|
|
|
even more remote control
Tuesday, December 24 2024
At some point today it occurred to me that there are a bunch of values used by each of my remote control ESP8266s that it would be useful to change remotely while they are in operation. These include the time to wait between each poll of the backend as well as the time to wait between each recording of probe data (which now is still mostly weather-related). But the single most useful parameter to change is one that thus far I've never used: the amount of seconds to go into deep sleep after doing the things it needs to do. Overnight I'd tested an ESP8266 running my firmware with a 10 amp-hour battery and found that it could only run 24 hours with that amount of power, suggesting it burns 500 milliwatts constantly. But if I could get it so that it's only wide awake for one tenth of one percent of the time (which it would do if I logged data only once every half hour), it could run for a thousand days on that same battery. The benefit of switching it into such a low-power regime remotely is that I could then switch it back out again if I needed more detailed data or wanted to remotely-control devices with less command latency.
To implement this new functionality, I had to make some configuration values into variables (that is, no longer declare them as constants). I also had to change the timing of the logging so that it happened immediately after the device starts up (that is, not waiting for the first data_logging_granularity number of seconds before logging), because once it goes into deep sleep, it resets, meaning it has to have gotten everything done it needed to get done or it will never get to those things. This also means that commands forcing it to go into deep sleep need to never be marked "done," since the device will need another such command once it re-awakens, as all dynamically-changed states revert back to defaults after the device comes out of sleep. (I could change values in an emulated EEPROM in the flash storage, but I'd like to minimize wear on that, at least at this point.)
This evening Gretchen and I went to the Garden Café in Woodstock for dinner. It was a special reservation-only night, and originally we'd planned to go there with our friends Cathy & Roy, but their daughter (who is currently in an exhausting medical residency) was having trouble, and they needed to support her. The menu at the Garden was full of specials, including a Thanksgiving-style feast called "a Christmas miracle." Most of these didn't interest me, so I went for a simple TLT, though I did have the mushroom skewers with peppers appetizer special. Gretchen got a drink special that was some sort of wine with cranberry sauce and a skewer of candied cranberries that she was very excited about. Because it was a special occasion, I ended up drinking two Ommegang Abbey Ales. Over dinner we discussed my continued joblessness as well as my mother-in-law asking me to do the cover art for her self-published memoir. It will have a title consisting of the names of two historical figures as well as the name of an endangered species of great ape, which are meant to stand for the things that were important in her life. But I think she's taking things a bit too literally by wanting the art to include depictions of those three named things. (This reminds me of YouTube videos where everything mentioned in the narration has to be reflected in short, unrelated video clips from stock libraries, which I find disrespectful of the intelligence of the viewer and a production choice that I do not think will age well.)
On the drive home, Gretchen took us a bit out of the way to drop off cookies at the log cabin off 28A where her friend Greg the Librettist spends his time upstate.
For linking purposes this article's URL is:
http://asecular.com/blog.php?241224 feedback
previous | next |
