|
|
|
a circuit board from my 13 year old self
Saturday, December 7 2024
location: rural Hurley Township, Ulster County, NY
While I was dinking around this morning, Gretchen made us both coffee, so we ended up playing collaborative Spelling Bee for awhile before I packed up the Forester and began my drive to the cabin. I knew the cabin would be cold when I got there and that keeping warm would be a challenge just for me, so I decided again to not bring the dogs. Since I was carrying pretty much all the food I needed, I didn't bother to stop at the Hannaford in Cairo, though I did stop at the Stewarts nearest Exit 20 in Saugerties to buy a large VooDoo Ranger Juice Force IPA as a road beer, a large paper cup of coffee, a bag of Spicy Sweet Chili Doritos, which I have come to love. (There is just something about that sweet battery-acid burn on every chip!)
Further on the way, I stopped at the Noble Ace Hardware in Johsntown to get a male hose fitting for the trap in the cabin's kitchen sink. That trap has a plug with that shape, and I wanted to have a way to easily drain the trap without having to remove that plug. The plan was to attach hose to that plug and run it into the basement, where I could have a drain tap to open as a step in the winterization process.
There was snow on the ground along the entire drive, but it got deep as I drove up into the Adirondacks. Woodworth Lake Road had been plowed only as far as the fire department antenna, which meant I had to drive the last 465 feet through six inches of snow. This probably would've been impossible had others not driven through before me and left some tracks for my tires. At the entrance to our driveway, I took out the snow shovel (which I'd remembered to bring) and dug myself out a parking space, and then parked the Forester in that. The walk from there to the cabin was even more of a slog than it had been last time, as there were now about eighteen inches of snow on the ground and there was little evidence of the track I'd made last weekend. I was wearing a pair of black galoshes that have been badly delaminating for a couple years now, but they were better for walking in this kind of snow than my nice high-top rubber boots because their tops cinched shut to keep out the snow.
The thermostat inside the cabin said the temperature was 32 degrees, which was a half-degree warmer than what my upstairs weather sensor had been reporting. As you know from Monday, I was paranoid that perhaps I hadn't drained the warm water system sufficiently. But it was looking like water still hadn't frozen in the cabin; when I stuck my finger into the small hollow at the bottom of the downstairs toilet bowl, I found the water in there was still liquid.
Using lots of cardboard, I started a fire in the woodstove and then fired up the generator to replenish the depleted cabin battery. While running that, I decided to also run the minisplits. I couldn't run the boiler, though, as it was drained for winterization and I wanted to leave it in that state so as to avoid all the work of repriming the pipes only to drain them again tomorrow.
The two things I took out of winterization were the drain of the downstairs toilet (though I would only be flushing it with buckets of water) and the drain of the kitchen sink, where I wanted to implement that easy-to-rever trap draining system. Unfortunately, though, there was something wrong with the fitting I'd bought at Noble Ace Hardware and it didn't seal the plug hole correctly. So I was going to have to plug it and unplug it the conventional way.
Even with a constant stoking of the fire and the minisplit, it took a long time for the temperature to rise to 33 degrees. After that, though, it climbed about a degree every half hour. This meant temperatures remained in the 30s for hours. This wasn't too bad when I was puttering around or under a blanket watching episodes of Alone. But my feet gradually started aching and then went partially numb in places. So at some point I drained some water out of the hot water tank, heated it up on the stove, and then put it in a five gallon bucket for me to soak my feet in. That's definitely the way to deal with the problem of cold feet!
Last evening, I'd managed to successfully get someone else's minisplit-controlling Arduino code to finally compile, though it hadn't been easy. (I was forced to put all the files in one directory and correct the hardcoded paths to account for this.) I'd then worked that code into a branch of my SolArk Co-pilot code (which seemed like the ideal base, since that version doesn't blather away constantly on the serial line, which leaves it open for minisplit communication). This would allow me to send commands to the minisplit using the command-sending architecture I'd built out a couple days ago. Now that I was at the cabin, I wanted to see if this would actually work. So I flashed the code onto an ESP8266-based NodeMCU and used a micro-USB-to-USB-A adapter to plug that device into the USB port on the kitchen minisplit air handler. But when I sent it commands, it not only didn't do anything, but it didn't even mark the commands as "done." When I took the NodeMCU out of the minisplit and plugged it into my laptop, its serial port never registered. That was odd. Maybe its firmware had somehow been scrambled? There was no way to know initially, so I went and got another NodeMCU (they cost less than $2 each!), flashed the firmware onto it, and tried again. Again it failed to do anything. This time, though, as I removed it from the minisplit, I could feel that the CP2102 USB chip was hot to the touch. Had Pioneer gone and monkeyed around with the wiring of that USB port? So I fetched a USB cable with stripped wires at the end and tested to see where the +5v was showing up. Sure enough, the port had been wired so that the green D+ wire and the black ground wire had been swapped. This was no accident either, because the dongle that Pioneer had included to plug into this port also expected to find ground where the D+ line was supposed to be. That's definitely the kind of wiring game that could destroy a device not wired to expect it, and it now looked like I'd bricked two NodeMCUs. Whatever was wrong with them was more severe than dead CP2102 chips. Now they couldn't even boot up and log onto the local WiFi network, something that those chips play no role in. What's particularly galling about this is that the USB port is actually labeled with the official USB logo, something that is almost certainly a trademark violation. So I immediately posted about this in Reddit in an effort to cause as much trouble for Pioneer as I can. People accepting shitty behavior like this allows corporations to let their sociopathy flag fly.
I then proceeded to wire up an adapter to allow my conventionally-wired ESP8266 equipment to plug into the fucked-with Pioneer "USB" port. But by the time I was done making it, my heart really wasn't into that project, at least not for now.
So I turned my attention to the other big project I wanted to work on at the cabin: creating a way to remotely add another 12v marine battery to the battery-backed 12v DC that allows the Generac generator to start even in a power failure. (This same battery also powers the SolArk Co-poilot and the East Basement Remote Controller.) The idea for being able to remotely add (and remove) another battery is to be able to keep a fresh battery in reserve at all times so that the generator can always be started. Without that battery, the various ESP8266s and whatever is being drawn by the generator itself eventually depletes the generator battery, and it can no longer start the generator. There's probably a long period during which it has enough power to run the ESP8266s but cannot start the generator, and it is in those conditions that I can remotely add the second battery, fire up the generator, and then replenish all the batteries with enough power to run for a week or more. The key to getting this right, though, would be to connect the second battery through a fat conductor using a beefy relay. In this case, I intended to use both poles of a relay rated for 30 amps to connect the backup battery's +12v line. I figured 60 amps of power through the relay would be enough to start the generator (or provide enough of a boost to the better-connected 12v battery inside the generator itself).
To control such a relay, I needed to wire in a ULN2003 (or its equivalent). That's a sixteen-pin integrated circuit, so it needed a circuit board to organize the wires going to it. I thought I had a suitable virgin circuit board somewhere, the kind one used to buy in a Radio Shack. But I couldn't find it anywhere. So I ended up using a circuit I'd probably created when I was about thirteen years old that had somehow found its way to the cabin with other macroscopic electronic scrap (which I keep on hand to scavenge parts from). The circuit already had a 16 pin socket soldered to it, along with four six-pin optoisolators. The socket contained a 4017 CMOS decade counter IC, which is basically a digital sequencer that turns on ten different wires in a round-robin manner in response to a clock. I know the kinds of things I was doing with electronics in the years before I got my first computer at the age of fifteen, and they included hardware hacks of calculators to make them behave more like computers. I'd probably used this device to simulate the pressing of a sequence of four buttons on a calculator. The button presses could be faked with an optoisolator, which can bridge any two wires on command. And the specific sequence would be determined by what pins of the 4017 the optoisolators were wired to. (I'd been intrigued by the potential of building a general-purpose computer this way, but had eventually been stymied by the inability to electronically read — and then act upon — the results of calculations.) I removed all the scrappy wires from the board, ones that dated to a time when I was scavenging connector wires from an old tube radio my dad had given me after it died. Then I added pin headers for all the ULN2003 pins, since that chip requires no support circuitry. As for the optoisolators, I left them on the board, since they can be reused to control low-voltage signals in this same application. Even across 43 years, my electronic projects still follow a similar pattern.
Once I had the ULN2003 board wired up correctly and not shorting out, I decided to call it quits for the evening. I mixed myself a series of drinks and huddled beneath a blanket on the couch in the great room watching episodes of Alone on my laptop. The plan was to sleep there tonight so I could be near the warmth of the fire. By then, it was about 11:00pm and temperatures at the thermostat in the hall were in the low 40s.
Me in the still-winterized cabin this evening, drying wood from the outdoor woodpile and trying to stay warm.
Click to enlarge.
That scrap of circuit board from my thirteen-year-old self, after I removed the wires, added pin headers, and plugged in a ULN2003. The small black boxes on the right are some sort of six-pin optoisolator. Evidently those optoisolators all share a common ground on pin 2, so they're probably a 4N25 or similar. (I left the solid uninsulated jumper wires in place.) Note that I have strongly color-coded two pins, one for +12v (yellow), and the other for ground (black). That's labeling from today. The numbers for each optoisolator were written there 43 years ago.
Click to enlarge.
For linking purposes this article's URL is:
http://asecular.com/blog.php?241207 feedback
previous | next |
