Thanks. The diode is correct.

I've not been clear enough in how to measure the voltage, but the bit that you've said that you get 2-3 volts across the relay when the spindle should be on... that means you're dropping 3 volts across the relay coil which at 650R (roughly what it should be, given Jazz's quote of 36mA draw current) gives a switching current of around 4.5mA - bottom end of the performance of the opto-isolator of the NVEM and certainly not enough to actuate the relay.

If you are prepared to PM me your address I'll pop a gadget in the post (it's just the PNP transistor set-up... but I'll test it works before posting). Or, if you're happy to source/wire one up yourself then go for it.