My Motley Meanderings

I was hoping the weather would be fine this weekend because I had a long list of maintenance jobs I wanted to do on Fenchurch. That is why I've spent much of today and yesterday working outside in close to zero temperatures with a strong windchill factor and a constant light drizzle. I did manage to do most of the stuff on my list before it started trying to snow and I came back inside. It's unfortunate that the garage under my office that I did the respray in now contains a large trailer. The garage we have next to the house is just big enough to park the car in and squeeze around one side (if I clear a load of junk out first); it's too cramped to want to use it for jobs that involve taking all the wheels off.

When I wasn't getting cold and wet outside I've been doing a lot of reading about Nixie tubes and neon indicator lamps. I'd heard of Nixie tubes and even have an old piece of test equipment with a few in it, but they were obsolete before I was born so I've never worked with them. Recently there has been a resurgence in their popularity, usually in the form of digital clocks, as people come to realise that they actually look a lot prettier than the seven segment LED displays that replaced them. You can still buy them as New Old Stock even though the last factories that made them closed down shortly after the fall of the Soviet Union in 1991.

Neon lamps are still commonly used as a power indication light in mains-powered equipment such as multi-way socket adapters. I didn't know that Nixie tubes are basically glorified neon lamps, and I had no idea how versatile the lamps are. They were even used to build digital logic gates in the days before semiconductors took over. Don't believe me? Take a look at this!

You might be wondering why I've suddenly developed an interest in obsolete optoelectronics. The reason is that I've decided to use Nixie tubes for the score display and an array of 200 cheap neon lamps as the main display of my retro-futuristic Tetris machine. I originally wanted to build a mechanical display but it proved impractical and would have cost far too much for the parts.

Next weekend is starting to look a bit busy. I'm going to the Morris Minor Owners' Club AGM in Derby on Saturday, a vintage slot machine fair in Coventry on Sunday, and at some point I need to pick up a heater for Fenchurch from a bloke who lives near Birmingham. I'm thinking of trying to stay in the Stratford upon Avon YH on Saturday, and if I do that it might be worth also trying to see a play at the Royal Shakespeare Company's Courtyard Theatre (they'll be performing Henry V).

Inspired by Pac Gentleman and Pongmechanik and of course The Great Crystal Cyberdrome Exhibition, I just spent an hour in the bath trying to devise an electromechanical computer that could drive a Tetris fairground game. Of course I could cheat and hide a microcontroller in the base (I've already implemented Tetris in C), but there's something terribly appealing to me about the idea of a clicking whirring machine that could theoretically have been built as far back as 1835 (when the relay was invented).

I've just been reading about the Parallax Propeller 8-core microcontroller. I was never a big fan of their BASIC Stamp, but I have to say the Propeller sounds pretty cool, in an eccentric sort of way. The thing has 32K of shared RAM, 2K of local RAM per core, a 32K ROM containing an interpreter for a high-level language called SPIN, and multiple video generators (PAL, NTSC, or VGA). It doesn't have interrupts (the propeller philosophy is to dedicate a core to polling each input device), internal non-volatile memory (the monitor in ROM supports booting from an external i2c EEPROM), or much in the way of peripherals other than the video generators (want a UART or a PWM generator? program one of the cores to emulate one).

One big reservation I have about the Propeller is that it's so different to anything else available in the embedded systems market that if you developed an application for it you'd be locked into the platform with no practical way of porting it to another manufacturer's processors short of starting again from scratch. I also have some doubts about how fast it will be given the round-robin multiplexed fashion in which its cores access main memory. I suspect it will turn out to be excellent for rapid prototyping, one-offs, and hobbyists, but it's unlikely to make a significant dent in ARM, 8051, PIC, etc. sales for medium to high volume applications.

Playing with electrickery. Designing electronic gadgets, to be specific. It started off as an idea to make a simple speed controller (SSC) for Beyond Cyberdrome, but after I got it working I got a bit carried away and have now also developed a dual radio controlled switch (DRCS) and a single radio controlled switch (SRCS), and am considering trying to sell all three designs as self-assembly kits (thereby hopefully dodging some of the burdensome legislation that applies to electronics manufacturers). This is far from the first time I've had ideas along these lines, but I usually don't get as far as a working prototype. Pictures after the cut...
( Read more... )