Bit more progress but no screenshots.
Went through the RS-422 connection as it'd be nice to have both options if either go bad again, and traced it through the same way. Starting at near the port on the pcb inside the TNC itself, then inbound via transil diodes.
Unlike the rs232, inbound and outbound are handled by different ic's, because its differential pairs.
The two ics are 75als194 & DS34C96TM's, the DS34C86TM is the reciever, and the 75ALS194 the transmitter, both convert RS-422 to normal TTL serial signals.
These hook up via optocouplers to the UART, but on different pins, RXDB(11) and TXDB(13) on the uart, its a dual channel device, so RS-232 is on RXDA/TXDA. Same pinouts as the RS-232 on the opto's so tracing much easier.
Nothing on the scope at this point.
Then added jumpers for CTS->RTS & -CTS->-RTS with a loop in the plug, and DTR to DSR pair pins the same. Despite being software handshakes I think it NEEDS these mandatory. Nobody tells you this online, its all "its your cheap usb adaptor at faut" etc
Then I started to see activity on the UART pins, weird out of sync waveforms coming into RXDB and confused responses coming back to the PC via TXDB.
It was then I found the smoking gun I believe, the wire in the usb adaptor was broken internally and barely making contact. I remade the lead and it went clean and connected.
Like the RS-232 its flaky as hell at 115200baud data rate, but I've managed to make it negotiate at 9600 baud with no config changes. I think it needs the optocouplers replacing the same as the RS-232 to work at 115200 baud as they've gone soft with age. But I have proved the cheap usb RS-422 dongle works now, and know the correct wiring configuration.
I'm going to get one of the RS-422->ethernet boxes and mount it in the machine cabinet next to the tnc for remote access. It already is wired for ethernet. The serial bridges are about 30 quid and saves messing round with laptops and dongles balanced on tables nearby. Now I have a known working RS422 combo, I can see if the adaptor itself is the cause of any problems. And I can keep the lead length very short to try to cut down on noise.
Get a machine with a commercial control they said, it'll be good they said...
Went through the RS-422 connection as it'd be nice to have both options if either go bad again, and traced it through the same way. Starting at near the port on the pcb inside the TNC itself, then inbound via transil diodes.
Unlike the rs232, inbound and outbound are handled by different ic's, because its differential pairs.
The two ics are 75als194 & DS34C96TM's, the DS34C86TM is the reciever, and the 75ALS194 the transmitter, both convert RS-422 to normal TTL serial signals.
These hook up via optocouplers to the UART, but on different pins, RXDB(11) and TXDB(13) on the uart, its a dual channel device, so RS-232 is on RXDA/TXDA. Same pinouts as the RS-232 on the opto's so tracing much easier.
Nothing on the scope at this point.
Then added jumpers for CTS->RTS & -CTS->-RTS with a loop in the plug, and DTR to DSR pair pins the same. Despite being software handshakes I think it NEEDS these mandatory. Nobody tells you this online, its all "its your cheap usb adaptor at faut" etc
Then I started to see activity on the UART pins, weird out of sync waveforms coming into RXDB and confused responses coming back to the PC via TXDB.
It was then I found the smoking gun I believe, the wire in the usb adaptor was broken internally and barely making contact. I remade the lead and it went clean and connected.
Like the RS-232 its flaky as hell at 115200baud data rate, but I've managed to make it negotiate at 9600 baud with no config changes. I think it needs the optocouplers replacing the same as the RS-232 to work at 115200 baud as they've gone soft with age. But I have proved the cheap usb RS-422 dongle works now, and know the correct wiring configuration.
I'm going to get one of the RS-422->ethernet boxes and mount it in the machine cabinet next to the tnc for remote access. It already is wired for ethernet. The serial bridges are about 30 quid and saves messing round with laptops and dongles balanced on tables nearby. Now I have a known working RS422 combo, I can see if the adaptor itself is the cause of any problems. And I can keep the lead length very short to try to cut down on noise.
Get a machine with a commercial control they said, it'll be good they said...
