1. Home
  2. General Robotics Forum
  3. using a servo/encoder for motion recording and playback

using a servo/encoder for motion recording and playback

Hi all,

I'm working on a project which requires that I be able to 'record' externally driven rotations of a driveshaft and then 'play' them back using a servo/stepper motor. It'd be nice if I could just use the encoder in the servo/stepper for recording, but I'm not sure about backdriving servos. Its also very important that during recording, the external driver experience as little friction as possible. This makes me think that I need some kind of clutch, but I haven't been able to find anything suitable despite my best googleing efforts.

Are there any off-the-shelf parts (i.e. servos with a clutch) that could be used to do this? If I were to use a sufficiently powerful motor such that it didnt need to be geared down at all (or very little) would it be okay to just backdrive it?

If its useful, the motions that I need to record will vary, but will likely not exceed oscillations of about 5 degrees at 15Hz. My torque requirements are likely to be modest, but I'd imagine than what typical RC servos can deliver.

thanks in advance, Graham

Reply to
grindlay
Loading thread data ...

Some servos will indeed be unhappy if you backdrive them, but others are made for it. Many of the newer robo-one kits use this as the primary means for building new animations: you simply pose the robot, and then click "capture" in the software to record all the servo positions.

I've been looking at a lot of servos, and the Robotis AX-12 seems to me to be by far the best value for the price. It's $45 from CrustCrawler, offers 16.5 kg cm of torque (which is quite a lot!), and is a "smart" servo -- that is, it speaks a two-way serial protocol, and can report various things like its current position, current draw, etc.

Here's a community wiki page about it:

formatting link
And here's a demo of one AX-12 being manually driven as the input to control another AX-12:
formatting link
So, this certainly should be doable.

Best,

- Joe

Reply to
Joe Strout

I looked into some digital Hitec's recently for someone on the forum. I searched for the post and haven't found it though. Anyway...

I looked into the spec and found they (Hitek iirc) had a mode (very short pulse) where the servo would answer back with a pulse indicating where they were positioned. When they were in this mode, they stopped driving the gear train as well. So they are ready made to be positioned and read. Then with regular pulse lengths, you can play back the pulse which was read and it will cause a positioning.

This sounds exactly like what you were asking for, however, the new digital servos are not the cheapest possible way to acheive this, and you'll have to either have a special driver to position/read them, or do some fancy programming.

-- Randy M. Dumse

formatting link
Caution: Objects in mirror are more confused than they appear

Reply to
RMDumse

Thanks to both of you; this info is very helpful.

Randy - Any luck locating that post? I'm quite interested in the posibility of using the HSR-5995TG as I think that it is powerful and quick enough for my purposes. I did, however, find the following discussion on getting position info out of Hitec servos:

formatting link
The first post links to an interested pdf containing analysis of dissected HS-5475HB, HS-5245MG, HS-5645MG, and HSR-5995TG servos. In that document the author states that these servos use an H-bridge configuration that only allows the motor to be in drive or brake mode (not freewheel) when the power is applied. But perhaps the mode you mentioned removes power from the bridge entirely?

In anycase, any other information about the position readout mode you mentioned would be greatly appreciated.

thanks aga> snipped-for-privacy@gmail.com wrote:

Reply to
grindlay

I was begging to wonder if I was dreaming. But then I searched on my posts without any qualifiers and 17 pages of scanning later...

formatting link
which refered to this page:

formatting link
"The first thing to note is that the 50=B7S pulse which initiates the position feedback also disables the motor drive."

Sounds like they only have 10% accuracy though.

HTH's

-- Randy M. Dumse

formatting link
Caution: Objects in mirror are more confused than they appear

Reply to
RMDumse

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.