Test program for AM312 PIR works on Pi Pico W but not on Pi Pico 2 W

[emailman]

The test program below for the AM312 PIR works as expected on my Pi Pico W running on MicroPython 1.21.0 from Anvil Works, but shows no activity when run on my Pi Pico 2 W running MicroPython 1.25.0 from micropython.org. I've tried more basic test programs, different pins, and a spare AM312 with the same result. I've also tried using the built-in pull-down resistor and an external resistor with the same result. Any ideas what to try next would be appreciated.

import machine
import utime

# Define the PIR sensor pin
pir_sensor = machine.Pin(2, machine.Pin.IN)

def motion_detected(pin):
    print("Motion detected!")

# Set up an interrupt to trigger when motion is detected
pir_sensor.irq(trigger=machine.Pin.IRQ_RISING, handler=motion_detected)

print("Waiting for motion...")
while True:
    utime.sleep(1) 

[Josverl]

narrow in on the problem by looking at 'one change at a time' :

• what is the behaviour with a Pico W with v1.25.0 ?

[robert-hh]

I tried that code snippet on a PicoW with both a generator and a similar Sensor, and it worked well. Easy enough. So you could test if the sensor fails, or another Pin at the Pico. Be aware that the number supplied to Pin() is a GPIO number, not a board pin number. So board pin 2 is GPIO1, and GPIO2 is board pin 4.

[GitHubsSilverBullet]

The test program below for the AM312 PIR works as expected on my Pi Pico W running on MicroPython 1.21.0 from Anvil Works, but shows no activity when run on my Pi Pico 2 W running MicroPython 1.25.0 from micropython.org. I've tried more basic test programs, different pins, and a spare AM312 with the same result. I've also tried using the built-in pull-down resistor and an external resistor with the same result. Any ideas what to try next would be appreciated.

If you have a series resistor between the sensor output and the GPIO input pin, keep in mind that the RP2350 has a tendency to latch up once in the HI state. One needs around 8kΩ or lower towards ground for the GPIO to be able to toggle back to LO.

That's a different behavior than with the RP2040.

[GitHubsSilverBullet]

I did see one transition from low to high, but it never returned low and I can't repeat the results.

That's the Latchup I've explained (quite unsuccessfully) earlier.

[emailman]

This problem has taken a turn to the weird and possibly a solution. I took all the excess wiring off of the breadboard and just wired the AM312. Using the program below (note the use of the built-in pull-down), I could get the signal to go low by touching the jumper wire on the sensor's output. Further discovery was that it seems to work OK if I use a long jumper wire rather than a short one!? Maybe it picks up interference from the Wi Fi. If no one has any other suggestions, I'll consider the long jumper wire as the (weird) solution.

from machine import Pin
from utime import sleep

# Define the PIR sensor pin
pir_sensor = machine.Pin(16, Pin.IN, Pin.PULL_DOWN)

print("Waiting for motion...")
while True:
    if pir_sensor.value() == 1:
        print('Pin High')
    else:
        print('Pin Low')
    
    sleep(.5) 

[robert-hh]

I understand that the input goes low if you touch the wire and it works better with longer wires. If it is just the wire connected to the Pin, that that is not surprising. What you pick up is the 50/60Hz hum from the mains supply of your room with your body as antenna. The pull-down may bring the input to a level where is is close to the low/high switch level at about Vcc/2.

The typical wiring of a board with AM312 sensor which I found has a 20k series resistor at the output. That is to large to overdrive the latch-up effect @scruss mentioned, and with a 5k pull-down the input level stays low. So do not use a pull-down resistor less than ~40k, which is the value of the internal pull-down. If that is the fact, you have to lower the output impedance of the sensor be either:

• replacing the resistor, which could be tricky.
• adding a driver at the output to lower the impedance, That can be a single transistor like a 2N7000.
• some software means. In the discussion here https://github.com/orgs/micropython/discussions/15621 is was mentioned that the latch-up disappears if one does a ADC read on this pin. There also a reset Pin code shown there at the end of the thread.

[emailman]

@robert-hh Thanks for all the info and suggestions about this issue. After reading the discussions/15621 you suggested, I realize my issue is just one case of several where the Pico 2 behaves (much) less well than the Pico 1. Since I'm interested in using the Pico for teaching programming, I'm just going to recommend in my exercises that the original Pico be used.

[GitHubsSilverBullet]

Especially for education purposes, it might be a good idea to explain to the students, that there's no guarantee that an input (any input on any controller for that matter) is guaranteed to work with only a few micro amps sourcing or draining. Personally I see this as a very nice chance for a bit more in depth education. Along the lines of ›Always use capacitors at any Vcc pin‹ and ›Do not expect a rats nest on a breadboard to work as intended‹
And most important ›Do not assume, measure and verify‹

Anyway, with less that 500µA current it works quite well.