lte.deinit(dettach=False) with machine.deepsleep

[kjm]

I'm running the latest firmwares on a gpy & its 4G modem connected to Telstra cat M1 here in AU. The good news is the new modem firmware drops current to near zero in deepsleep with the modem disconnected from the Telco. Just for the hell of it I tried deepsleeping with the modem still connected. It didn't work of course, the deepsleep current stayed up at 50mA and there was no connection after deepsleep had expired anyway.

Remaining attached in near zero power deeplseep is a great improvement but will it ever be possible to be remain connected as well? I ask this because the time it takes to disconnect before deepsleep (around 7s) is much greater than the time it takes to send the data (less than 1s).

[Davidjames3321]

This post is deleted!

[andreas]

Dear @kjm,

@kjm said in lte.deinit(dettach=False) with machine.deepsleep:

We've found it's not a good idea to routinely reset the modem in battery powered/deepsleep applications. If the modem is not reset attachments take only a second or two but after a modem reset they take 3-4 times longer.

Sure. Sorry that I haven't told you that we are currently not using the modem at all and just want to turn it off as quickly and thoroughly as possible.

When actually using the modem and going to deepsleep, it should probably not be reset and detached so it will keep its association with the cell infrastructure.

Thanks and with kind regards,
Andreas.

[kjm]

@andreas We've found it's not a good idea to routinely reset the modem in battery powered/deepsleep applications. If the modem is not reset attachments take only a second or two but after a modem reset they take 3-4 times longer. I'll have to check all these things I've learned all over again of course with the 1.20.1.r1 in case the rules have changed.

[andreas]

@kjm said in lte.deinit(dettach=False) with machine.deepsleep:

the time it takes to disconnect before deepsleep (around 7s) is much greater than the time it takes to send the data (less than 1s)

We just mitigated this delay through [1] by not invoking the "detach" operation on lte.deinit() like

lte.deinit(detach=False, reset=True)

[1] https://github.com/hiveeyes/hiveeyes-micropython-firmware/commit/ec7c80e1219fb119c2881a0019cecdfad220f29f

[kjm]

Getting 20uA on a gpy(version='v1.8.6-849-dfa1176 on 2018-11-12) with catm1(39529). Or at least I was, won't attach now. Every time I get something working on this beast something else breaks!

[danielm]

@kjm
What deep-sleep current level are you able to achieve? I can't get it under 100uA. My G01 is running Sequans firmware NB1-40343.