Thanks for that, I came looking for the exact same issue, and feeding 5V into the W terminal worked fine, except for one gotcha:
The calibration procedure for this meter is to power it up with a 20V supply and 1A load, while this W terminal is pulled to ground. I discovered this because I'm using it to monitor the DC input of a dual power supply for a cycle race timing system, with the 5V supply to the W terminal derived from the AC PSU so the meter flashes its undervoltage alarm when a battery isn't connected, rather than just appearing blank. (Similarly, I have an audible alarm on the AC side to warn of generator failure.)
It turns out that my RFID reader and Ethernet switch is close enough to a 1A load to put it into calibration mode when powered up from a ~20V source
Resolved by feeding the W terminal via a diode, so the 5V regulator can't act as a pull-down when unpowered. The voltage drop doesn't appear to be a problem (indeed, the display contrast was slightly worse when fed with the full 5V).