Here are links I found useful. The boards in the photos look exactly like what I have. If you want me to take photos of my board at better resolution or something let me know and I'll do it.
In my opinion the issue of this device behaving badly is unique to the DA-2000 Numitron clock board. The Numitrons pull much much more current that any of the Nixie or VFD clocks (up to 750mA) and I think what I am seeing is voltage drop through USB power leads (the wires are awfully thin). Perhaps also some of the trashy USB charger/power supply modules I experimented with are poorly regulated and collapsing under load. This would explain the voltage at the sensor board being consistently low.
Also, when the Numitrons suddenly blank during transitions to and from Date display the voltage at the sensor spikes up to full 5 volts briefly causing the sensor to falsely signal that motion is present.
It appears that everything on the sensor board is running off 3.3V except for the "brain" chip. There is no 3.3V 3-pin regulator on the board. The only thing I can figure is that the chip contains a (very poor) 3,3V regulator. Power in on Pin 8, 3.3V out on Pin 11. I think.l
Attached a couple of pics of the sensor I am using. On the lab PSU set to 5V, it works fine - reliable range in the open is at least 4 metres. Behind about 10mm of wood (in the case) it is still at least 2 - 3 metres. It works reliably down to a voltage of 3.4V.
It also works fine oa a cheap Poundland USB PSU.
I had it working once on the clock - I remember thinking it was neat the way the display fades in and out, but I can't seem to get that now!
Curiouser and curiouser said Alice!
Just plugged the sensor back into the clock, and it is working fine on the lab power supply.!
Tried the USB PSU and it won't boot up properly when the sensor is plugged in.
I have added a simple 4V7 Zener regulator to the latest edition of the board. The quality of the sensors seems really variable, and just because it works on my set up doesn't mean that it will work with everyone's.
Another thing perhaps that might influence the problem: My clocks run in fairly dark rooms, and rarely have full brightness. I wonder if you dim them, the still false triggering happens?