Warning! Don't unplug feeders with power on...

[Mike]

So after a 2AM repair session to get the machine working for a job that _HAS_ to be done over this weekend....

I discovered what an utterly  terrible idea it was to mix mains and RS232 on the feeder connectors.
Job was all set up for PCBs arriving next day, decided to move a feeder, and suddenly, all feeders not working.
I figured it would be a serial port issue, ran feedertest on a laptop to verify the feeders were OK, then troubleshooting the dead PC serial port.

Mine is the type that has the two COM ports on the DSP board, and instead of proper UARTs, they use 87C51 microcontrollers for some bizarre reason. For a 9600 baud tx-only interface FFS - what planet were they on?
It took out not only the RS232 driver but this MCU. Luckily I had a pile of scrap boards ( Thanks Cubik!), and I got one to work with armtest, verifying that it had a good MCU for COM1.
All the DSP boards I had which didn't hang the PC seemed to have the same fault, so clearly a definite issue.

One thing I found - the feedertest utility doesn't seem to work through the DSP board UARTS - I initially tried to bypass the DSP board and run via the PC COM ports - I had feedertest running via the PC COM port but not in RVPlace. I'm not sure if the ports on the DSP even act like normal COM ports.

I assume there is some way to get the RV to not use the DSP card ports and use the PC ones - anyone know details ?
I was getting too desperate and sleep-deprived to wait through yet more WIN98 reboots to experiment.

I can't understand why they did the feeders like this - only one ever needs power at a time so a 24V supply from the main box would have been a much better solution - cheaper, safer and would allow cheaper connectors. Not to mention using RS232 and +/- supplies for the receivers when RS422 would have actually used fewer wires and no need for oddball bipolar supplies.
I guess there is probably some history there but it boggles the mind how they came up with such an obviously terrible design given that the rest of the machine is pretty reasonable.

I've now put a local mains power switch in line with the feeder cable, and a tranzorb on the serial line.

Rant over...

[Cubik]

Hi Mike, I once had a live to RS232 short in the back of one of the shells on a link cable. It killed pretty much everything including the PC PSU but amazingly, the PC was ok once I changed the PSU. On this system, the com ports were on the MB and unsurprisingly were not found by W98 any more, so I managed to add a 2 com port bord and got the thing up and going again.

It was a terrible idea! We ended up doing a periodic inspection on all feeder cables as a precaution. Im glad the box of bits came in useful!