Just home so opened the Node case and glad to confirm despite ageing my recollection was accurate.
This screenshot is common to the PR6/7 using CH2P if it is a undirectional flow.
Looking at the circuit afresh I see no reason why the internal input pullup of IRQ should not work either through a drop resistor or possibly direct, since all it does is sink the signal, which the IRQ will pick up as Change/Falling/Low. From vague memory of the sketch, mine is triggered on LOW.
1 - Unlike Reed switches (my gas reed duration is ca 6 seconds) mosfets are fast, so the pulse duration is so short it will not register on your multimeter hence the perceived uninterrupted 5v. The PR7 spec you posted says 10ms duration...
2 - If envisaging a standalone radio node such as deployed here, you may be better going with a 3.3v PM to begin with rather than mess about with level converters later...
Hope this helps...
