by Steven K. Roberts
Nomadic Research Labs
Santa Clara, California
January 14, 1996
Hello from the Microship lab, where attention has mercifully turned from moving and unpacking to project-related activities. In this quick update, we begin chipping away on many fronts at once…
First, this was MacWorld week, and spending a day there was top priority. One day, of course, is woefully inadequate to truly assimilate a show of that scale, but we made the rounds in cruise-mode — visiting with our sponsors at Digital Ocean (Groupers and Tarpons), Capilano Computing (Logic Works and Design Works), and Farallon (networking goodies). We also flirted with a few new possibilities, sat in on a Newton programming session in Apple’s Developer Central, bought a 4-port serial switch to save my back, ran into about 10 old friends, and witnessed first hand the health and vitality of the Macintosh community despite the media’s gloom ‘n doom obsession.
Logistically, since we get quite enough city driving on our epic speaking tours, we elected to have a go at public transport. That translated into San Jose light rail down to the Cal Train station, and thence up to the city. As the show drew to a close Wednesday, we decided last-minute to spend the night with our friends in Sausalito, so we hopped a bus across the bay and found ourselves on the….
Sausalito Field Trip
The original motive here was to spend some time with Eric & Sue, who are contemplating buying my Fulmar for adventures in and around Tonga. A remarkable evening of interesting people ensued, as expected, aboard the boat they call home… a tempting lifestyle if ever there was one. Next week, we go back with Fulmar in tow to spend a day on the water.
Eric led us on an overview tour of the Sausalito nautical community, and we spent the next few hours primarily chatting with folks in Marine Recycle (where I bought a slightly used Navico 1800 Tillerpilot for $150) and visiting Bay Riggers. At some point, soon, the Rigging Project is going to have to commence, so we’re making the rounds. The first step, of course, is input from a local marine architect, and we’ve initiated contact with Jim Antrim as suggested by Gino Morrelli and just about every Bay Area multihuller we know… stay tuned for more on this. I’m hoping to be starting non-electronic sea trials this summer.
Firing Up The Tillerpilot
The little Navico tillerpilot I bought this week is not the ultimate solution for the ship’s steering, though it will give us a quickie autopilot for early sails and a hacking platform for the eventual neuropilot (begun last year at UCSD, with the intent of using a neural network to integrate ship motion as part of the steering input — not just compass heading). The simple unit here on the bench has a gimballed flux-gate compass, a small microprocessor, trivial user interface (buttons for left, right, and set/auto), and a motor driving a linear actuator that attaches to the tiller. Basically, you point the boat in the direction you want and tell the tillerpilot to keep her pointing that way — but it has no intelligence with regard to the effects of current, wind angle changes, waypoints, the boat’s actual course, or anything else. Nor does it accept NMEA-0183 inputs from the rest of the nav system. (I’ll have to name our autopilots and windvane Ike, Tina, and Ted…)
Anyway, this one had reportedly been removed from a boat for sale, with the connector simply lopped off (duh). I opened it up on the bench, being unwilling to guess the polarity of a blue/brown wire pair, and puzzled over apparently healthy logic but only short motor impulses until realizing that my bench supply was current-limiting at 400mA. On the 30 amp ferro-resonant monster, it worked fine… with about 150 mA average (operating no-load at a duty cycle reminiscent of Homer on our all-nighter to Catalina), about 45 mA when static, and 1-3 A when pushing or pulling the actuator, depending on load.
This now goes on the shelf for use in test sails, later to be hacked as an “end effector” for neuropilot development if the 36kg thrust and 12.5-sec lock-to-lock response time of this low-end unit prove adequate for the ship.
Control Network Redux
A good many of these status reports last year centered around the development of the control network of New Micros FORTH 68HC11 boards, dealing with tasks ranging from crosspoint switching to power management. The first critical box on the Control System PERT chart, from which everything else dangles, is to get the network back up to the level it was before we shut down the UCSD lab. This sounds easy.
Actually, the hub came to life immediately, and I even added another sensor to make it do more than just blink its green health LED. A simple task displays textual annunciation of the opening of the lab door and the breaking of an IR beam — these inputs will be part of the test configuration of the HyperTalk “Watch” view that will run on all connected Macs.
But for some reason, communication with other nodes is flaky. I finally got the audio and video crossbar systems running their multitaskers and controlling their switch matrices tonight, but it took numerous retries. The scope revealed some pretty bad line reflections, so I followed RS-422 implementation advice and added long-overdue terminating resistors to clean it up. The transition overshoots disappeared, but nodes still tend to lock up during file transmissions. So although a few blinkenlights are enlivening the midnight blackness of the lab, I can’t quite color in that critical task just yet… <sigh>
HF Automatic Tuner Order
We’ll have more on this when the hardware actually arrives, but I just ordered the kit of parts corresponding to the automatic HF antenna tuner designed by Dwayne Kincaid , WD8OYG, as described in the January issue of QST. For $150 and a bit of assembly, we should have the functionality of boxes selling for $600-1500 — magically matching an unbalanced antenna (like a backstay sloper) to the current transmit frequency. There’s a 68HC11 that watches the SWR, then switches some combination of eight series inductors and eight parallel capacitors into an L network (131,072 possible combinations) until reflected energy is minimized. Nifty. I’ll keep you posted…
Technomadic Home Base?
Finally, I’ve just emerged from a meeting with a woman who may become not only our new base office manager but also our partner in the business of offering a whole suite of services to technomads. I’ve wanted to do this for years… I’m not the only one who needs competent management, a level of indirection in all business dealings, datacomm paths, and active assistance in speaking, publishing, and development projects. More on this as it unfolds!
Back to it…