continuous rotation servos

[koran49]

My project is going to be a robotic arm on custom fitted onto a slider actuator and the robotic arm is going to be controlled by an 8 servo controller board. Now the problem I'm going to run into is how I will control the actuator back and forth, I thought I would be able to use a continuous rotation servo but you guys seem to be against that?

You mentioned Roboticinfo something about PWM, do you think I can use a regular DC motor to move the actuator and use something like https://www.pololu.com/products/pololu/0410/ and plug it into one of the servo outputs?

Robotic arm: https://www.imagesco.com/catalog/robot/arm.html

Actuator: https://www.directindustry.com/prod/slide-style-rodless-screw-actuators-102981.html

Servo controller: https://www.imagesco.com/catalog/motors/servoboard.html

 

[Roboticinfo]

reply

For the linear you can make it like a car, basicly you have 2 dc motors under the base plate and it would have for wheels two would be the drive wheels and the other two are for statability.

Anyway if you want the arm to go left you make the motors go forward and if you want it to go right then you put the motors in reverse.

The h-bridge you shown me only handles 2 amps if you use 1 motor, It depends on the motor you use. If you use a motor with with more then 2 amps then you have to find a h-bridge that can handle more amps.
The are h-bridges that are ICs which I prefer because they are way cheaper then the motor mind b I believe that you shown me.
Choose you motors first then report back to use.

With pwm you can make the arm travel at exact speeds and vary of course. Have you deceided what microcontroller to use?

Also how many degrees of fredom do you want your arm to have?

Also if you choose a dc motor look for the rpm generaly you want 20 to 36 rpm for a robotic arm.

 

[koran49]

what parts and from where do you recommend?

 

[Roboticinfo]

reply

First choose 2 dc motors with low rpm, the rpm should be around from 20 to 36 rpm and try to find 2 dc motors that doesn't have to many amps.

 

[Nigel Goodwin]

I thought I would be able to use a continuous rotation servo but you guys seem to be against that?

Not particularly 'against' the idea, the point I've been trying to get through all along is that a 'continuous rotation servoMOTOR' (note the MOTOR in the name), isn't a servo - it's just a geared DC motor with an internal H-bridge, who's direction of rotation you can change by altering the width of the pulse you feed it.

It's a VERY useful device, and (as mentioned many times) is commonly use as drive motors in small robots - simple to control, and easy to mount - plus it's fairly cheap.

With a servo you can move the position of the output (in your case a linear movement) to a specific point - so you could quite simply instruct it to position itself 6 inches from the end of the 24 inch travel, then move it to 3 inches from the beginning. This is called 'absolute positioning', and presumably what you are after? - certainly it's what I would want!.

If you use a stepper motor to drive the actuator you get what's called 'relative positioning', where you could move the actuator a specific distance left or right, but you have no way of knowing where it is - plus if the stepper misses steps, you've no way of knowing or compensating. One method of trying to overcome some of the limitations is to have limit switches, so you can identify when it's at either end - limit switches are a good idea anyway!.

The third option is to use a simple DC motor (or, a continuous rotation servo motor - basically the same thing!) - in this case you don't have any positioning at all. You can move the actuator in either direction, but you've no idea how far it might move, or even if it has moved at all!.

Throughout this thread I've presumed you imagined a 'continuous rotation servomotor' would give you absolute positioning? - which is completely untrue, and I suspect would seriously handicap your project?.

 

[Roboticinfo]

I thought I would be able to use a continuous rotation servo but you guys seem to be against that?

Not particularly 'against' the idea, the point I've been trying to get through all along is that a 'continuous rotation servoMOTOR' (note the MOTOR in the name), isn't a servo - it's just a geared DC motor with an internal H-bridge, who's direction of rotation you can change by altering the width of the pulse you feed it.

It's a VERY useful device, and (as mentioned many times) is commonly use as drive motors in small robots - simple to control, and easy to mount - plus it's fairly cheap.

With a servo you can move the position of the output (in your case a linear movement) to a specific point - so you could quite simply instruct it to position itself 6 inches from the end of the 24 inch travel, then move it to 3 inches from the beginning. This is called 'absolute positioning', and presumably what you are after? - certainly it's what I would want!.

If you use a stepper motor to drive the actuator you get what's called 'relative positioning', where you could move the actuator a specific distance left or right, but you have no way of knowing where it is - plus if the stepper misses steps, you've no way of knowing or compensating. One method of trying to overcome some of the limitations is to have limit switches, so you can identify when it's at either end - limit switches are a good idea anyway!.

The third option is to use a simple DC motor (or, a continuous rotation servo motor - basically the same thing!) - in this case you don't have any positioning at all. You can move the actuator in either direction, but you've no idea how far it might move, or even if it has moved at all!.

Throughout this thread I've presumed you imagined a 'continuous rotation servomotor' would give you absolute positioning? - which is completely untrue, and I suspect would seriously handicap your project?.

I would have to agree with you nigel, also why do you need accuracy?

 

[koran49]

With a servo you can move the position of the output (in your case a linear movement) to a specific point - so you could quite simply instruct it to position itself 6 inches from the end of the 24 inch travel, then move it to 3 inches from the beginning. This is called 'absolute positioning', and presumably what you are after? - certainly it's what I would want!.

Thats exactly what I want Nigel! Is there a servo out there that can turn beyond 180degrees so I can get my full 24inches of linear movement and still keep my 'absolute positioning' as you put it? Or am I limited to only 180 degrees with using a servo (which would only move my actuator about 1/2inch). Or am I doomed to use a DC motor which could greatly hinder my project.

Thanks for the lengthy explainations I can feel my brain expanding already

BTW I'm not sure on this but I think what I'm going to have the arm do is build those bigger sized legos or wood blocks make a wall or something of that nature.

 

[Roboticinfo]

reply

The most a servo can do is 270 degrees that is it, which was stated early next time pay attention.

Why do you need it to be that accurate?

If I couldn't go with servos I would go with steppers.

Usually steppers do skip a couple of steps sometimes, but if that would be a problem then you can use limit switches as also stated early.

What don't you get?

 

[Nigel Goodwin]

Thats exactly what I want Nigel! Is there a servo out there that can turn beyond 180degrees so I can get my full 24inches of linear movement and still keep my 'absolute positioning' as you put it? Or am I limited to only 180 degrees with using a servo (which would only move my actuator about 1/2inch). Or am I doomed to use a DC motor which could greatly hinder my project.

The potentiometer in a radio control servo only has a maximum possible movement of 270 degrees, and usually only allows about 180 degrees maximum movement of the servo horn.

As I mentioned earlier in the thread, by removing the potentiometer from the servo, and wiring an external one moved by your actuator, you could produce a servo that gives a full 24 inch linear movement.

However, the problem is how to convert the 24 inch linear movement to move the potentiometer!. If you could find a 24 inch slider potentiometer, that would make it pretty simple for you! - however, I suspect that would be very difficult to find?. Another possibility is to try and convert the linear motion to a rotary movement - this could be done with a system of gears, and you could use a multi-turn potentiometer (usually ten turn) to make the gearing simpler - but this does introduce problems of backlash!.

Basically, your problems are more mechanical than electronic!.

 

[bloody-orc]

hi there.
I see you have an interesting project in hand.
anyways, this Nigels idea is i think the best and the easiest in the future(using it is very simple)
if you cant find the 24'' linear pot, then why not build one yourself:
take a 24'' long wire what has enaugh resistance and then "slide" on it and get your positioning. If this is too hard, then you could also get an adc value out of it and read it with a pic adc input and from there calculate the place where you are with your robotic hand. this stepper is bad idea, cause it as mantioned earlyer skips few steps and is not acurate enaugh. If using this wire then you can use both servo and DC motor (depending on what tecnique you use) you get pretty accurate results. (more accurate at least than using stepper).

But if this is still imposible to do, then why not try to use an Ir sensor os a ultrasonic module, and calculate the position from there .

hope it helped.
Rain

 

[mechie]

Missing the easy answer ?

A 24 inch travel would be easy if ...
The motor turns a threaded rod (a lead screw).

The lead screw has a nut on it which cannot turn so when the lead screw is rotated the nut will travel up and down the lead screw.
The nut has a 'bit of string' (drive cord) tied to it, this is looped over two pulley wheels as shown.
One of the pulley wheels is fastened to the shaft of a 10-turn potentiometer (or 5-turn or whatever).

As the motor turns,
the nut travels up or down the screw,
The drive cord is pulled one way or the other,
The potentiometer shaft rotates,
## FEEDBACK ## :!: connected to the servo positioner

Old radio tuning dial sort of thing

 

[mechie]

Controlling the beast

I wonder if there isn't something much cheaper than the control board you are looking at ...

Maybe the same money (or less) could buy you a PIC programmer and a fistfull of bits, there are PIC-based servo controller plans scattered around the web, free for the taking.
You could then consider a custom keypad to control your robot ???

I know this is a radical rethink from M$ Visual Basic and may not be your preference but it could possibly be done - even with some sort of simple 'scripting' programmability.

ps.
There is an add-on for a LEGO RCX controller that allowed radio-control servos to be driven ... ?
http://www.mindsensors.com/index.php?module=pagemaster&PAGE_user_op=view_page&PAGE_id=42
or
**broken link removed**

 

[koran49]

Those are some good ideas to consider thanks for the suggestions, I found this controller board https://www.pololu.com/products/pololu/0290/ and from the looks of it this motor controller https://www.pololu.com/products/pololu/0410/ will interface together fairly simple then I could just program the 2 together what do you guys think?

 

[Nigel Goodwin]

Those are some good ideas to consider thanks for the suggestions, I found this controller board https://www.pololu.com/products/pololu/0290/ and from the looks of it this motor controller https://www.pololu.com/products/pololu/0410/ will interface together fairly simple then I could just program the 2 together what do you guys think?

I don't really see what you're asking?. One board is for controlling servos from a serial input, and the other for controlling small DC motors from a serial input - there's no way (or reason to want to) connect them together.

 

[koran49]

I thought it was going to control my screw actuator but I guess it was a bad idea.

 

[Oznog]

Thats exactly what I want Nigel! Is there a servo out there that can turn beyond 180degrees so I can get my full 24inches of linear movement and still keep my 'absolute positioning' as you put it? Or am I limited to only 180 degrees with using a servo (which would only move my actuator about 1/2inch). Or am I doomed to use a DC motor which could greatly hinder my project.

A servo usually IS a DC motor.
I'm not sure if anybody mentioned it yet or not. This type of linear positioning is extremely common. It's not commonly done with pots. Rather, you have a stepping motor and a "home position sensor". Sometimes, a DC motor with an optical encoder is used by steppers are much more common. It initializes by backing up until it hits the home position sensor, so it knows the absolute position, and from there counts the turns. It's quite accurate. You still have the issue of backlash in the screw, if well made it can be low. In any case, pots aren't exactly 100% accurate either and typically have a very short lifespan in this sort of duty.

 

[koran49]

Could you point me in the direction of where to buy stepper motors and controllers please someone?

 

[Nigel Goodwin]

Could you point me in the direction of where to buy stepper motors and controllers please someone?

The easiest place to find them is in old printers, or in old 5.25" floppy disk drives, or you can buy them at any decent electronics store.

As for controllers, all you need is simple drive transistors (or ULN or L series driver IC's), and do the controlling from a micro-controller - it's simple to control a stepper from a uP.

One other thought! - how fast do you want the actuator to move?, steppers are very slow! - unless you can feed them from high voltages probably something under one revolution per second!.

 

[koran49]

One other thought! - how fast do you want the actuator to move?, steppers are very slow! - unless you can feed them from high voltages probably something under one revolution per second!.

Well less then 1 rpm per second is kind of slow considering I just bought a screw rod that has 16 threads per inch. I can't help but feel like I'm going in circles How about an Optical Encoder that was mentioned earlier what can you guys tell me about that?

 

[chiba]

Many years ago i used chopper modules and surplus Bi-polar nema-34 stepper motors to build a cnc milling machine. I powered them from an old computer power supply modified to run @ 17v. They were running about 330rpm with 4 tpi screws, used it mainly for "chopping" through aluminium and pcb.

google for GSD200S, cost about $US60ea.

https://www.selectronic.fr/includes_selectronic/pdf/Thomson/GS_D200S.pdf