docs: compiling the runtime for Raspberry Pico W

[aleppax]

I've compiled the micropython runtime with the picosleep module for a RP Pico W, the only difference from the documentation is that I've issued BOARD=PICO_W at make time.
MicroPython v1.2-pico-deepsleep on 2023-01-09; Raspberry Pi Pico W (sleep) with RP2040
But I'm confused because the current measured when I try to run picosleep.seconds() never goes below 15 mA. This is far above what I was expecting and what others have found. Is it a problem of compilation? Has something changed?
maybe the CYW43 chip requires his own command to enter in sleep mode, and based on the comment of an user on Stack Overflow wlan.deinit() should do, as far as I can understand from the driver source, but in my test it does't reduce the absorbed current below 12.85 mA.
Are there precompiled binaries for the Pico W that perform better?

[aleppax]

I don't know what I was measuring yesterday, I'm sorry. but I've tested again everything and now the power absorption both with picosleep.seconds(), machine.lightsleep() and machine.deepsleep() is 1.2 mA.

[jmodrako]

Have you been able to get lightsleep working with WiFi?

[aleppax]

strange, I thought it was solved, but I recompiled and connected via usb cable I only get 13mA.
I'll try later with a battery
look at this video
https://youtu.be/GqmnV_T4yAU?t=462

there's also this modified firmware with a different implementation and uses the custom internal module
import picosleep
picosleep.seconds(60)
https://github.com/ghubcoder/micropython-pico-deepsleep/releases/tag/v1.2-pico-deepsleep

[jmodrako]

In my case it behaves like the flow of the program does not even stop in the place where light sleep is called. And thanks for the video, I can see some additional calls to gpio to actually shutdown Wi-Fi radio. I need to check it.

[aleppax]

sometimes I reach 1,4 mA. Maybe it's related to the time in between one command and the other because I succeded only with commands issued one by one via REPL. But can't do it again without rebooting the board.

[jmodrako]

I think that I've finally managed to put Pico W into lightsleep. I think that the crucial part was to set low/high on the Pin.23 (WiFi on/off).

Here is my code:

import time
from machine import Pin, ADC
import network
import urequests

ssid = 'ssid '
password = 'password '

wlan = None
wlanSw = machine.Pin(23, Pin.OUT)

def connect():
    global wlanSw
    wlanSw.high()
    
    time.sleep_ms(500)
    
    global wlan
    wlan = network.WLAN(network.STA_IF)
    wlan.active(True)
    wlan.connect(ssid, password)
    while wlan.isconnected() == False:
        print('Connecting...')
        time.sleep_ms(500)
    
    print("Connected!")
    print(wlan.ifconfig())
        
def disconnect():
    print('Disconnecting...')
    
    global wlanSw
    wlanSw.low()
    
    time.sleep_ms(100)
    
    global wlan
    wlan.disconnect()
    wlan.active(False)
    wlan.deinit()
    wlan = None
    
    time.sleep_ms(100)
    
    print("Disconnected!")

def measure_vsys():
    Pin(25, Pin.OUT, value=1)
    Pin(29, Pin.IN, pull=None)
    reading = ADC(3).read_u16() * 9.9 / 2**16
    Pin(25, Pin.OUT, value=0, pull=Pin.PULL_DOWN)
    Pin(29, Pin.ALT, pull=Pin.PULL_DOWN, alt=7)
    return reading

def mesureVsysAndSend():
    vs = measure_vsys()
    r = urequests.get("http://192.168.2.214:8855/?vsys=" + str(vs))
    r.close()

def doLightSleep(sleepSeconds):
    time.sleep_ms(100)
    machine.lightsleep(sleepSeconds * 1000)
    time.sleep_ms(100)

while True:
    connect()
    mesureVsysAndSend()
    disconnect()
    doLightSleep(sleepSeconds=60)

In the lightsleep there is 1.5mA +-1mA current draw, so I'm happy :)

[aleppax]

Yes, good job! It works.
time.sleep_ms(500)
this half a second of sleep after wlanSw.high() seems to be important.
According to the datasheet of CYW43439 (Figure 15. WLAN Boot-Up Sequence) something around 80-100 ms should do.

I always obtain 1,87 mA with a cheap multimeter, but for me is ok.