LoPy LoraWAN OTAA Deepsleep Example

[jcaron]

Are you using the most recent firmware? Are you using nvram_save (before going to sleep) and nvram_restore (after coming back from sleep)?

[roadfox]

@jcaron i was reading about nvram_save and restore and i also found a code snippet,how to detect if the system is booting or coming from deepsleep but its all puzzle pieces and new to me, so i was hopimg for a bit more complete example

If boot
 lorawan join
else
 nvram_restore
send data
nvram_save
deepslep 2min

Is this approach correct?

[jcaron]

You only need to join once, so you may do that manually at the same time you upload the code. Once you've done that, it's as simple as:

nvram_restore()
send your data
nvram_save()
deepsleep

[roadfox]

@jcaron ok thats what i have now:

  def connect(self, app_eui, app_key):
    """
    Connect device to LoRa.
    Set the socket and lora instances.
    """

    app_eui = unhexlify(app_eui)
    app_key = unhexlify(app_key)

    # Disable blue blinking and turn LED off
    LED.heartbeat(False)
    LED.off()

    # Initialize LoRa in LORAWAN mode
    self.lora = LoRa(mode = LoRa.LORAWAN)

    # Join a network using OTAA (Over the Air Activation)
    self.lora.join(activation = LoRa.OTAA, auth = (app_eui, app_key), timeout = 0) #login for TheThingsNetwork see here: https://www.thethingsnetwork.org/forum/t/lopy-otaa-example/4471

    # Wait until the module has joined the network
    count = 1
    while not self.lora.has_joined():
      LED.blink(1, 2.5, 0xff0000)
      print("Trying to join LoraWAN with OTAA: " ,  count)
      count = count + 1
      sleep(2.5)

    print ("LoraWAN joined! ")
    # save the LoRaWAN connection state
    self.lora.nvram_save()



  def send(self, data):
    """
    Send data over the network.
    """
    # Initialize LoRa in LORAWAN mode
    self.lora = LoRa(mode = LoRa.LORAWAN)
    # restore the LoRaWAN connection state
    self.lora.nvram_restore()

    # Create a LoRa socket
    print("Create LoRaWAN socket")
    self.s = socket.socket(socket.AF_LORA, socket.SOCK_RAW)

    # Set the LoRaWAN data rate
    self.s.setsockopt(socket.SOL_LORA, socket.SO_DR, 5)

    # Make the socket non-blocking
    self.s.setblocking(False)

    try:
      self.s.send(data)
      LED.blink(2, 0.1, 0x00ff00)
      print("Sending data:")
      print(data)
      # save the LoRaWAN connection state
      self.lora.nvram_save()
    except OSError as e:
      if e.errno == 11:
        print("Caught exception while sending")
        print("errno: ", e.errno)

i call connect after a hard reset or boot, otherwise i just call send
so far everything is working beside the fact that LoRaWAN frame counter is always at 1 and the packets are dropped in TTN unless i disable frame counter checks.
should nvram_save/restore not also save the framecounter?

[jcaron]

What kind of deep sleep are you using? Deep Sleep Shield? Pysense? Pytrack?

How do you detect whether you're coming from deep sleep or some other form of startup?

[roadfox]

@jcaron i'm using the default machine.deepslep, it's not for current it's to see if it all will work whenever i get a deepsleep shield. my understanding is that the lopy boots no matter what kind of deepsleep i'm using

thats how i detect the boot and send the data

  n = LORA()

  # Setup network & sensors
  if machine.reset_cause() == machine.DEEPSLEEP_RESET:
    print('woke from a deep sleep')

  else:
    print('power on or hard reset')
    # Join LoRaWAN with OTAA
    n.connect(app_eui, app_key)

  # Send packet
  response = n.send(data)
  print("Received data:", response)

  # put the device to sleep
  machine.deepsleep(sleep_time*1000)

[johncaipa]

@roadfox I also have the same problem, I do not know if they have already found a solution. it seems that the frame counter is not stored

https://forum.pycom.io/topic/1668/has-anyone-successfully-used-lora-nvram_save-and-restore

[JuanMa]

I had the same problem and I only could solve it using confirmed messages.

[jcaron]

Try these two to see if it changes things for you:

• make sure the socket is blocking
• add a time.sleep(2) after sending, before saving and going to deep sleep.

[papasmurph]

@roadfox Late response, but this doesn't look like a correct approach, as the join (and send) might have failed. Next time it will detect that it comes out of deepsleep and assume that it has joined, which it hasn't, so "if boot" must be replaced by something like "if successfully joined and not boot".