Software > Software
How to stop a continuous rotation servo?
Duane Degn:
My reading of Futaba's pdf lead me to believe analog servos gets a shot of power once per pulse with the amount of power being proportional to how far off target the position of the servo is with respect to the target position.
Is my understanding of the pdf incorrect or is the Futaba's description of how servos work incorrect?
Here's a link to the pdf.
http://www.futaba-rc.com/servos/digitalservos.pdf
Edit: I see I didn't word my description very well. Yes, power isn't only applied during the pulse but the pulse triggers the application of power to the servo. If the refresh frequency of the pulses sent to the servo is reduced the amount of power to the servo is reduced. (I think jwatte and I agree on this.)
Instead of saying "Analog servos only receive power during a pulse", I should have said "Analog servos only receive power when they receive a pulse". When fewer pulses are received, the servo receives less power, thus "starving the servo".
Again, this applies to analog servos. Digital servos how their own uC and send a continuous PWM to the servo's motor.
Duane Degn:
--- Quote from: jwatte on May 01, 2013, 11:23:46 PM ---For some hobby servos, you'll actually break the output gear before you manage to move the inert motor through the gear box.
--- End quote ---
Do you have an example in mind? I probably have a dozen or so different models of servos but I have never found one that wont let you turn the motor my turning the servo horn. One does need to start moving servos slowly (and some very slowly) but I haven't found a servo that wont move at all.
johnwarfin:
there are "old world" brand servos and then modern mainstream chinese "micro" servos which, in terms of sales and manufacturing, outnumber the former 100:1. i do have some old oddball types but 99% of my stock purchased last few years (hextronic, towerpro/tpro/tp, gws, turnigy, corona, hk, em, bms, vigor, bluearrow, etc) are based on aa51880 or clones. they behave similarly and as ive described above. one way to tell whether its old or new tech is to look throught the plastic shell. if you see transistors or other discrete components then this is not the modern rc servo referred to here.
if you remove pulses it does not move to zero or any other extreme. it simply stops and in addition resists further efforts to rotate. just because it is unpowered does not mean the arm now moves freely. depending on the speed you try to force it and gear ratio theres a good chance it will strip. as i mentioned before this is the main reason they fail in real world use. control surfaces jamed or jarred cause stripped gears. very few actually "wear out" or sustain electrical failure.
i highly recommend picking up a few cheap ($2) generic units from places like dealextreme, hobbypartz, aliexpress, hobbyking, or ebay and playing around with a servo tester . the testers are only a couple bucks or you can throw one together in few minutes with pic or avr. you will quickly see how the majority of modern servos behave in regards to torque/friction and removed pulses.
im sure if cantilevered far enough some servos can be made to sag or break. but in terms of sumo, tank, or other wheel robots my suggestion to stop by removing pulses is good advice with no undesirable side effects.
jwatte:
--- Quote from: Duane Degn on May 02, 2013, 12:00:40 AM ---
--- Quote from: jwatte on May 01, 2013, 11:23:46 PM ---For some hobby servos, you'll actually break the output gear before you manage to move the inert motor through the gear box.
--- End quote ---
Do you have an example in mind? I probably have a dozen or so different models of servos but I have never found one that wont let you turn the motor my turning the servo horn. One does need to start moving servos slowly (and some very slowly) but I haven't found a servo that wont move at all.
--- End quote ---
Mainly, the problem is when you try to apply force too quickly. Typically, this happens when colliding with something, rather than when trying to turn by hand.
Also, if you overheat servos, parts of the motor may expand and get out of alignment, so the motor has more internal friction, and if the gear ratio is high enough, you won't be able to turn it manually from the output spline.
I've managed to do that to some Dynamixel MX-64s, which was not a fun experience :-/
When it comes to "remove pulses, and the servo stops," I think we are all actually meaning the same thing, and just discussing the accurate description of how it works.
If you stop providing pulses to the servo, then after some timeout (50 ms, 500 ms, whatever) then the servo motor will stop providing any torque, and the only forces acting on the output will be friction in the motor and gearbox. For strong enough servos and/or small enough robots, that friction may very well be sufficient to stop the robot and hold it in place, unless someone pushes it hard; especially on a flat surface.
johnwarfin:
--- Quote from: jwatte on May 02, 2013, 12:30:40 PM ---I've managed to do that to some Dynamixel MX-64s, which was not a fun experience :-/
--- End quote ---
yes, i managed to strip a couple of blue arrow 2.5g that way back when they were $20 a wack. now we can get 1.7 gram digital servos for 1/6th that price. i do find these devices one of the most interesting aspects of buidling small robots. other interesting discussions might deal with how to modify for continuous rotation and also providing optical or hall position feedback. imo robots, like rc models, are among the most fun and educational hobbies. it seems to mix so many interesting technologies.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version