Croatia

Wow - I'm way behind in my blog postings. I'd better start catching up now or I'm never going to get it done. I guess I'll start with my vacation in Croatia.

I flew to Dubrovnik by way of LA, Frankfurt, and Zagreb on June 18th. Arrived at the apartment (just across the street from the right-most tower in the picture) a little before midnight on the 19th. My friends Noelle and Rachel arrived separately on the 20th, from Iraq and Florida respectively.

The old city is absolutely beautiful. We spent a day visiting all of the local museums, another island hopping in the Elaphiti Islands, took a day trip to Mostar in Bosnia, went horseback riding, attended a very touristy 'authentic village dinner' party, went on a scuba trip, and after Noelle and Rachel left on the 27th, I made a day trip by myself into Montenegro to see Kotor and Budva. The Bay of Kotor is gorgeous, but Budva is way over-touristed.

A couple of chance occurrences on the trip happend to contribute to my decision to leave my day job when the vacation was over. First, the guy seated across from me at the dinner party in the village turned out to be the former head of research for a major water utility in the UK - and had experimented with some of the same approaches we were working on in our project, and abandoned the effort because of the same problems that have always made me doubtful about our prospects for success. And over a bottle of Havana Club back in the apartment, Noelle described her personal experience with the Army system we were in the process of writing a proposal for. Turns out the system was an even bigger boondoggle than I'd guessed. And something she mentioned about another, entirely unrelated program, turned out to be directly relevant to what the company later tried to convince me to stay on to help with.

I had a 7-hour layover in Zagreb on the way back, and a local OpenTracker user picked me up at the airport and took me out for a beer. I found out a little more about commercial use of my trackers in the country, and maybe even got myself a legitimate business expense deduction for part of the trip.

I'll post pictures to my Flickr account when I have time, but I think that's enough for tonight.

More Machining Fun

Here's the scarab I mentioned before - I scanned the original with the MDX-20 and cut a new one out of machinable wax. It was finished with a 1/16" ball-nose end mill, which did a surprisingly good job of reproducing small details. It's also a lot more resistant to breaking than the smaller, more expensive tools I'm learning to use. So far I've broken two 1/64" end mills, both in their first 60 seconds of operation.

Here's another experiment. It's a model of Mount St. Helens, based on USGS digital elevation models. I'm sure there must be an easier way to generate stereolithography files from DEMs, but for this test I had to make a VRML file and convert it to STL format using an evaluation copy of Rhino.

I still haven't had much opportunity to mill PCBs with the machine, but I'm still finding other useful things to do with it. Today I fabricated a replacement lens for my Maglite out of Lexan. I figure that saved me at least $2. Or rather, it would have if the original lens had actually been broken. But at least I've got a spare, just in case.

Water, Water, Everywhere

I suppose it can't quite compare with the flooding in the Midwest right now (we were in Des Moines for the 1993 installment, and I hear it's worse this time) but we had quite a mess here over the weekend. The hot water hose on the washing machine burst on Friday afternoon, and fortunately I was home at the time, but I was working upstairs and didn't discover it for half an hour or so.

By that time the water was about two inches deep in the laundry room and had run all the way through the office, under the door, and down the driveway. I got the water shut off and started a siphon with the garden hose. Once that was going I grabbed the wet/dry vac from the garage and started slurping up five gallons at a time and dumping it outside. Took two days for my back and leg muscles to recover from that, but it got most of the water out.

Tory made it home about then and started making phone calls while I continued damage abatement. It took the water damage cleanup people three or four hours to show up, and we spent that time continuing to haul stuff out of the office and storage area while Charlie worked the vacuum, extracting as much water as he could from the carpet.

For the next five days that part of the house was sealed off, full of blowers and a giant dehumidifier. The temperature got up to about 105 F in there, so I didn't get a lot of hardware work done over the weekend.

In the end, the damage wasn't too bad - maybe a hundred bucks worth of packing materials ruined, some cables destroyed, and a bunch of camping and emergency gear made very soggy.

The lost time is the biggest inconvenience, I suppose. That was supposed to be my last free weekend this month - this weekend we're going to Magic Mountain, and I'll be in Dubrovnik for the two weekends after that.

It's over!

Hamvention 2008 is done! Dayton is my most stressful annual trip by a wide margin. This year went fairly smoothly, and I think the display was the best we've had yet. Jason did a great job putting together the nüvi / T2-135 display and it drew a decent size crowd.

I was definitely more busy this year than last - I got out of the booth a couple of times on Friday and managed to see at least some of the other exhibits, at least in passing, and even had time to eat a cold hot dog over the course of about an hour. Saturday was a little more hectic, and I don't recall even getting a bathroom break for about eight hours, and no food at all. Can't say I really noticed until it was over, though.

Financially I think it was probably a wash, as usual. Between booth space, airfare, hotel, shipping, and so forth it's hard to turn much of a profit at Dayton, but it's worth doing for the exposure.

I'm still about a hundred emails behind thanks to the trip, and I've got a ton of other stuff to get done this weekend. I did finally get a chance to try making a PCB with the MDX-20 today, though. And broke my first tool in the process - not one of the $6 ones, either. It was my sole $48 1/64 inch end mill, of course. I had it cutting just a bit too deep and a bit too fast, and it lasted about 5 or 10 seconds. Did a nice job right up until it snapped, though!

Somehow I wound up with a bunch of 0.062" Garolite G-10/FR4. It's the same stuff the PCB substrate is made of, but colored. I cut some shapes out of one of the black sheets and they look cool, but they seem to be very prone to scratching.

I'll have to pick up some more polycarbonate and play with that. I've only got a 1" thick slab, which is a bit more than I want to start with.

I got the scanner option running the other day, too. We put Charlie's little stone scarab carving in it and generated a model with 0.008" resolution. Took a few hours. I then had it cut a new scarab out of machinable wax. It came out pretty good, but I'm still learning how to properly set up roughing and finishing processes.

I even managed to make a useful production item with the machine. It took me a few tries, but I got it to do the required cutouts in the ABS end panels for the ADS-SR1 repeater. I have a couple hundred aluminum panels already fabricated and paid for, but I should be able to do the next batch in-house. It's amazing how much more mess a small amount of ABS makes, compared to the machinable wax. I think it must be a result of its electrostatic properties. It coats every surface in the machine, while the wax just piles up nicely to be vacuumed away.

New toy

I'm totally swamped with Hamvention preparations, but my new Roland MDX-20 CNC milling mchine came in today and I had to try it out. It's been working great, at least since I figured out that one of my serial ports was bad and was causing the machine to try to throw chunks of wood at me.

Right now it's running through one of its sample files - a banana, which it's cutting from machinable wax.

Anyway, back to flyers and price lists and demo hardware!

Mila Rainof, 1980 - 2008

No project updates this week, just a farewell to a friend. I met Mila through Tory - they were in the same draw group at Stanford and had been very close ever since. I didn't know Mila nearly as well, but she was one of those people whom it's almost impossible not to like, and news of her death came as a shock.

Mila was finishing her last year at Yale Medical, and was set to start her residency as an emergency room physician back here in California. She was struck by a car and killed while crossing the street on the way to her apartment last week.

My last email conversation with Mila was about trauma cases we'd both seen, oddly enough. She was on anesthesiology rotation at the time, which she said mostly involved jealously watching other people sleep.

What a horrendously unfair end for a wonderful, bright young woman who should have had a long and productive medical career ahead of her. It's been a week and still the refrain from Depeche Mode's Blasphemous Rumours runs through my head when I think about it. A sick sense of humor, indeed.

You'll be missed, Mila. The world is without a doubt a worse place without you in it.

Making Progress

The simplex repeater project has been coming along more slowly than I'd hoped. That's been partly due to the difficulty in getting the software-based DTMF decoder working, but I think it's good enough at this point for a beta release at least. But mostly, everything else imaginable keeps getting in the way.

I've got what must be my tenth cold of the season. Charlie brought this one home. Fortunately it's not a bad one, and I haven't been totally out of commission.

I started buying parts for the first production run last Friday. Ran into two problems - first, the flash ICs are being phased out in favor of a new model. It's not the end of the world, since I made sure there were other parts out there with the same pinout that'd work. Trouble is, the new chip doesn't have one particular feature I'm using, and coding around it would take a bit of work. Adapting the code to an entirely different vendor's chip would take a significantly larger amount of work, and it'd probably mean a major rewrite of the SSTV encoder, which uses the same chip. Fortunately I found a good source for the old part, so I've got 500 on the way, due in on Friday.

The next problem was the switch. I designed around a model that I had several examples of on hand, and I was careful to make sure it was available from at least two other vendors. Only I didn't think to check on availability of the caps. Who would have thought that sourcing a stupid plastic cap of the appropriate size would be such an ordeal? I spent two or three hours hunting online and calling vendors. The closest thing I've found is too big to fit in the hole, and I'd rather not throw out a couple hundred custom-made end plates. I think I can modify the holes to widen them by about 1 mm without ruining the finish, though. Might have to move the switch back a bit on the next PCB revision to accommodate the longer cap, but there's tons of space for that.

No major hardware glitches with the PCB so far. The footprint for the flash chips is wrong, but it's close enough to work with a little care. I routed the external squelch input to an I/O pin with no interrupt function, which pretty much kills my power save scheme when operating in that mode, but if you've got an external squelch then you're probably using a mobile rig and have external power anyway. Besides, if you've got any announcement or ID timers enabled it's not going to be in deep power save mode anyway. Worst case, it'll mean 30 days of idle battery life instead of six months.

The prototype pictured above has a laser-printed paper label on it. It was my first attempt, but I think it'll do. I don't want to clutter it up with more text and graphics than necessary. I'll add a page to the manual with a quick reference guide on a strip to cut out and fold up to fit in the battery compartment.

The 10-mil Lexan graphic overlay on the production version will cost nearly as much as the case itself, at least to start with. Still, I think it'll be worth doing, and I'd like to get experience working with the material.

Oh, and after all of the component troubles, I was finally ready to get some work done on Sunday, but there was a fire in the lot of the packing company down the street from my office. It took out power to a good portion of the town, along with cable TV and Internet. Fired up the generator for a bit, but didn't get a lot accomplished.

Charlie's off at Outdoor School this week. I remember my own week there, two decades ago. I hope he has an easier time on the 9-mile ridge hike; the wind must have been blowing 50 miles per hour at the top, and was driving rain and hail on the day I made the hike. I think it was also uphill both ways in those days.