Actually I have to bow my head in shame. It’s been more than a month since I posted my original article. What can I say? I just got so snowed under with work that there was no time to do anything. I’m sorry. Take me outside and spank me now!
So here we go. As you may recall, I had three hardware/software bundles to play with:
In fact I was inundated with suggestions for each of these offerings. That’s part of what made my making a decision so difficult. The other part was picking projects that I think are actually “doable” and also the people who I think will actually get around to doing them (I would love to see the final products working).
So, with all of this said, the following are the project suggestions I’ve decided to go with:
Winner of Bundle #1: LCD and Touch Sensing
There were a wide range of suggestions for this bundle, but I decided to go with Buck who wrote as follows:
I have a Pickit2 and I really like it. I started to sketch out an antenna rotor controller for my ham radio hobby. A more or less normal setup would be to have the antenna bolted to a mast which in turn is bolted into a motor called a rotor or rotator. The rotor is commanded to turn clockwise or anticlockwise by a simple switch and relay arrangement. The current position is read out by a multiturn pot and displayed with a simple meter. Say the antenna is currently pointed due east, 90 degrees and we want to point south (180). In the shack, we press the UP button which closes the clockwise relay and sends the antenna gradually higher. When the needle gets near 180, we release the button, the relay opens and the motor stops turning.
That's a normal setup. It's perfectly useful for typical yagi antennas, which need a pointing accuracy of something like 10 degrees or so. I have a microwave station. My beamwidth is on the order of 3 degrees, so I need to point to within one degree. The typical error is greater than that in the pot/needle system, and the motor probably won't stop where I want it to without some fooling around up and down. Too much mechanical slop.
So I have a different system. My dish is bolted to the face of the tower. The tower is bolted to a very high torque prop pitch control motor from an airplane. The prop pitch motor is geared down quite a lot, and has an optical encoder. All I need to do is count the rotations, convert to degrees and Bob's your uncle.
If I had the Microchip LCD and Touch-Sensing bundle, I'd use the daughterboard as my up/down control and would do all of the math and display control with the PIC. It'd be a sweet setup, especially because I could make a 'configuration' mode to set the rotations/degree as well as the current pointing alignment out in the field! Because I use this tower away from home, during VHF contests here in North America...
Winner of Bundle #2: Ethernet
This was a tricky one, because there were suggestions for everything from water irrigation systems to remotely powering up PCs to perform backups to... trust me, the list goes on and on. There were also a surprising number of suggestions for home automation in the form of heating and air conditioning control. The one I selected was the one that seemed to me to have been the most thought out came from Dustin, who wrote as follows:
Problem to be solved: Our new house is unevenly insulated. With a single furnace and a single thermostat in a room that is rarely occupied, the rooms that we spend most of our time in are either too cold or too hot, often at the same time. Overheating/overcooling unoccupied rooms wastes a lot of fuel and money.
Requirements: A device is needed to automatically sense if any room in the house is occupied or not and what the temperature of that room is. A master device needs to determine if heat or AC should be directed to the zone that includes that room. A zone in our house is a block of rooms fed by common ductwork into which I intend to install an electronically controlled vent.
Situation: As I plan to retrofit Ethernet jacks into each commonly occupied room in the house near the areas we sit, that is the perfect place to add the sensors and provides a means of relaying the data via Ethernet. I only want to run 1 line to each room. That line must provide a link for the connected microcontroller/sensors at the same time as serving as a full function Ethernet port.
Proposed Solution: At each room's wall plate will reside a microcontroller with temp sensor, light sensor, and Ethernet interface mostly IEEE 802.3af PoE compliant with data sent on the spare pairs. The master microcontroller decodes the data and decides when to turn on the furnace or AC and which vents to open. The assumption is, if the room light is on, the room is occupied. The exception to this is the bedroom which will also have a reed switch on the door. If the door is closed or the light is on the room is occupied. If the door is closed and the light is off and the time is later than 7PM that condition will trigger night mode. In night mode, the temperature set-point is lower by about 5 degrees F and the bedroom is the primary zone. In night mode the bedroom is the only zone being regulated, unless a new light comes on somewhere in the house, a room is approaching the freeze point, or the time of day is 90 minutes before the wake up alarm time.
Winner of Bundle #3: Projected Capacitive Touch-Screen Technology
You have no idea how hard it is to select amongst all of these projects. Some suggestions come from students who need all the help they can get with regard to obtaining low-cost (free in this case) hardware and software. Others are for projects that are of particular interest to me, including some very innovative musical instruments. But in the end I decided to go with the following project suggestion from Mike:
Hi Max, I've retired after 20+ years of high-end computer design. These days I just want to use my knowledge to help kids; in particular autistic kids who need help just to have a good interaction with their peers because they are non-verbal. I've been looking for cost effective techniques to package an intelligent text-to-voice system in a way that is still "cool" to avoid ridicule ... by their classmates.
Needed: a robust touch/display surface; a "tune-able" voice transducer to match the age/sex of the child; and compute power / non-volatile memory so the child can build his/her vocabulary; associated processing power; and compact enough to be the equivalent of an arm guard in size.
So there you are – I’ll be emailing everyone who sent me project suggestions to let them know if they won (or not). Hopefully we can persuade the winners to write up their projects (with photos) so that we can see how they get on.
Thanks to everyone who sent me suggestions, and I hope you’re not too disappointed if you didn’t win. More as it happens... watch this space...
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.