Servo Best Practices?

[adrenalynn]

{edit}

Doh!  Of course I find my answer here: http://www.societyofrobots.com/axon/axon_function_list.shtml#hardware_defines after I post...   Sorry for missing it!

(20ms suggested as optimum safety margin as I read it.)


Hi All,

I did a quick search and didn't see this addressed anywhere, so I'm surfing for opinions:

I'm looking to do a wireless serial link betwixt the Axon and the SSC-32, the Axon taking analog inputs and translating them into servo positional commands for the SSC-32.  Ok - so far, so good, that's all easy.

The question that I now face is whether I can run that translation in a tight loop.  If I blast the same pulse repeatedly out to the servo is it going to get cranky or do I really need to test for change in position value before instructing the servo(s) each time through?  Ideally I'd like to get my lag down as far as possible so I would just stupidly pound the servo:  "go to position 900, go to position 900, go to position 900, go to position 1400, [...]  Am I going to get into trouble with the servo doing something like that?

Thanks in advance for any input!

[adrenalynn]

Apologies for replying to myself, but after that last edit. . .

Any issues, assuming a 20ms delay, with sending the same position command to the servo repeatedly when it's already in that position?

TIA, Again.

[paulstreats]

No issues.

Thats how you keep it in the same position. Otherwise the servo goes to sleep and the output shaft is free to spin loosely. By repeating the same position signals, the servo holds the position with its motor.

Also, sometimes when you are moving servo's from 1 position to another it can take more than 1 position signal before it reaches its position.

[adrenalynn]

Well, yeah, duh.  I'm a moron. [where's the "blush" emoticon?]  Ms Obvious Answer - that's me. [insert "blush" emoticon again]

Thanks for your assistance!  I think what I'm looking to do is actually pretty trivially simple - I'm just really good at overcomplicating projects.  If it's worth engineering it's worth over-engineering, right? [liberal application of "blush" emoticon required]

[Iron Man]

it's all about learning.

[Gertlex]

Correction: if you're using a servo controller, you only need to send that byte once.  Doesn't hurt to send it over and over again, though.  The servo controller's purpose is to move servos and *keep them in that position*.  They send constant pulses regardless of whether they receive constant updates.

[adrenalynn]

Thanks!

I did notice that today when I put it on the 'scope.  It also doesn't seem to matter, practically, how frequently I issue the servo position - the SSC-32 doesn't want me to be too stupid and burn-up a servo *that way*.  Of course, it does let me send positions that are out of range of many (read: all of mine) servos and could quickly toast a servo that way.

I decided that putting the controller on the oscilloscope for debug was really "the better part of valor".  Which is a nice way of saying that I'm "chicken".

[MadMax]

also, it's good practice to make the servo slowly change it's position, this makes the servo last much longer. However, you'll need to make a little more effort (timing and coding the thing in a right way)

[Admin]

Thats how you keep it in the same position. Otherwise the servo goes to sleep and the output shaft is free to spin loosely. By repeating the same position signals, the servo holds the position with its motor.

This is true for analog servos, but digital servos always hold the last position it was commanded to go to.