#linuxcnc-devel Logs

Mar 20 2023

02:16 PM -!- #linuxcnc-devel mode set to +v by ChanServ
03:37 PM -!- #linuxcnc-devel mode set to +v by ChanServ
06:14 PM -!- #linuxcnc-devel mode set to +v by ChanServ
07:33 PM andypugh: pcw-home: Just to report, I have the relay board working from the 7i96 UART now, but at the moment only for writes.
07:34 PM andypugh: Reads return an error, and an overrun with no stop bit.
07:35 PM andypugh: But I didn’t get very long on it this evening.
07:39 PM pcw-home: Maybe RX setup doesnt match TX
07:40 PM pcw-home: might also be BIAS (the will generate an overrun)
07:41 PM andypugh: I think that the PktUART linuxcnc driver might need work. Perhaps a function to clear sticky fault bits, for example.
07:42 PM andypugh: Am I right that with a 100Mhz ClockLow and a 9600baud comms, the filter needs to be 255, as the closest I can get to 104,000 ?
07:43 PM pcw-home: for a synchronous protocol like MODBUS you could actually reset the RX UART before every XMIT
07:43 PM andypugh: And should I be able to set up the setup registers once at the beginning and not touch them again?
07:45 PM pcw-home: Yes but I think it may be prudent to clear the RX (UAR?) every TX/RX cycle
07:46 PM pcw-home: I am not sure if there might not be some combination of RX garbage that confuses the RX FIFO logic
07:47 PM pcw-home: if you enable echo you could check RX with nothing connected
07:58 PM pcw-home: You might try adding bias resistors (300-600 Ohm from +5 to RX-,TX-, 300-600 Ohm from GND to RX+,TX+)