The usual control mechanism for a whole-house fan is an on-off timer and a high-low switch. You set the speed to high or low, then twist the spring-driven timer to set the length of time the fan should run. A certain amount of guesswork is involved in setting things before you go to bed. Err to one side and you wake up with the house warmer than you would like: you ran the fan on low, but it should have been on high. Err on the other and it can be a tad on the brisk side: a cold front went through and it's 56. A year-and-a-half or so ago, my wife decided that she wanted more control over how the fan behaved. It should be capable of switching speeds, shutting off if things got too cold, etc. As no such controller appeared to exist commercially, I was charged with building one.
First, I suppose, is that it works. No more 56-degree mornings. And in some modes, the software decides how long to run on high before shifting to low while still having a good probability of reaching the desired temperature. Certainly I have a much better idea when I go to bed of what the temperature in the house will be in the morning than I did before.
Second is that the device is characteristic of the sea change that has happened with control applications over the last 15-20 years. The device consists of a single-chip microcontroller, some sensors, and a few actuators. Everything else is software. While that provides a great deal of flexibility, it also means that the whole approach is dependent on the continued availability of large-scale integrated circuits. In the near term, that's not a problem. In the longer term, with constraints on energy availability, limits on the scope of trade, preserving the production capability may be more difficult. Certainly if you believe in anything approaching the Mad Max scenarios, integrated circuit fabs will disappear quickly. Are we building ourselves into a corner?
Third, and this is mostly whining, concerns the fact that it looks... tacky. The problem of mounting the various parts in a good-looking enclosure is, in fact, one of the most challenging parts of hobbyist projects like this one. What I would have liked to buy was the LCD panel and touch screen all mounted in a nice plastic box. One where everything was properly centered and aligned, and with enough room behind the display for the small custom printed circuit board. Granted that this type of hobbyist market is small — how many times was I told that I shouldn't be building a box to mount on the wall, I should be writing a smart-phone app to control the fan? — but it seems like it would be big enough to support such a widget.