DPRG List

 [DPRG] Why high resolution encoders are desirable Message index sorted by: [ date ] [ thread ] [ subject ] [ author ] Previous message: [DPRG] Counting encoder tics quadrature vs non-quadrature Next message: [DPRG] Why high resolution encoders are desirable Subject: [DPRG] Why high resolution encoders are desirable From: Clay Timmons ctimmons at cadence.com Date: Fri Aug 24 11:11:00 CDT 2001 ```Some discussion has come up about why so many encoder tics are needed. The robot has to go 20 feet, Do you really need 100 tics per inch? Suppose you only have 1 tic per inch. Assume your robot can stop within 1 tic of deisired position. You program it to go 20 feet = 240 tics. It goes 241 tics (my algorithm only overshoots, never undershoots). Still the audience is impressed since the average spectator's eye cannot discern a 1 out of 240 error (0.4 % error). 1 tic per inch seems to be plenty of accuracy. Now lets talk about turns. Say you program your robot to do a 90 degree right turn. Assume your robot has a 12 inch wheel base. Each wheel must go (12 * pi) / 4 = 9.42459 inches. This would be 9 tics. First of all you have already introduced error by rounding to an integer number of tics! Next lets say you robot overshoots to 10 tics. 10 tics = 95.49 degrees 9 tics = 85.95. Either way your off by 5 out of 90 degrees or 5.5% error! Still the audience is probably impressed untill your robot starts driving 10 feet. Starting 5 degrees off your robot will be 120 sin 5 = 10 inches off ! Make one or two more turns and your robot will be hitting the walls! Bottom line. More accuracy is better! Turns need much more accuracy than straight moves. Later, -Clay Timmons- ``` Previous message: [DPRG] Counting encoder tics quadrature vs non-quadrature Next message: [DPRG] Why high resolution encoders are desirable Message index sorted by: [ date ] [ thread ] [ subject ] [ author ] More information about the DPRG mailing list