|
|
|
big snow and capacitors
Sunday, January 25 2026
Due to an unusual alignment of weather patterns, we experienced an epic snowstorm today of the sort we hadn't had in over a decade. The storm extended across the eastern two thirds of the United States, with snow stretching from Texas to the Atlantic Ocean. Much of that region was expected to get ice, though temperatures here were so cold that the snow fell as a dry powder with almost no cohesion. This made it tend to roll off surfaces that weren't horizontal. So our roof accumulated little snow, while the ground and deck below it was piled with round-topped mounds. Out on the laboratory deck, the spaces between the planks were wide enough for snow to slip through, so each plank developed a steep ridge of snow running down its center.
We kept indoors most of the day, though Gretchen managed to walk Charlotte up the Farm Road and back this morning. Later this afternoon, though, I was suddenly inspired and wanted to see what the snow was like. So Gretchen and I ended up shoveling out around the cars and then a narrow path all the way out to the road, where I dug through the massive ridge of compressed snow created by the snowplow. At the time, though, a plow hadn't been through in awhile and there were two inches of snow on Dug Hill Road, a depth I don't think I'd ever seen. (Since the plows are kept further up Dug Hill Road, they tend to keep it well-plowed in even the worst of winter storms.)
This morning we'd done the usual living room with coffee, fire, and Spelling Bee routine. One of the panagrams in the Spelling Bee today was "capacitor," which Gretchen got before I did. She thought that was particularly ironic given that she has no idea what a capacitor actually is. So I explained that a capacitor typically consists of two metal plates located close to each other but not touching and that a charge placed across them can be stored. She wanted to know what use that was, so I explained how it could be used to filter out annoying frequencies, allow frequencies to pass through, or to store electricity to backfill momentary voltage sags (like tiny batteries). Later I retrived a large 11,000 microfarad capacitor from the laboratory, charged it with a 9v battery, and demonstrated what would happen when its terminals were shorted (a visible spark was produced).
Later tonight, I went out to measure the snow in a place in the yard reflecting its natural depth and it came to 18 inches. But the storm was far from over.
From this afternoon until well into this evening, I was back working on my I2C bootloader project. I was in such need of complex debugging data from it that I briefly considered making my own bit-banged serial system using a clocked serial system (similar to, but much simpler than I2C) that would allow me to receive debugging data from the bootloader on a special Arduino programmed to receive it. This would avoid all the timining issues that seemed to be plaguing my attempts to use the Arduino's serial hardware. But then I realized that the place where I'd put the code to initialize the serial hardware was never actually being reached in my particular Arduino. Once I put the initializing code there, the serial port started working correctly, albeit at twice the baud rate it should've been. But then the code stopped showing serial data for unknown reasons and I became obsessed with making the Arduino use its internal 8 MHz crystal instead of its 16 MHz crystal. I figured if I could get that part right, I could then unwind the baud discrepancy, which I assumed was the cause of all my garbled serial data. But after issuing one of the fuse commands suggested by ChatGPT to the Arduino Nano, it became unresponsive. I should've known better than to then ask ChatGPT what to do next, because it was full of enthusiastically-delivered bad ideas. It came up with at least two different sketches run by another Arduino that would supposedly make my bricked Arduino Nano responsive again, but none of them worked. It also kept suggesting to me that if only I applied a 16 MHz signal to one of the crystal inputs, it would come back to life. But it had also said that once certain fuses are set, it ignores the crystal and signals from the pins. I kept pointing out discrepancies between the things it was telling me, and it kept being ignorant of things it had said. But then at some point it fessed up and said that I'd been right to be skeptical, that there was no solution to my problems other than a high-voltage reset. At some point I decided to research high-voltage reset the way I used to research things before ChatGPT, that is, with internet searches. There I learned that high voltage reset requires sending 12 volts to the reset pin, but that the reset pin must not be connected via other circuitry to anything else when this is done. In the Arduino Nano, the reset pin is connected via a one kilohm resistor to the VCC pin and via a 0.1 microfarad capacitor to the DTR on the USB interface chip. I didn't want to bother setting up a high voltage resetting system tonight, but I did fire up a soldering iron to remove those two tiny surface-mount components so I can zap it later. (I will then replace them with larger through-hole components somehow wrangled onto that tiny congested circuit board.
Later I moved the serial-initializing code to a different part of the bootloader and serial started working again. So I decided to leave it there and be happy with what I had.
Today Gretchen decided to cook up a bunch of barley and beans in separate batches using the InstantPot. Later tonight she made cheesy barley with beans, and it was okay, but it had that weird aftertaste that things made with soy milk frequently have, and it was a little hard for me to eat because of it. After Jeopardy!, we watched the penultimate episode of season one of The Rehearsal, and it was pretty good, though it was also a little taxing for our meagre human brains.
For linking purposes this article's URL is:
http://asecular.com/blog.php?260125 feedback
previous | the future |
