This is a graph of solar power currents over about 12 hours. It began near the end of the day with the battery mostly charged from the morning's light.

The battery current explained:

The battery just has a small trickle charge of about 10 mA to maintain it. About ¼ of the way into the graph, the sun begins to set and the battery current begins to waver as the solar panel can no longer maintain the trickle charge. About half way through the graph the light has dropped enough that the solar panel is just able to break even. At 0 mA the solar panel is able to supply the computer but give nothing to the battery. After this point the battery current becomes negative as the computer is now drawing on the battery for what the solar panel cannot provide. The last part of the battery trend shows the battery discharging steadily. At this point, the sun is no longer providing power—just the battery.

The computer current explained:

The computer current is mostly flat, except for some fetchers right at the beginning. Because I got this setup running near the end of day light, the current numbers were pretty small. I needed to be sure things were connected properly, so I had the Sun Dragon start running some heavy computation to increase it's current draw. I did this three times, visible as three current spikes. Aside from these the current is pretty steady. There is, however, a slight increase as the sun begins to go down. This is because as the sun sets, the system voltage drops from around 14.5 volts to around 13 volts. The power required by the idling computer remains the same. So when voltage drops, current must increase. Thus the slight increase in current as the sun sets.

The solar panel current explained:

The solar current seems to track the battery current, except for the spikes when the computer is drawing higher current. The spikes at the beginning show the solar panel has enough light to fully supply the computer and the battery trickle current. In fact, much of the solar energy now received by the solar panel is unused because the battery is fully charged—there is no need for the excess energy.

With this set of data clearly working, it was time to wiring the A/D converters to the Sun Dragon.

My three ADS1115 boards showed up the other day, and I was excited to get them running. The first thing I did was wire up a module and use the Adafruit_ADS1X15 library to see it function. Using a my bench power supply I could see it doing exactly what I expected. Setting the device in differential mode I hooked it to one of the shunt resisters. Soon I was getting data.

Then the first hurdle. The ADS1115 boards I ordered have the address pin always grounded, meaning all three boards have the same address. I cannot talk to them individually like this. So I had to figure out a way to lift the address pin and manually wire it. My first attempt resulted in the address pin breaking off. Not even 2 days and I've destroyed one piece of new hardware. The second attempt went a little better. So I have two functional devices. They have been wired to each of the shut resisters and I am allowing the setup.

   Body art.

   The reason the fire department showed up for Cinco this year.  A little too much wood to get the fire started.  Only lasted about 5 minutes, but by that time the cavalry was on their way.

   Xiphos.

   Sara.

   House swimming this week.  During the winter we walked to the pool, but now that it is warmer we have been biking.  The first time I did this the ride home was rather slow.  At the end of the swimming night I do two sets of endurance: the first with fins swimming as hard as I can for 10 minutes, and the seconds with arms only for 10 minutes.  The swim with my legs I can do nonstop without too much issue.  The arms only is nowhere near continuous.  When I first started doing it I could do about 2 laps before having to stop and pant for awhile.  Now I can do about 5 laps.  Trying to bicycle after this endurance swimming is almost amusing.  I want to peddle my normal speed, but my body just won't do it.
   Pictured is Pinko de Mustache.  He didn't survive long after this photo was taken.

   I had been growing out facial hair for the past month for this date.  When it arrived, I shaved it into mutton chops.  My initial thoughts were that I looked rather hipster, and my roommates agreed.  So I decided to fit the part.  I found I could fit into my roommate Reven's size 3 pants, and turned myself into a trashy hipster.  Pictured are the results, and I find them scary.

   The Garage Crowd use to hold a large gathering called Cinco de Mustache every year.  After our group stopped holding it at a it's usual venue, recent versions of this event haven't been quite the same.  So we at Elmwood Park thought we'd try hosting it at our place.  Our planning was poor (i.e. all but non-existent) but nonetheless we had a decent turnout and a really good time.
   James and Xiphos built a fire pit in the backyard, and this was the first time we used it.  The initial fire had far too much wood and the first 5 minutes produced a fire about 8 feet tall.  That attracted the fire department and local authorities.  But Elmwood Park is not in the bowels of Wisconsin and policed by power hungry rednecks, and so they were really nice about it.
   We had a piñata filled with candy and mini bottles of liquor, and I got to keep people from hitting it by controlling the rope.  It survived about 5 people before being destroyed.
   The backyard was the center of the gatherings.  We had one tent (with room for plenty more) and we lite the yard from an ellipsoidal reflector light mounted high in our maple tree.

I listened to my first Amiga module file sometime in either 1992 or 1993, and I was hooked to this music format almost instantly. Some of the composers went on to have successful careers in music. Bjørn Lynne (a.k.a Dr. Awesome), Markus Kaarlonen (a.k.a. Captain of Dance Nation), Peter Hajba (a.k.a. Skaven of Future Crew), Jonne Valtonen (Purple Motion of Future Crew) all went on to produce music.

Playing Amiga modules was always a little tricky on a PC. The format was created for an Amiga, which had a build-in hardware that could take 4x8-bit PCM channels over DMA. This offloaded much of the sound processing to hardware. In addition, the format based it's speeds off clocks readily available to the Amiga. To do the same for a PC required raw processing power to pick up for what it lacked in hardware capability. I'm not sure what the minimum requirements were for PC mod players. The Amegia 1000 was a 7.16 MHz 68000-based machine. I was able to play module files without a problem on a 16 MHz 386SX-based machine.

Speed wasn't really the problem with module file playback. The problem was correctly emulating the various effects. There were essentially 32 effects, such as sliding pitch, setting the volume, or jumping to a location in the sample. It is likely the file format was reverse engineered, so the effects did not always do the same thing with different players. I had found some players would not do pitch-shifts or volume slides if the particularity note did not select the associated interments for each beat of the slide. Some players would not play a note if the instrument was not selected in the note, while others used the last instrument selected. And this got worse as the original module format was improved and turned into updated formats such as Scream Tracker (S3M) and Extended Module (XM). Sometimes this had unintended side effects that ended up sounding interesting. Most of the time it meant the song just didn't sound correct. Having always been a PC owner, I often wondered how many songs I heard incorrectly, but only know them in that state.

I've used several players over the years. I didn't let go of Windows 3.11 until the late 1990s because I hated both Windows 95 and Windows 98. The player I used during that time was Mod4Win. This was a pretty decent player but did make a few mistakes—particularly with S3Ms. As of this writing, the website for Mod4Win is still running with a copyright date of 1997. Impressive. Mod4Win was a Windows 3.1 program, but functioned on Windows-based operating systems through Windows XP. I keep a virtual machine running with Mod4Win.

The module player that I used the most was a program called Dual Module Player (DMP) by Otto Chrons. I was having a hard time getting this program to function on modern hardware. DMP is a DOS program, so I thought DOSBox should be able to handle it. However, the version of DMP was trying was version 4.0. This version added support for the DPMI—a method for DOS programs to get access to upper memory when run from Windows. For whatever reason, this wasn't working at all in DOSBox. The reality was, I wasn't using version 4 when I was using DMP. Luckily, I had left some traces of what version of DMP I was using. It took me a bit of searching to track them down, but I used version 2.55 and version 2.82. After a fair bit of searching, I managed to find these versions.

Aside from the nostalgia of having one of my favorite module players again functioning, I am now also able to record MP3s from a few modules for which I could not previous get good recordings because of player error. DMP also had the ability to select “surround sound” panning on specified channels. This was a short delay between right and left which had the effect of making something sound like it was in the background, but an effect I really enjoyed.