Hi, I would like to use a NO reed switch in a situation that requires an NC switch. It is supposed to send a pulse through a cap like for the Arduino RST, so I need a certain fast current inrush. The setup is battery powered with the aim of long life, and the switch is normally inside a magnet field. So I fear a simple circuit with a pullup will either drain my battery too fast or will limit the current too much to cause a decent pulse (in the simulation, the voltage doesn't go up enough). Any idea? Would a curcuit with a PNP(-Darlington-)transistor (to boost the current after a pulldown setup) be an option? I feel I didn't quite get it right yet, and it is hard to test which values would work reliably without a digital oscilloscope.
Ok, I played a little more. Would this be feasible and stable in practice? As I said, without an oszilloscope, I guess it would be hard to test in practice if a solution just barely works, or gives a stable signal.
The slow time constants would not be a problem, and the consumption of 3.3uA would also be fine.
You could use the pulseIn function to detect either a rising or falling edge, so you don't need an invertor...
Sorry, I was unclear. The whole project is going for long battery life on an ESP8266. It is supposed to be a door contact. I posted a more general advise question in the Project Guidance forum, but I didn't get any response, so I posted a more concrete electronics question here. The uC is supposed to sleep all the time. I need the switch inverted to send the wakeup pulse over a capacitor to the RST pin to wake up when the door is opened ("out" behind cap in the simulator). While the door is open, the ESP will wake up periodically to check if it is still open (by checking "out" before cap in the simulation). Since this is an apartment door, times in which the door is open should be short, detection failures could be dangerous (left door open) and exact time is not critical, so I am willing to spend some power there instead of adding a second reed switch (or circuit on the first one) to wake up when closing.
So you have a NO reed in a magnetic field - so it's actually normally closed. Hence it will normally draw current, even if small.....
You can buy changeover reeds - that would avoid the problem
Use a hall sensor? They don't wear out or stick. Wake up and power your circuitry every few minutes to check the door. Jamming the reset pin like you suggest sounds dodgy to me, using the watchdog timer to wake periodically is fairly common practice.
"Jamming the reset pin" is in general exactly how you wake up an ESP8266 from deep sleep. The timer for timed wakeup does exactly the same, that is why you have to connect pin 13 to reset to use wakeup from deep sleep on the ESP. The pulse via a cap is also exactly how the Arduino reboots when flashing a sketch. I fail do see how any of that is dodgy. Using a hall sensor, I would burn MUCH more power (the ESP is a power hog, anyway) with considerably worse resolution.
Using the reed spdt is probably the way to go.