Certainly in an industrial environment, the situation is different. For one thing, you are not necessarily in control of where things are used and given that there might be a mix of three-phase supplies plus single-phase supplies coming from different phases, "ground" potentials all over the place, then relying on a connection between grounds on two bits of kit might not be a good idea. After all, it's one reason why ethernet is galvanically isolated. There is also the "make it idiot-proof" requirement. In my case, all my kit is being fed off a single 13A multiway extension off a single 13A socket, and I am the only idiot allowed to use the kit! After all, similar reasoning says that I haven't fitted door-open trips on my control box where that would be entirely unacceptable in industry. A relay is definitely the proper way to go, but I've run out of mounting space and I don't have one to hand. In addition, the UCBB outputs are each opto-isolated on the board and feed distinct mosfet output devices rated well above my requirements, so while a fault might take out one channel, I have spares. And a VFD fault major enough to cause such a failure is likely to be a major VFD failure anyway.

For those who might be wondering, the digital inputs on the HY VFD are sufficiently isolated that a direct connection to the UCBB works fine. I have had the thing running this evening. Just like the analogue output from the UC300ETH which connects to the VFD analogue in and ACM pins. I mentioned above that I was going to need a 16-pin IDC connector to pick up the analogue signal but as I finished typing that entry, I suddenly realised that the 2.5mm pin spacing, etc, matched that used on PC motherboards for USB connections, etc, and I was able to find a surplus connector that I could use. I think that this is the equivalent of the connectors that someone else (forgive me for forgetting who at this point) mentioned - thanks for reinforcing my own thoughts!

But I have to confess that I did shoot myself in the foot with the UCBB - stepper driver connections. I was replacing a motion controller with differential outputs, but for convenience wanted to reuse the UTP cables between the two devices. I used the "solid" colours for the signals and just trimmed back the "stripe" connections, and made up a set of jumper cables to provide +5V to the drivers. Bit of a shame that without thinking, I "assumed" that the solid colours connected to the -ve side of the stepper driver inputs so having removed the "stripe" connections I replaced them with the +5V feed. I really should have looked at the markings on the drivers as I had the polarities reversed and it took me most of the evening to realise it. I knew it was something silly that I had done, but didn't imagine just how silly...

You are allowed to fall off chairs laughing