what i thought would be a great solar day turned out to be so-so. what was supposed to be a clear day turned out to be cloudy. the sky cleared up by afternoon but our production was already on a downward swing. despite it all, we still managed to production 26.62 kWh (i was predicting 30 kWh at least). after just one day of not biking i could feel my endurance atrophying again. i put my bike lights in my jacket pocket but my headlight was missing by the time i arrived in belmont. it was the rechargeable amotaios 300 lumens light i got back in august (paid $8). fortunately i bought a spare for my mother but since she doesn't ride her bike at nights (nor during the cold months) i ended up taking her light (still in original package). i hate losing things. i'd feel slightly better if another biker found it and could use it, rather than lose it on the road and have it get crushed in traffic.
i wanted to drill additional holes in the backyard tree stump before adding the stump remover powder (which should've been done back in the fall) but my father said it was too cold (temperature in the lower 40's). i didn't think temperature made a difference but that area of the yard was awfully muddy. if we don't treat the stump come spring, we won't be able to grow a new tree in its place. instead i helped my father fix the kitchen sink drain pipe which started leaking a few weeks ago after we initially thought we fixed it back in october when we replaced all the corroded plumbing. it wasn't difficult, the pipes needed to be re-oriented, and he replaced a brass section with a slightly longer plastic replacement. afterwards we upgraded the fluorescent light above the kitchen sink with a similar length LED light. my father thought the LED light would be more energy efficient but it still uses 12W of power compared to the 20W of the fluorescent.
i showed my father the sense monitoring system, something i only recently discovered myself. for $299 you get this easy to install box that effortlessly attaches to your electricity switch box and can then monitor your electricity usage in real time, down to the minute, even knowing when particular appliances are on (e.g. washing machine). even better, for $349 there's an option to monitor solar production as well. this box would solve all our monitoring problem, we'd never have to go down to the basement and read the meter like what my father currently does now.
the project i was working on was setting up one of my raspberry pi 3 with an RTL-SDR dongle so it could receive wireless transmissions from the smart meter in our basement so we can collect electricity data without having to go downstairs every night, which my father has been doing for the past month. i learned that the elster smart meter broadcasts data at 900MHz 4 times a day. first i needed to set up the microcomputer. i started it up with the HDMI port connected to the tv because i needed to toggle out of OSMC and have it load into raspbian instead. once booted, i had to connect to the wifi network manually, using a wireless keyboard. next i opened the admin page of the router and opened the SSH port so i could connect to the raspberry pi remotely from home and work on it from there. i transferred the SSH key so i could log in without password.
i followed this guide to set up RTL-SDR on the raspberry pi 3. the only times i've successfully used the RTL-SDR dongle has been from my android phone, where i managed to track some planes overhead. i tried installing it on my macbook pro one time, but it involved running a listening server and i thought it was just too complicated so i never bothered. i never set it up on the raspberry pi (though i knew it was possible) until now. i installed rtl-sdr and gr-osmosdr (sudo apt-get install rtl-sdr gr-osmosdr). rtl-sdr seems to be the driver for the dongle, but i wasn't sure what gr-osmosdr was. i knew that the final step was to eventually install gnu radio, so i could use the radio block that would decode the elster meter data. the initial installation failed, until i issued an sudo apt-get update and the second time around worked. lsusb could see the usb dongle, but when i performed an rtl_test it failed.
my father left for the airport before 5pm to pick up my 2nd aunt, returning home after having spent almost a month keeping my grandmother company in taiwan. after my father came home we had dinner. my mother made this curry-like rice dish with these big slices of beef that was hard to eat. i persevered and finally finished the whole plate even though i wasn't very hungry. i left for cambridge immediately afterwards.
back at home, i connected remotely to the raspberry pi 3 to try and debug the RTL-SDR issue. i finally fixed it by setting some files and rebooting the rpi3. when i issued an rtl_test it was able to ping the dongle successfully.
i was just doing some research, trying to find more information about belmont's installation of smart meters. they made the upgrade a few years ago with little fanfare, even though in other towns some residents were against it for data security concerns. i then went to the belmont (municipal) light website and something caught my eye that i hadn't seen before: there was a button to "ACCESS SMARTHUB". i clicked on it and couldn't believe what i saw: apparently belmont's smart meters has the ability to monitor electricity usage, not just by billing cycle, but by months, days, and even hours! belmont - which i've always considered backwards and unsophisticated - was doing something cutting edge for a change. it's weird that this feature is not more advertised, or maybe it just recently came online. here was a treasure trove of data that i was trying to collect the hard way via raspberry pi and SDR dongle, all laid out in nice graphs, accessible via browser and even complementary apps. it asked for an e-mail adress and password. i tried logging in as a new account, but it said i'd already created one. so i tried guessing the password my father probably created and managed to get in after a few tries.
i spent the rest of the evening just playing around with all the data. electricity can be broken down into usage, production (e.g. from solar), and net. there's even a handy temperature trend line. it also allowed me to download the raw data, but the formats - CSV or XML - were a mess. instead of nicely delimited, everything was jumbled together in a single column. the data is still useful but needs a lot of manual massaging to get it in working condition. i then tried the app. not as many settings as the browser version, but it still pulls in some good data. one thing i noticed is neither browser or app version can produce live data; the most correct information you can get is for yesterday. that's why a rpi3-sdr pair that can read the info directly from the smart meter is still useful to have. the very earliest tracking data started september 19 2016.