DC Motor or Servo for rack and pinion drive
2008-09-30 19:38
Hey guys...another question...which would you use and how?
I want to make a rack and pinion armature system for my bot. But I'm unsure of which direction I should go...Let me explain my ideal setup.
Two parallel rails that run in parallel with the rack. the Motor would be on a platform that uses bearing to run in the rails and the pinion to drive the platform back and forth. similar to a gantry.
Heres the kicker...I would like to have a platform travel speed of about 16 inches per second and have lots of torque in the motor for the mechanics that cause about 3lbs of resistance to the platform. So at this point I'm thinking DC motor.
However, in combination with speed control and constant torque, I would like to be able to dynamically set the travel distance of the platform from 0 to 16 inches of travel, also I would like to have the platform oscillate from end to end and back repeatedly, with the smoothest motion possible at a constant speed...I know thats a tall order, from speed to dead stop then reverse direction.
I'm stumped....anyone got any ideas?
Umm...
I can see... in your future...
Ah, Yes! You are dropping 200 bucks on a ready-made linear actuator!
Seriously, a good actuator would do the trick, be wicked strong (maybe not quite fast enough) but would have all the bells and whistles you need i.e. feedback, stops, reverse etc...
Alas..
pictures?
This seems like a complicated setup you're after. Any pictures/sketches yet?
About the smooth ride. That calls for a controlled acceleration/deceleration. Do I understand correctly that your entire track is 16 inches (0.4m) long and that you want your top speed to be 16 inches/second? That means accelerating from 0 to top and decelerating to 0 again while traveling 16 inches.
If the acceleration is constant and if the top speed is reached exactly in the center of the track (a=16 inches per second per second), then the total travel time would be 2 seconds. This could be shorter if the acceleration would be better.
16"/s2 is little more than 0.04 g. I guess you could find a stronger machine.
Ok well my first gen design
Ok well my first gen design created the linear motion using a reciprocating piston. Much like this animation.
I was able to change the stroke length of the piston (armature) by change the distance of the connection for the connecting arm to the radius of the wheel. However I had to use a servo motor to change the distance, this caused an unbalanced wheel and cause alot of vibration. Also the design is too long. about 20 inches...8 for the wheel, 12 for the connecting arm. I would like to reduce the overall length to about 10 inches, so I thought about the gantry idea where the armature is on board and the platform does the traveling.
the track is about 10 inches and I would like to make a complete circuit in > 1 sec.
variable piston stroke ehh?
So what you are really after is a variable stroke length for that piston? While still able to give it a proper push, hence the torque required.
If the overall length is an issue, maybe you could "fold up" your design? Connect the piston to the arm via a lever (inverting the direction).
Hmmm, expanding on that thought: by controlling location of the pivot point in the lever (sliding up and down), you could effectively control the traveled distance. This needs more sketching I'm afraid!
must the motor be on board?
must the motor be on board?
maybe place a stepper motor along the track, like a matrix printer, using a toothed belt?
No the motor doesnt have to
No the motor doesnt have to be on the board...I think a toothed belt or drive chain could work
no motor no inertia
weightmass would be less, your acceleration would be much better.What about something like
What about something like this: i assume this is something similar to what you meant for the motor not being on board
(please excuse the crudity of the drawing I'm at work and dotn have access to drafting software)
The gears position would be fixed, the spindle would be attached to the belt but allowed to rotate in the fitting that attaches it to the belt. As the belt moves the gantry is pulled forward, as the spindle reaches the far gear, the spindle rotates around the gear and pulls the gantry back the other direction. The only interuption of linear travel is when the spindle rotates around the gear, however that can be minimized my using smallish gears. Now however to change the travel distance of the gantry, lets say from full stroke to half stroke, the system would need to know at all times where the gantry is.
however the problem with this design is lets say that the travel distance is very short ~ 1 inch , as the motor oscillates is to create the linear action in the chain the spindle will also rotate in reaction to changing from pushing to pulling...gah why is engineering so hard
A little less accurate, but...
Pneumatics, maybe? --The cylinders are super cheap but you are going to have a few bucks in tanks/ hoses and valves. You would, however have your acceleration. I mean 16" / sec is pretty darn fast.
I just thought of another one... If you have the room, you could do a car's power window motor with a long arm. The arm would need a long slot cut in it for a pin to slide in (as the arm arcs, the pin sliding in the slot would allow for a transfer to linear travel) Would have to be a long arm, though. I remember using one of these motors in a similar set up back when I was doing custom car stereo work... Worked well, and sure did have the speed, I'm sure you could get easily into the 16"/sec range.
the problem with pnuematics
So this design is out of the question?
So this design is out of the question?
http://www.instructables.com/id/Low_Cost_Hobby_Servo_XY_Table/
http://www.teletoyland.com/Projects/Sandbox/index.php (java)
My current design is similar
I'm actually using those exact rails in my first gen prototype, the problem is that they provide a fine track, however the shortest versions are 14 inches long collapsed. Now you could say, well just cut the top segment to a small length, problem there is that the ball bearing slide travels independantly in regard to the top track..So if you cut down the top, it will pop loose from the ball bearing slide and fall apart.
On the other hand if you were talking about the toothed belt, thats almost exactly what I meant, however I still need the torque and speed, doesnt look like that servo table is very high speed
Write LMR
Arrick Robotics
Anyone Used a magnetic rotatary encoder for a DC motor
I found a link for converting any DC motor into a psuedo servo motor using this kit:
http://hackaday.com/2008/07/16/turn-any-motor-into-a-servo-with-repraps-new-board/
has anyone every used one of these? and if so, how do you actually hook it up? how do they work?
Hall effect encoders
I may be way off the mark -
I may be way off the mark - but why not use a small dc motor instead of a servo (more strength & speed) with threaded rod in a block. The benefits of using a threaded rod as a worm drive is the fact that when the motor stops the gantry will lock. I've done this many times and the results have been pretty satisfactory. Here is an example with the up/down tilt control of my bots head
http://www.youtube.com/watch?v=L6VJOS-ZJvI - instead of a pivot (in this case) you could mount it on rails
You could use IR or halls effect encoder to do positioning
I thought about that, usign
How about this?
Hey guys...came up with a slightly different design using 2 different motors
One standard DC motor to use for rotation of the flywheel, speed set through PWM
The other a servo with an arm on it that rotates in a 90 degree arc. the arm is under the flywheel with the servo above. the the stroke length of the piston can be changed by adjusting the angle of the arm relative to the line between the servo motor and the axis of the fly wheel. Heres a bad picture of my thought
Would a servo motor have enough torgue to maintain its postion while the flywheel is spinning and a resistance force is being applied to the piston?