Novice with question about servo control

Archiver

I would lie to proportionally control a servo using a variable
potentiometer. I have tried using a 555 timer and a r/c circuit to
collect a variable but I am having a hard time converting the
variable to a useful pulsout number. Can anyone give me a clue? Thanks

Jeff

Archiver

DLC, Thank you very much! A very elegant solution that also addresses servo
response time which is very critical in this application. I do need to
operate 2 servos and was thinking about using a stamp for each as the servos
will have to operate simultaneously. My initial thought was to include both
the checking of the pot and the pulsout to the servo in the same for next
loop so that there would be a servo correction for every pass through the
loop. I will give your method this a try this evening.

Thanks again

Jeff

Archiver

Jeff,

You can use the pot with a capacitor and the rctime function to get
the pot values. You'll then need to scale it in some fashion to whatever
resolution you want with the servo control and use pulsout to the servo.

Remember that the rctime function takes some finite amount of time to
complete - use the formulae in the Stamp II manual to find that time and
factor that into the pause for the pulsout command.

I don't recommend that you do the "usual" servo commands that look like:

for x = 1 to 15
pulsout SERVO,value
pause 20
next

You would have just used up over 300ms where the Stamp was doing nothing
but twidling its thumbs. Rather, use some form of state machine to track
where you are in the 20ms repeat cycle needed by the servo, for example:

SERVO con 7
LOOPS con 20
v var byte
s_state var nib

s_state = 0

main:
gosub get_pot
gosub move_servo
goto main

move_servo: 'moves servo to v location
if s_state = 1 then dservo
pulsout SERVO,v 'do pulsout
s_state = 1 'set up for decrement
goto sdone
dservo: 'decrement loop
sloop = sloop -1
if sloop > 0 then sdone
s_state = 0 'reset to do pulsout next
sloop = LOOPS 'reset loop counter
sdone:
return

get_pot:
..

I arbitrarily set LOOPS to 20, timing Stamp instructions is kind of non-
deterministic on occasion, so to figure out what your loop count needs to be
I suggest that you set it to some high value where the servo doesn't work well
and reduce it until it does work. Note, you can add several routines to the
main loop and get the affect of having a multitasking setup just by reducing
the number of times the move_servo loop is executed between pulsout commands.
All the routines in the main loop can be constructed similarly and you will
interleave actions to get more efficient use of the Stamp. For instance,
suppose you wanted to sample 2 pots to position 2 servos, you could have the
get_pot routine sample one pot on one entry through and the other pot on the
next time through, then you can have the math for scaling done on a third
loop through that sets the "v" value. Each servo could have its own move_servo
function as well. This will allow you to get a WHOLE lot more done without
forcing the Stamp to sit in useless "wait" loops.

DLC

> I would lie to proportionally control a servo using a variable
> potentiometer. I have tried using a 555 timer and a r/c circuit to
> collect a variable but I am having a hard time converting the
> variable to a useful pulsout number. Can anyone give me a clue? Thanks

Archiver

Jeff:
Maybe I don't understand the big picture, but I built a servo tester that
controls servo position with a pot. All it uses is a 555 timer controlled
by the pot. It consists of a 555, 2 resistors, a diode, a cap, and a pot.
If this is useful to you, I can describe the circuit.

Ray McArthur

---

Original Message

---

From: Jeff & Julia <EL-JEFE@P...>
To: <basicstamps@egroups.com>
Sent: Monday, May 22, 2000 12:03 PM
Subject: Re: [noparse][[/noparse]basicstamps] Novice with question about servo control


> DLC, Thank you very much! A very elegant solution that also addresses
servo
> response time which is very critical in this application. I do need to
> operate 2 servos and was thinking about using a stamp for each as the
servos
> will have to operate simultaneously. My initial thought was to include
both
> the checking of the pot and the pulsout to the servo in the same for next
> loop so that there would be a servo correction for every pass through the
> loop. I will give your method this a try this evening.
>
> Thanks again
>
> Jeff

Archiver

Jeff,

De nada. If you are interested in some examples that may help you design
such code check out my web page http://www.verinet.com/~dlc/botlinks.htm
and bonk on the article download under the BoE-Bot picture. Also, the
Parallax web site has a PDF of this article under their robotics section,
I think its called experiment 6 or somesuch.

have fun,
DLC
> DLC, Thank you very much! A very elegant solution that also addresses servo
> response time which is very critical in this application. I do need to
> operate 2 servos and was thinking about using a stamp for each as the servos
> will have to operate simultaneously. My initial thought was to include both
> the checking of the pot and the pulsout to the servo in the same for next
> loop so that there would be a servo correction for every pass through the
> loop. I will give your method this a try this evening.
>
> Thanks again
>
> Jeff
>
>
>
>