Servo motor help

Archiver

Hello all-

Thanks to Al Williams' programming chops, I've been working on getting some
constantly rotating servo motors to raise and lower sculptural objects
(about pound each) with some basic Futaba S3003 servos. Now that I have
them in the studio and hooked up, the timing on the motors seems -way-
loose. By themselves (no load) maybe a couple of seconds at slowest speed
(151) over 30 seconds. Today I took the whole rig to the sculpture studio
and hooked up an "elevator" (up and down) motor on the floor. String on a
spool up to the ceiling, through a pulley and down to a suspended weight
(>8oz) and what I noticed is if the motor does 151 for 1000 clicks, pauses
and down for 1000 clicks the distance traveled stays about the same (with a
couple of inches of 50") but the whole distance descends by about 25% per
elevator trip. As in it starts up at 30" then next starts up at 24", then
18" etc though not quite that mathematical. So I proportioned the amount of
up (1000) to down (around 790) and it is somewhat helped it working but
still with a range of 6 inches out of 50-odd traveled. A weight of 16 oz
completely screws up the motors. They are rated at 40 oz. torque which I
thought was enough.

Any motor suggestions?

I thought these motors, being used for robots, would have a closer
tolerance. Any observations, experiences or advice??

Thanks,

John

////////
'''''''' John Phillips / sound artist: http://terragizmo.net

Archiver

John,
I'm glad it's doing something, anyway. Some ideas:

1. Higher torque servos will likely cure this gradual reduction in throw.
The drawback is more $$ - a good quality high (~130 oz/in) torque servo
from a major manufacturer can be $150.00. I have <definite> opinions on
which ones are the best if you choose this route...they'll be more
reliable, too. This type of servo can usually be run at a higher voltage
which will make it even more powerful.

2. There are servo-winches for RC sailboats which can take up and spool out
long runs of cable. These are also a bit pricey, but tend to be more robust
than standard servos w/ sealed cases, etc. Another way to do this is to
modify a standard servo for continuous turning and attach a spool to it.
This mod involves removing the hard stop inside the case and installing a
fixed resistance in place of the potentiometer. Specifics on this are
available all over the web now.

3. I should tell you that any RC servo will only last for about 100 hours
of continuous use. This figure has been proven over and over again in
display-type applications...it's quite reliable. If you need a servo in a
continously running app - make them replaceable.

There are of course very heavy duty large scale servos with aluminum
cases and several hundred oz/in of torque. These are really expensive (ie.
350.00+).

HTH, Duncan

Archiver

Hi John,

Has to do with raising and lowering a load. It takes more power to raise the
load than to lower it. Have you thought of using a limit switch or opto
interrupter at the top of the travel. The motor runs until it hits the top
limit, then you could use a simple counting and time technique to lower the
object. The phenomena you observe is called hysteresis or backlash. We use
this system of de-hysteresis in many of the motion control devices we build.
I know it would make the setup a little more complicated, but you wouldn't
have the problem with different amounts of travel in each direction.

Best regards,
Gary

G. Shearer
Free Electron Laser Research Center
Vanderbilt University
Nashville, Tennessee

---

Original Message

---

From: "John Phillips" <jpch@t...>
To: "Stamp list" <basicstamps@egroups.com>
Sent: Wednesday, August 16, 2000 6:31 PM
Subject: [noparse][[/noparse]basicstamps] Servo motor help


> Hello all-
>
> Thanks to Al Williams' programming chops, I've been working on getting
some
> constantly rotating servo motors to raise and lower sculptural objects
> (about pound each) with some basic Futaba S3003 servos. Now that I have
> them in the studio and hooked up, the timing on the motors seems -way-
> loose. By themselves (no load) maybe a couple of seconds at slowest speed
> (151) over 30 seconds. Today I took the whole rig to the sculpture studio
> and hooked up an "elevator" (up and down) motor on the floor. String on a
> spool up to the ceiling, through a pulley and down to a suspended weight
> (>8oz) and what I noticed is if the motor does 151 for 1000 clicks, pauses
> and down for 1000 clicks the distance traveled stays about the same (with
a
> couple of inches of 50") but the whole distance descends by about 25% per
> elevator trip. As in it starts up at 30" then next starts up at 24", then
> 18" etc though not quite that mathematical. So I proportioned the amount
of
> up (1000) to down (around 790) and it is somewhat helped it working but
> still with a range of 6 inches out of 50-odd traveled. A weight of 16 oz
> completely screws up the motors. They are rated at 40 oz. torque which I
> thought was enough.
>
> Any motor suggestions?
>
> I thought these motors, being used for robots, would have a closer
> tolerance. Any observations, experiences or advice??
>
> Thanks,
>
> John
>
> ////////
> '''''''' John Phillips / sound artist: http://terragizmo.net
>
>
>
>
>

Archiver

John:

How are you keeping up with the position of the motors? Are you using a
closed loop feedback with an incremental encoder or just sending out the
signals to the motors and hoping they will position correctly?

Randy A.