long (~1 second) wake up times for mDot

[Eric Tsai]

I have a simple program that wakes up from sleep and starts timing how long an I/O pin is high. The pin happens to be wakeup pin. I’m consistently at least a second short on this duration. I think it might be the wake-delay. How long does it take the mDot to wake from sleep and start executing user code? Is this documented somewhere? Is there any interaction between the wake-delay and user configuration on the mDot?

Prior to actually starting the timer, I’m doing these configuration items. I didn’t expect them to have much of any delay:

1) dot->setLogLevel()
2) dot->setTxPower()
3) dot->setAntennaGain()
4) dot->setAck(1)
5) dot->getNetworkJoinStatus()

LibmDot: 14
mbed-rtos: 94
mbed: 107

Thanks,
Eric

• This topic was modified 7 years, 7 months ago by Eric Tsai.

[Andrew Lindsay]

Not sure about your versions, but for libmDot 14 I’ve been using mbed 121 and mbed-rtos 110 as the latest version combinations that work without error. Later versions of mbed-rtos wont compile at all.

There may be some delay in instantiating the libmDot library depending on what join mode you are using. If using AUTO OTA then this is likely to be reading the preserved session data from flash. You might like to try different join methods and see if this makes a difference, for example MANUAL might be quicker, but without access to the code we cant see exactly what its doing internally.

If you’ve saved the config then I wouldn’t expect you’d need to set the TxPower, AntennaGain and Ack each time.

Andrew

[Author]

Posts