Your approach is pretty well though out.
I took a look at one of my nano-10 boards and traced the output circuit. They appear to use a desecrate power MOSFET with a 5.6K pull up with a red LED. The interesting observation is that the pull up is not tied directly to the 24V input, but to a diode connected to the 24V input.
The single diode will prevent current from the 48V supply from flowing through your load to the 24V supply. The motor won't turn and the output LEDs won't pop because of the reverse bias.
But, there is only one diode in the nano output circuit and it is shared by all 4 outputs. If you have a load connected to 48V and the output is
off, the other outputs will be pulled up to your 48V supply through your load. This may cause some issuses (output LEDs may melt and/or the 5.6K resistors may momentarily emit light/smoke when the other outputs go active. The current that flows thru the other 3 output's LEDs will be sourced through your motor load.
You still may want to consider an external solution. If you are already using a SSR to handle the motor direction issue, you might want to use another SSR for the pwd issue. You may want to look at something called an "h-bridge driver".
Gary d