Servo signals vary widely among manufacturers. I took an article on faith that claimed 0.5 to 2.5 milliseconds was the right signal range. Worked great on one servo. Bought another servo, and it was acting a little funny around the end of travel. the second time it hit the end stop it went "POP!" and never worked again. OOPS it was a 1 to 2 ms servo. You can NEVER tell this info from the packages, you have to just call the mfr and ask their tech heads. Don't guess. SILVER LINING: So I took apart the servo, and removed the busted pot and the stops. Replaced it with an external 5K linear pot. Now my servo drives a little nonlinear mechanism (cams and stuff) that is linearized because the feedback is an external linear pot! Much improved. I have another servo that I replaced the pot with fixed resistors and removed the stops. Instead of position, the input signal determined speed and direction. If you hack a servo this way (they are EMINENTLY hackable!) you can go outside the mfrs. signal specs quite a bit. I fell in love with servos about a year ago, having ignored them all this time. Servos are great little packages of cheap mass produced technology! They were born to marry the PIC. -----Original Message----- From: Quentin To: PICLIST@MITVMA.MIT.EDU Date: Tuesday, December 28, 1999 9:04 AM Subject: Re: Servo control signals to analogue voltage conversion >"Paul B. Webster VK2BZC" wrote: >> >> Philippe Jadin wrote: >> >> > Well, I put together some informations about servos (controlled by lpt >> > port on a pc running win 95, but still interesting if you know nothin >> > about servos) : >> >> > goto http://users.swing.be/philippe.jadin/servoen.htm >> >> OK, firstly your webpage omits the ground connection to the printer >> port connector. I think you should have shown that clearly as some >> people will otherwise complain of very peculiar behaviour. > >You also talk on your web page of pulse lengths of 1.5 to 2 millisecs. >AFAIK 1.5 millisecs is servo centre with 1 millisecs minimum and 2 >millisecs maximum position of the servo. Also don't use 1 millisec and 2 >millisecs as absolutes, as it can damage some servo's. Rather keep it >between 1.2 and 1.8 millisecs to be safe. > >Quentin