Sorry to dump a bunch of big images in this thread but… This is the log using SLEEP_NETWORK_STANDBY and set to call in every 60 seconds so it’s a bit easier to see what’s going on timewise. As you can see the third 60 second interval message after the “device came online” is missing, regular as clockwork.
Not sure about the extra data being a cloud artefact. The following connection always has a much higher TX/RX session usage. If getDataUsage(data) is reading from the modems internals then presumably it’s real spent data, isn’t it?
I’m actually wondering if SLEEP_NETWORK_STANDBY is really working as advertised.
A little more investigation: I noticed that this sequence of three calls (and one missing call) between “device came online” messages has distinct varioations in timing.
The first call (which is immediately after “device came on line” takes longer to connect, but still connects, transmits and disconnects in about 12 seconds.
The second call connects immediately, transmits, but then stays in breathing cyan mode for exactly 1 minutes before disconnecting. This seems to be the reason the second message is missing (it’s gone past the next 60 second cycle).
The third call connects and transmits in a second, then stays in breathing cyan mode for about 12 seconds before disconnecting.
Although I have seen as many as 5 messages between “device came online” messages, the three message sequence will repeat for hours without changing.
The staying connected after tranmitting the message is obviously going to be very wasteful in power, and the frequent “device came on line” connects is very wasteful in data. I’m not sure the result is going to be any better than not using SLEEP_NETWORK_STANDBY. I’m going to make that my next test, although without SLEEP_NETWORK_STANDBY data usage doesn’t work so I’ll just be checking on time.
That’s an interesting find for which I have no explanation, but maybe @BDub has.
Just to clarify one thing, you have actually got a production Electron and not a Beta device?
The Betas do have some pull-ups missing to keep the modem reliably “powered” during sleep.
I got this Electron from Mp3Car Store via Amazon Prime since that was the quickest way to get it (I’m very impatient to try cool new things;). But maybe they’ve been on the shelf a while. How can I tell the age of this board?
The former uses that 60 seconds as a clock including time spent in connection, while the later doesn't include the time spent in connection. So the logs show (more or less) accurate 60 second intervals (except for the missing message) with the former, but regular 90 second intervals with the latter.
So not using SLEEP_NETWORK_STANDBY seems to give a pretty much expected result, while with it it's a bit weird.
I wish I could tell what the data consumption was without SLEEP_NETWORK_STANDBY but, data usage doesn't work without it. In theory SLEEP_NETWORK_STANDBY should be the ideal compromise between data and battery usage, but I think it has some foibles!
I just brought a 2G electron (G350) to Japan with me on a business trip and tried registering the SIM by entering the CCID and creditcard for the first time. Still getting blinking green light after about 10 minutes and a cold reboot.
I havent used it in the US yet. Will be in the US tomorrow and can test there… But would be nice to know if it works in Japan.
The 2G (G350) Electrons don’t work in Japan. They’ve already turned off 2G service in Japan to re-use the frequencies for newer services. The 3G Europe/Asia/Africa model does work in Japan.
@JethroNull I just setup the same code at 60 second intervals and i’m seeing the same thing as you. I’m not liking what I’m seeing here so I’ll have to investigate this.
@BDub, Glad you’re on the case. Happy to help out too if you want me to try anything. If my charting is any help to you I can share the webhook so you can set up your own really easily? (See, I’m stealthily trying to get everyone to try Grovestreams
Any ideas?