the position potentiometer voltage to give the position error. The error voltage is then used to generate the digital control to an H-bridge supplying the DC servo motor.

There is usually no need for a brake with servo motors. The average current through the servo motor is only proportional to the torque applied to counter the load. It is usually only safety issues that require a mechanical brake or lock. The mechanism should be designed to be counterbalanced by mass or through springs and linkages. Take a look at an anglepoise lamp.If the mechanism is balanced it will not collapse.

Only a crude bang-bang controller will use full power all the time. PWM derived from a simple PID controller can minimise settling time and power consumption.

 

[Paul Nicholson]

I need 50A because of the current used by all the servos in the robot.

I was originally using these Power HD 1235MG servos for the shoulders:

These 7.4V analog servos have a holding torque of 40 kg.cm, move at human speed and cost US$60 each. But they sound like a chainsaw, incompatible with most servo horns and brackets and are quite large. They require 9 amps to hold their maximum force / change directions quickly before they catch fire.

At the moment I'm experimenting with these Hitec 9380TH servos instead:

These 7.4V digital brushless servos have a holding torque of 33kg.cm, move at human speed and cost US$165 each. They are quieter but massively expensive. I'm not sure how many amps these need to hold their maximum force / change directions quickly (I'm guessing 3-4amps) before the catch fire.

I am unsure about the hips, neck to wall connector and lower back at the moment, but I am likely to use these Invanscience i00800 Torxis servos in a couple of places:

These 12V analog servos have a holding torque of 57kg.cm, only move at half of human speed and cost US$290 each. They can be very loud but they are able to hold their position forever - there are no brakes but rather they use heavy duty steel gears that won't move without a pulse signal, even if power is switched off. This also means they never jitter or overheat. Unfortunately they are as big as a carton of 200 cigarettes so there is no practical use for them. Also, the horn is held to the shaft really badly and simply slips off when hung from the ceiling without support. I originally used these for the ceiling plate but the entire robot slipped off the shaft and smashed to the floor. They are incapable of allowing a slip ring to pass wires from the wall socket to the robot, no matter what kind of housing is built, so are limited to about 200 degrees rotation as a base/wall/ceiling plate.

I needed a motor to allow the wires to reach the robot. I eventually found this Chamberlain CR600 for US$190:

These DC gearmotors probably have an AC to DC transformer but I haven't dissected it to check (I use a 240V wall socket). They have a holding torque of 204kg.cm. That's MUCH higher than the DC motors inside 7.5V hobby servos. I may want to put one of these inside the robot for the hips at some later stage, at which time I may need a lot more than a 50A slip ring. I am going to test using this as a rotational ceiling/base plate. Ceiling plates are best because it leaves the legs free to run without a large amount of floor space for a 150cm arch support base. These garage door openers contain circuitry that stops them from rotating after around four rotations, which means that circuitry is an obstacle. They also do not have a potentiometer and servo circuit connected to them. But they are hollow, so a slip ring can be used easily to get wires to the robot without twisting.

Now that I know about hollow core DC motors, I am tempted to find a bunch of them with the torque and sizes I want, then add servo circuits, potentiometers and brakes to them myself (if its easy).

The microcontroller is an EZB-V4 that will be powered separately from the motors to prevent overloading:

It can handle up to 20A and needs at least 7V and at most 16V to operate. Most motors in the robot need 7.5V, but many of them catch fire at this voltage. Most small servos in the robot can handle 4.5-6V. Therefore there will need to be voltage regulators for all the different motors and for the microcontroller. 6V, 7.5V and 12V, plus 240V AC for the base plate. All power after the slip ring will be DC.

I would have used a lazy Susan (actually called a thrust bearing) if the garage door motor didn't already handle this. Who knows yet - it'll probably brake off if I don't rebuild it more robustly first.

"You might consider a "lazy susan" rotation bearing. With the servo motor offset from the bearing you might use a stepped belt or bicycle chain and chain wheel to drive the rotation. One advantage is that the gear ratio of the drive increases motor RPM, so it reduces the torque and current requirements."

Its easy for you to say that because its the first logical solution. I've spent HUNDREDS of hours searching for the right parts and wound up in hospital from the stress of it, so if you don't have links with the best prices including delivery options to Australia then its counter-helpful to me at this point in development. Sorry to say this again, but its all about the product links, images and videos. I think this rule is true for every commenter on every forum. I have driven all over Sydney trying to find belt and pulley systems as a counter idea to finding a hollow core motor. I would rather NOT build a belt and pulley system with tensioner etc for the next 12 months if I could just buy one or something different. If you know how to build one properly and easily with step by step instructions then I'd love to know, but every link to every part must be available or it will eat at least 6 months out of my development time.

Here was my belt design but I've worked with belts before and they are a nightmare so I would vastly prefer a hollow core motor with slip ring if its easier and cheaper to prototype. If I can buy a belt and pulley system that is cheap and works well and is already assembled then that would be a viable alternative to a hollow core motor. The garage door motor has complicated wiring and it will probably be difficult to attach a servo circuit and potentionmeter to such a system without getting electrocuted. I'm still open minded about a belt system if I could source cheap parts with proper links to buy them. Australia doesn't sell many things. We must import everything or have it fabricated. This design of mine will cost $3000 even though suitable belt and pulley systems exist in a lot of places that I don't know about - I'm just hoping someone can give me clear instructions on where to find them and how to make it work with what I'm doing cheaply and easily.

I thought the PWM RC servos low pass filtered the PWM signal to produce a voltage. That voltage would be subtracted from the position potentiometer voltage to give the position error. The error voltage is then used to generate the digital control to an H-bridge supplying the DC servo motor.

There is usually no need for a brake with servo motors. The average current through the servo motor is only proportional to the torque applied to counter the load. It is usually only safety issues that require a mechanical brake or lock. The mechanism should be designed to be counterbalanced by mass or through springs and linkages. Take a look at an anglepoise lamp.If the mechanism is balanced it will not collapse.

Only a crude bang-bang controller will use full power all the time. PWM derived from a simple PID controller can minimise settling time and power consumption.

You're probably right. I have no idea how a servo motor works really. I don't care to know. I just want to know how to get this robot plugged together within the next few days before I end up in hospital all over again. If I can rip the guts out of a servo and hook them up to any dc motor safely and EASILY with INSTRUCTIONS, then I might do that, but I'm not putting springs through the robot at this stage. I have a pile of abandoned spring robot frames in the cupboard that are the stuff of logistical nightmares. Servo motors can be made into tilt and pan segments. They need holding brakes, and while I'm quite certain this will incorporate small springs, balancing a frame through 3D printed parts is something I will not revisit.

Most of your comments have proved your knowledge to be superior to mine, but the impracticality of your explanations leads me to conduct a lot of research, which I'm trying to avoid at this stage. The summary of motors you see above is the result of real-world stress and efforts. I have had to learn to 3D design and print THOUSANDS of DISCARDED components. Step by step instructions with existing part links is all I want now.
This is my robot so far:

 

[Paul Nicholson]

I just had a great idea for a slip ring that's extremely cheap, you can find them in most homes, they can handle huge current and they also act as a one handed click-connector. By the way, this is the kind of lateral thinking I'm usually looking for:

Edit: I looked into these. They only tolerate around 13 Amps because household wall sockets only provide around 10 Amps. But wind turbine slip rings allow for 180 amps for US$25 or in my case, AU$35 plus $35 delivery plus 3 weeks waiting time. Also, they don't come with any kind of connectors:

 

[Baluncore]

I just had a great idea for a slip ring that's extremely cheap, you can find them in most homes, they can handle huge current and they also act as a one handed click-connector.

They are not designed for continuous rotation with current flowing. They arc and burn contacts if moved while turned on. They are only designed to carry about 8 amps AC because they are 230VAC. De-rate that to less than 1 amp DC.

I certainly avoid driving around Sydney. It is cheaper to mail order from China. I can avoid Hobart because Australia Post delivers to me.

Brute force solutions can be expensive in many ways. If designing low-cost and efficient robots is making you sick then you really should look for an occupation that you enjoy. At least you have a Medicare card to cover some of the damage. Many people think that designing new products is in itself a sickness.

 

[Paul Nicholson]

You can't assume that I don't like building robots because the effort of it landed me in hospital. I'm grateful for your suggestions. You are right about everything you say, but I don't think you're aware of the waiting time for shipments from China. I'm saying that links and step by step instructions are all I'm looking for so that I don't have to do heaps of research based on general concepts. I don't understand why you say 'Many people think that designing new products is in itself a sickness.'

My biggest problem when searching the internet is knowing which words to type into google when I decide on what I need. Right now I would like to find an electrical hook that attaches into a slot on the back of the robot's neck. The slot does not snare clothing and is welded to the core or spine of the robot. The slot includes includes an electric connector so that when the robot is lifted and dropped onto the hook of the rotating ceiling pole, the power is connected at the same time. I don't know what keywords to type to achieve this, or how wide my expectations need to be with alternatives.

'Shelf hook'

'flag pole connector'

'racking slot'

'hotel swipe card'

These things are on the right track, but I need to find all the pieces of THIS jigsaw puzzle before I assemble it. One decision-change affects everything else, and my decisions are based on 1. How easy it is to source for my product 2. Cost 3. Safe to use 4. Easy maintenance and assembly 5. Quality

 

[Paul Nicholson]

The puzzle I have been trying to solve in this thread can be seen in what I have so far:

I had already constructed that frame last week so I had a ceiling to work with.
Then I bought that garage door opener because it was right in front of me and I was losing motivation. This has re-inspired me.
The quick connector robot arm containing the expensive hobby servos has been under development for months.
The mannequin I bought a couple of weeks ago.

The mission is to run wires from the ceiling into the back of the mannequins neck and hold her solid so she can rotate and dance at full human speed without wobbling.

This pole design from my neighbor's apartment building seems the perfect shape for a beam that would rotate from the outside of a 75mm diameter steel disc. This would keep her centered without getting in the way of her head and arms.

But given that a single pole is more likely to wobble, and the garage door opener has two poles coming out of it, I'm willing to consider a more rigid Y connector to the neck.

The robot needs to lift off the ceiling hook any time her clothes are changed, which in many cases would be daily. After the curved ceiling pole meets her neck, THAT connector needs to constitute the top of her rigid spine, and none of it is allowed to wobble, whether she is performing high speed Karate or slow motion Tai Chi.

A sturdy spine structure faces it's first problem at the collar bone: The arm inputs meet in the middle as seen here:

Removing the head each time clothes are changed and locking the robot downwards might be feasible in the structural design, but I am more inclined to create the internal structure in multiple parts. Cutting the torso into two clam halves is not allowed unless there is no other way. So the five entry holes need to be considered: top of the neck, back of the neck, two arm holes and base of the torso. Also consider how this internal structure can be attached to the torso without any visible screws. Right now the rule is that the holes are the only places screws can be attached or flanges tightened into position - I'm not building Frankenstein's monster, even though the finish on the arms might make you think that. These are fashion mannequins and must have a flawless finish.