Foiled by a fishing line?

A lot of progress was made this past week, some of it good, some not so good.

First the good: 
1. A new GPS board arrived from China in only 3 or 4 days, so I printed out a holder for it, and glued it to the Plexiglas board that is currently holding the Raspberry Pi circuit board, so it should be well out of the way now, and shouldn't short-circuit against any other stray components.

2. I added alarm software (email and texting) into the main AMOS program, and it seemed to work pretty well. Currently it is just being used for leak sensing, but it could also be used for voltage monitoring, navigation warnings, or even for just sending periodic data updates.

3. I Added the ADC converter chip to AMOS, and hooked up a resistor divider circuit to the battery voltage to enable simple battery voltage monitoring. I also added the necessary I2C code for reading in the ADC values and tested it out a couple of nights ago to see how much the battery voltage dropped overnight while running the main AMOS program:

The test started at 10 pm, and showed some semi-regular voltage fluctuations that I don't really have an explanation for... perhaps the Pi was pulling more load for certain tasks throughout the night, I'm not sure. At any rate, shortly after 6 am, even though the weather was cloudy and a bit rainy, the voltage from the solar panel kicked in, resulting in the sharp voltage increase.

Then the bad:
1. In an effort to improve the look and quality of the wiring inside AMOS, I thought it would be a good idea to use some strip lighting terminals to replace the soldered wire junctions that I had been using up until now. The terminals that I bought off Amazon just used little springs inside them, and didn't have any screws, so there wasn't really enough force to hold the wires in securely. I think I realized that they probably wouldn't work well, but I still persisted in undoing all of the soldering and re-wiring things to use the spring terminals.

2. After the re-wiring, a field test was attempted back at Woolastook. It was a cold, wet, rainy sort of day, but I was hopeful that AMOS would zip off to the GPS waypoints without any difficulty. Right off the bat though, it looked like something was wrong. AMOS seemed to be turning too much to the left. I had to run after it from the start and grab it before it smashed into the gravel bank in the corner of the cove. I righted it, and it seemed to progress mostly in the right direction, but would still sometimes pull a bit to the left. I paddled after it in the kayak, but after a few minutes it veered more sharply to the left and then crashed into the western shore. Here's a map of the failed route (shown in blue, waypoints are red):

It was a bit of a struggle to remove the hatch while the propellers were still going, and I noticed that one of the propellers wasn't really moving and was making an awful noise. A post-mortem back at the house revealed this rock to be the source of the bad noise:

It was lodged in there quite firmly, and had slightly damaged the propeller blades, but I was eventually able to pry it out. After that, I had to fiddle with the bad wiring inside AMOS to get reliable power to the propeller, but once I did, it turned quite well... except that I noticed some sort of stringy material was attached to the prop after it stopped turning. This material turned out to be a considerable length of fishing line... I think I was able to cut and pull it (and some other lake crud) all out:

So between the bad wiring, the fishing line, and the rock, I guess AMOS had some excuses as to why it kept veering to the left. I've partly finished re-doing my soldering, and have ordered some better screw terminals, and prototype boards for holding the ADC circuitry (right now it's just on a breadboard). Hopefully this weekend I'll get a chance to try the Woolastook navigation challenge again.