Newsgroups: comp.robotics
Path: brunix!cat.cis.Brown.EDU!agate!howland.reston.ans.net!usc!elroy.jpl.nasa.gov!decwrl!netcomsv!netcom.com!nagle
From: nagle@netcom.com (John Nagle)
Subject: Re: hobby servo motors
Message-ID: <nagleCu2nqE.6Av@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
References: <Ctyv0J.BKI@hpl.hp.com> <MAFM.94Aug4133451@parma.cs.uwa.edu.au> 	<nagleCu0r7I.Kz7@netcom.com> <MAFM.94Aug5152035@parma.cs.uwa.edu.au>
Date: Fri, 5 Aug 1994 17:11:02 GMT
Lines: 26

mafm@cs.uwa.edu.au (Matthew McDonald) writes:
>nagle@netcom.com (John Nagle) writes:
>>   The off-time doesn't even have to be constant, and it isn't
>>   constant in the usual R/C application, where the full-cycle
>>   time is the sum of the ON times of all the servos plus a
>>   delay at the end of the cycle.
>Interesting - the servos I tested jittered if the off-time varied from
>one cycle to the next. This happened with good Futaba servos and
>cheapo Sanwa ones, so it seemed like a generic problem.
>What I did that caused the jitter was disable interrupts in the posted
>code for only the on-time timing loop. The jitters were really big.
>Does this mean I did something wrong or are the servos just faulty?

      That's strange, because if you look at the waveform that goes
over the RF link to an RC receiver, what you see is a series of pulses.
The time from pulse 1 to pulse 2 is the ON time for servo 1, the time from
pulse 2 to pulse 3 is the ON time for servo 2, and so forth, for up to
8 servos and 9 pulses.  Then there's a period with no pulses that syncs
up the decoder, and that period has a fixed length.  So the length of the
whole cycle varies, depending on the individual servo settings.
This is from measurements on an Airtronics unit circa 1987, with
analog timing.  We're talking here about an allowable jitter in OFF time
no greater than the total range of 8 servos, which would be about 8ms.
More jitter than that might cause problems.

					John Nagle
