On Fri, Dec 28, 2007 at 11:39:38AM -0800, Kirk Wallace wrote: > This looks interesting: > > http://axis.unpythonic.net/01198594294 > > This looks like an inexpensive way to get fast PWM and encoder rates > except it doesn't work in realtime yet?
I have been debating whether to announce this to a bigger audience. I think this is more a proof-of-concept than anything truly useful. First, beacuse the particular board I used is on a USB interface, it simply can't be used for realtime purposes. Switching to a "traditional" PC serial port (there are, or at least were, similar board designs with RS232 or TTL serial instead of USB), you run into bandwidth limits -- you can't quite transmit 12 bytes every 1ms at 115200bps, and with the protocol I used, it takes 12 bytes to read or write all the pins. You also have to budget the time that the micro takes to actually perform its actions -- e.g., it takes at least 100uS to do one ADC conversion, and because the AVR's UART has only one byte of output buffering it can't efficiently be interleaved with transmitting the previous sample over the serial connection. The driver from my website has ADC in and PWM out (the avr microcontroller has both built into the hardware) but no quadrature decoder. Quadrature might be possible, but it would be fairly low speed because avrs do not have dedicated quadrature decoders. At best you'd get something good enough for use with an MPG, not for a spindle or axis motor (5kHz sample rate, maybe?). Jeff ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users