I’m trying to get graphics handled nicely (for at least Safari and Firefox), and it appears that web servers are behind the times.

Plan A, was to put up PDF files, which display great in Safari, but not at all in Firefox.

Plan B was to put up SVG files:

gs -sDEVICE=svg -dNOPAUSE -sOutputFile=`basename $1 .pdf`.svg $1 < /dev/null

but they were really big.

Plan C was to put up compressed GZ files (.svgz), but despite claims all over the web that this was the way to do it, it didn’t work, either in Firefox, or in Safari. However, Apache configuration is notoriously arcane and horrible.

The web server that handles these requests:

http://homepage.mac.com/dr2chase/WP/SVGTest.html

http://homepage.mac.com/dr2chase/WP/HV_ctlr_6_crystal_imon_filled.svgz

needs to include the Apache equivalent of

    AddEncoding x-gzip .svgz

I have tested this fix on my own laptop, running MacPorts Apache2,
serving files from my local disk. Failing to include that directive
yields either (for the direct reference to the svgz file)

This page contains the following errors:

error on line 1 at column 1: Encoding error
Below is a rendering of the page up to the first error.

or little broken-graphic boxes if they are embedded.

The MIME type IS being set correctly, but the compression is not.

I’m pretty happy with my bike light, but gosh darn it, it could be better!

Various features that I want, but demonstrably don’t need:

• Able to tolerate higher supply voltages.

  The Buckpuck tops out at 32 volts, and the desired voltage for best efficiency is probably closer to 2 times mph — 40v at 20, 50v at 25, and 70 at 35 (which I have surpassed on some hills). Obviously, eventually I am more concerned about the survival of the electronics than I am about perfect efficiency, but I’d like to be moving faster than 25mph before that matters.

  Furthermore, at moderate speeds, voltages above the peak efficiency level allow for lower drag at reduced power. This might be good for daytime running lights, for example.

  There’s at least two ways to get this; one is to disable the voltage doubler once an adequate voltage has been reached, and the other is to choose higher-rated parts.

• Standlight.

  A real one — everything needs to be on at least half power, meaning 2.5 watts. In terms of energy storage, 150 joules per minute, so an ultracapacitor is not a solution. A capacitor holds 1/2 C V2 joules, so a 1-farad ultracapacitor run down to 10 volts (50J) must start out at 20V (200J) — such devices seem not to exist, and certainly not at a sensible price.

• Charge-back into the battery

  Given that a standlight needs a battery, it would be nice to run the battery as long as possible, meaning to put some charge back into it.
  Getting a battery to 100% charge is not a job for amateurs, especially on a bicycle (varying temperatures, varying voltages), but perhaps we can do an approximate job.

• USB charger

  A USB charger needs to supply 500milliAmps at 5volts — 2.5 watts. This is possible either if the lights are low or at good speed. So why not, it would be fun.

• Daytime Running Lights features

  These include, mostly, the ability to blink and do other things to call attention to the bicycle. Reduced power, compatible with USB and/or battery charging, is also a good thing.

• Semi-open, so other crazy people can do more than my already ridiculous set of plans. So, attempt to follow standards set by the Arduino guys, make provision for a crystal good enough to run USB, be sure that one of the connectors could support not just USB charging, but USB communication. It’s not an exact match for the Arduino pinouts, because things just didn’t fit, and because it was important to get the two AC lines to the comparator inputs.

It’s still at the do-I-dare-build-this stage. Not only is it a lot of parts, there’s some surface-mount bits, and SOFTWARE. Said software has the option of setting some (non-essential!) parts on fire, and as a control system, there’s ample room for funny oscillations. However, the big-ass capacitor (4.7 milliFarads) takes almost 5 milliseconds to charge by 1 volt, given a 1 amp input, so there is some wiggle room.

Microprocessor inputs:

• Two ac inputs, to measure wheel rotation
• System voltage
• Rectified AC current
• Four analog lines, for controls
• Four digital lines, for controls/USB
• Programming header, and its lines

Microprocessor outputs:

• Three enable/disable lines to LED power supply, can be PWM’d
• One voltage-doubler disable line (ON by default, so that system powers up fast at low speeds)
• One battery-enable, for charging/discharging the battery. Software has to know what the charging voltage is; current to battery can be estimated somewhat by strobing the lights and battery and monitoring the change in the system voltage; as noted above, a goal of 100% charge is not a good plan.

Behavior, roughly:

• Off, is off. Need to roll the bike to turn it on.
• Slow. Use batteries to keep the lights on. Boost on.
• Medium. Batteries off, boost on.
• Daytime running lights. Lights on steady-dim, pulsing bright.
• Hi/low beams. Of the three LED power supplies, one is high beam, one is low beam, one is rear. Normally the lights are high or low beam.
• USB charging. USB charging will put a load on the supply, which should cause some dimming in order to stay within the power limits.
• Fast. Riding fast produces excess power, which means that it must be dumped somewhere. So, both high and low beams will be activated, as much as necessary to cap the voltage. (Need to put a zener clamp on this somewhere…). One possibility is to dump power into the batteries; for plausible descents, for plausible amounts of battery, the excess energy would “fit” in the batteries, and could then be deliberately run down by powering the lights on full after the descent for a corresponding amount of time.

The weird slots in the PCB planes are intended (not that I know this will work) to provide some noise isolation for the microprocessor and in particular, for its analog side. The switching supplies for the LEDs each put out at least 350mA, switch in the 100s of kilohertz, and could be a tad noisy. The slots under the inductor on the right are intended to reduce eddy currents; the inductor wires are perpendicular to the cuts.

(The PDF images will display in some browsers, or can be clicked for download in others. I need to make them be SVG. Aha, trying to get that done right, that is a problem — a problem caused by misconfigured web servers).

“Don’t worry as its completely SAFE…”
Brilliant, to pretend to be Adobe intruding into your browsing to update Flash (with no close button, of course), because that’s just the sort of abuse I’ve learned to expect from them. So maybe it’s Adobe, maybe not, who can tell?

Playing with my bike, playing with my camera.

and the importance of obeying it.Is it really necessary to point out the obvious — that people pick and choose the laws they wish to obey, and come up with all sorts of reasons why it’s ok that they break that particular law — but other laws, those are important.

I’m not arguing that traffic laws should be obeyed to the letter, or breaking them is unsafe — rather, I would like to point out that we already break traffic laws when we think it is safe — and bicycles, who are different from cars (slower, nimbler, better able to see and hear) might be able to do some things safely that cars cannot.

11:34:09

11:34:14

11:34:25

11:35:26

11:36:04

11:37:49

But perhaps stop signs are different. Or maybe they aren’t.

I fix my (car) brakes, and I fix my (car) stereo. The rest, oh well.
The new speaker did not fit in the old speaker mount, despite being the same “size”. Marine plywood to the rescue. This is the second one I’ve made like this, and it went unexpectedly well. A dremel is wonderful — and I just set the saber saw at 45 degrees, and free-handed a marked circle, and it came out nice. But, again, marine plywood wins big — strong, no cutting, no splitting. I don’t think I could have cut low-grade plywood or regular wood that close without cracking it in the process.

When you give a pumpkin field a year off, what do you think grows in it?

Free Pumpkins!

I like my bike, but these are just wonderful. I don’t think I can have all of them.
The Frances Small Haul. Cable steering, a canvas cargo area (got to be good for the wind resistance), pretty-pretty-pretty.

And the Chaise Cruizer. Because, why not?

Turns out, you do it with Quality of Service. Infinitely low, that’s the answer.

And don’t forget to click “Save” at the bottom of the page.

The chain case serves two purposes; keep my pants clean, and keep the chain “clean”, such as it is.
First, I added what I guess I have to call “pants ramps” or “cuff ramps”. My pants don’t usually catch on the mounting brackets, but last time they did, it was quite a surprise. I was out of the saddle and pushing hard and fast, and OOPS! Turns out they make Sculpey (bakeable plastic modelling clay) in “silver”, and I also had some silver paint. So, I put an 8mm socket on the mounting nuts, and mushed and carved till I liked it. With the soft clay sticking to the chain case, I took the whole thing off the bike and put it in the oven (275 F for 30 minutes). It cooled fast afterwards. I used super glue to fasten the hardened ramps to case, and then put a little silver paint on them to get them closer to the aluminum.

Second, I took a piece of plastic sheet from the hobby shop, cut it to fit the inside of the chain case, marked holes for the mounting pegs, and then cut holes for the pegs, bottom bracket, and the PVC mount. The plan is that this will keep most of the muddy tire splash off the chain.

In the good old days, I’d be nine years dead already, if I were lucky.
Most of my ancestors on my father’s side are buried in New England. Until the Pomeroy sisters showed up, life (for the men, at least) was short.

Pre-Pomeroy:

Isaac Newton Watts (no kidding) and his great-great-great-grandson Isaac.

The most hated math professor ever at Dartmouth College, and his great-great-grandson John.

My great-great-grandfather, made it all the way to 49.

Other uncles and cousins.

Then, Mary (Mary Fuller Chase, above) and Emma Pomeroy married into the Chase and Lord families, and things got a tad better. I knew all these people when I was a kid, they were in their 80s and 90s then, and still somewhat up and around (my great-grandfather was still working on his orange grove at age 88 or 89).

His brother. I ought to wander down to Milton, just for the sake of completeness.

And their cousin, Emma’s daughter, Laura Woolsey Lord Scales.

Terrafugia brought their car to the Unconference at Sun Thursday. I think I like the idea of the flying car, more than the actual thing; or rather, for that money, I like other things more. But still, a flying car!
Click the pictures for full size — you know you want to.

1300 pounds of carbon fiber, 100 horsepower of aircraft motor (but runs on auto fuel).

The weird triangles on stalks on the elevator are a “fix” on the prototype to improve the handling. (I asked.)

Nothing says “safety” like an explosively deployed rocket parachute.

And the wings really do fold automatically, it was fun to watch.

These bars are Too High. These bars are Too Low. These bars are Too Wide…
I’ve been tinkering with handlebars for years now. The current setup works better than anything I’ve had previously, but scores way high on the dork scale:

The main bar is a Velo-Orange Monmartre bar. I like skinny bars (I’ve tried wide bars, don’t like them) and these are skinny. There’s an upright section for riding upright, but on my old (dropped, 45cm V-O Nitto Randonneur, see banner at top of blog) bars I also spent a fair amount of time on the hoods, so I carried those over. Turns out sometimes I do like to be bent over (especially on a windy day) and the speed picks right up, so with these I can grab the center and rest my elbows on the grips. Maybe I should get some tri-bars, and complete the Franken-bar look.

The vestigial mountain bars below are so I can keep the HANDlebars clean, and so I can fit the SRAM shifter to something.

The only think I don’t like is the weird inward tilt of the brake levers. So this may not be the end; one possibility would be get some narrow mountain-gauge upright bars, put bar-ends where the hoods are, and deploy interrupter levers strategically.

That funny little switch on the left end of the baby bars? That’s my high-beam/low-beam switch.

It was tedious and fiddly to get it on right, but I am happy now.
The rest of this post contains some large photos.

If you go looking for instructions online, you won’t find much. The product page says it “works with modern Sugino Cranks”, and it does, but just barely. My crankset is a Sugino RD2, with a 38 tooth chainring, and a track (103mm?) bottom bracket.

First, the results:

The chaincase mounts on two pegs that strap to the seat and down tubes, and a sort of U-clamp that grabs the chainstay, if there is one.

My plan was to put the nuts on the outside; that might not be quite as pretty, but it was clear that this was going to need fiddling and adjusting, and I have a new chain and chainring in the mail (the old ones are badly worn), so all in all this seemed like the best plan.

The main problem is that the clearance between crankset and chaincase is very small, perhaps a millimeter at best. I spent about a day dry-fitting, not sure if it would work. I had to do the final fitting with the bike laid over sideways and propped up, so I could rotate the pedals freely.

A Big Dummy presents two additional problems, one minor, one tricky. The minor problem is that the down tube is really large, and the supplied mounting strap simply does not fit. Fortunately, I had some stainless steel of the appropriate gauge hanging around, so I cut myself a new strap.

The tricky problem is that there is no chain stay, so the clamp provided with the chain case does not work. The tube has a weird oval section, and is large, so it’s tricky. I started with a trick that my brother told me, which is to make the clamp you need from Fimo or Sculpey:

This worked well enough, but I decided that I could use the Sculpey as a guide to carving (with a Dremel) a piece of PVC pipe:

It’s necessary to drill holes in the chain guard to mount it, so I got as close as I could with the dry fit, and marked marked the chain case with pencil.

Once it was rough-mounted, I used an allen-wrench in the gap between the crank web and the crank arm to be sure it was centered, and tinkered with the mounts to get it right.

Because I wasn’t using the chain stay clamp, I had two spare nuts, so I doubled up on the mounting clamps, mostly because a nut is a smoother and less snaggy surface than a threaded end. I could also Dremel the excess threads off, but that would make it harder to work on in the future, so I decided to start like this and see how things went.

You’ll notice that the rod (below) looks like it is not perpendicular to the centerline of the bicycle; this is intentional, for whatever reason (skinny track BB, probably) if it is perpindicular, it is too long.

I put two holes in the rear part of the chain case above and below the top edge of the PVC; the idea was to pull it tight with zip ties so it could not rotate down. I pulled it tight, checked the clearances, and it was too thick, removed a few millimeters with a power sander, until it was just right.

And it works. I took it for a ride of a few miles, bumps and all, in jeans with no pants clip, and they did not snag (I was a little worried about the mounts) and it did not rub.

My first youtube video. I’d been meaning to do this for some time, the stop sign at this intersection seems to be often ignored. 60 seconds, six cars, no stops.

So if you need a rebuttal to one of those scofflaw-cyclist-ranters that are so common, here you go.

Friday morning I installed Tomato 1.25 on a Linksys WRT54G 2.2 router, and it went swimmingly.
Except for changes to the admin/root account info and dynamics DNS, it picked up all the settings from the previous (Linksys) firmware.

Other than tweaking the power slightly upwards, I have not looked into the bells and whistles much, except to study our bandwidth use. If it turns out that we’re not using, or not getting, the peak bandwidth promised by our cable company, I’ll have a look at DSL. I’ve checked performance at DSLreports and it looks good there, but I would be completely unsurprised if they noticed test connections to that site, and gave them priority (I have, in fact, worked on benchmarking in the past, and know that cheating is industry-standing behavior).

Well, gosh darn, I think I have to take back some of the bad things I’ve said about Comcast. To test bandwidth, I fired up Vuze (= Azureus = BitTorrent) and pointed it at the South-by-SouthWest 2009 showcasing artists, gigabyte 1 and gigabyte 2. Uploads were running at about 2Mbit/s, which I think is a lot more than is advertised. Downloads were something like 3 times that. The timing for the bandwidth was a little wonky; it went like gangbusters for a while, then it backed off (why?) but when I tried it experimentally just now, vroom, 2Mb/s, no problem.

And the mighty-fine tomato router software lets me look at this. And it also makes it pretty obvious why ISPs might be a little grumpy about P2P; the P2P parts of the graph just dwarf all the rest.

Further update: hard to tell, after two weeks, we had a spell of really wonky behavior. Despite my best attempts to debug it, cannot tell if it was the router or the ISP. Similar problems (and my reply) reported here.

It was supposed to rain, so I was driving to work, saw this out in someone’s trash, and it followed me home. $45 in parts from the Broadway Bicycle School, plus new bearings for the headset, and it’s a bike again.

I have no idea what is going on with the paint job.

I took it to the B’way Bike School on the Big Dummy, which turned out to be a good idea because that way I didn’t need to mess with parking; things were slow enough, they were happy to let me wheel the whole thing in, and then we took inventory of all the missing bits.

The headset was the worst I have ever seen, so I had to rebuild it. The top bearings were full of some crusty stuff, so I cleaned it all out with turpentine. I washed the bearings in turpentine, and when, for grins, I put a magnet on the bottom of the tin, look what stuck:

No, not The Blob. Magnetic dirt, that used to be ball bearings (and races, no doubt). I bought new BBs, from the nice guys at Wheelworks down the street.

There’s a trick for getting old rusted threaded stems loose without destroying the bolt; use a cold pack to cool the stem, and it contracts a hair, and then you can move the bolt. Greased up seven ways to Sunday on reassembly, of course.

(No, it did not get that much shorter, that is after loosening the bolt, on an older bike.)

Saw this just the other day driving home in traffic, three cheers for telephoto and anti-shake.

But what does it mean?

Changed jobs twice over the years while the family had a “pre-existing condition”, i.e., pregnancy. COBRA, which incorporated Kennedy’s “fairly small step” to extend health insurance coverage, saved our butt, both times.

It’s expensive, but it beats the alternative. Also, anyone doing this, be aware of the following details:

• You have to keep the old insurance running till the OB/GYN gives the ok at a follow-up appointment;
• The new baby is NOT a pre-existing condition, and emerges, covered by whatever new insurance you might have, poof!
• The hospital may get the billing wrong on this last detail, which can result in a telephone call after you arrive home with the baby, claiming that (s)he is not on the insurance. Tell them to bill it to the new insurance.

Thanks again, Senator Kennedy.

Front fork was aging fast on the old bike, my brother said his Big Dummy was great, I didn’t get laid off this spring, and the stock is up on solid acquisition rumors. So I bought myself a Big Dummy.
Or rather, I bought a frame and some parts, and reused some from the old bike. The tires, front wheel, handlebars, seat, seatpost, seat lanyard, lights, fender, water bottle cage, front rack, wide loader, and snap deck are from the old bike. Rear wheel, shifters, pedals, botom bracket, crank, chain, brakes, and stem(s) are new. The “chain tensioner” is recycled from another bike. KickBack, I bought from my brother, he notices the weight more than I do. I like it.

Here it is loaded:

Here is the (revised) cockpit. Since the new bike has a threadless fork, I decided to mount all the crap on a second stem. From left to right, bell, lights (hi/lo beams), odometer, coffee, shifter. The circuit-board with blue tape thing is on the rack; it is the power controller for the LEDs, and needs to find a better home eventually.

And yes, that is a fancy headset. The guys down at the LBS did a great job on the hard sell, and the clincher was the multi-year guarantee; given that longtails put double the weight on the front fork, this seemed like an excellent idea. Plus, it’s very pretty.

Chain tensioner. I decided to go with internal gears, eventually I am getting a chain guard or partial chain case (I would love to keep the chain clean in the winer), but the Big Dummy has vertical dropouts so the chain might not always fit right. Plus, with the long chain run, it doesn’t hurt to be sure it stays aligned going onto the rear cog. Using a single chainring and sprocket also meant that I could use a track bottom bracket and crank, and get my feet a little closer together (about 20mm closer together), which I had noticed, riding on a 3-speed, felt better.

Here is most of it, unloaded:

No problems with the internal gears yet; I’m monitoring the adjustment in the beginning (it’s EASY), and once I got the second stem located right, shifting was convenient, and I am not noticing any particular lack of speed. I could well be putting more power (no shock fork, stiffer bike, better Q factor) into a less-efficient drive train, but as long as the net is unchanged, I am happy.

I’m not yet as comfortable no-hands as I was on the old bike, but I expect I will get there. Lightly loaded, it’s fine.

Also: Xtracycle conversion kit (FreeRadical, parts, sideloaders, rear brake and cassette-ready 26″ wheel, plus NEW SnapDeck), one owner, rather used, looking for a good home. Installation help is available.

I mean, it’s been weeks.
It’s worth remembering, when you hear anti-universal-care propaganda, that they are lying. We pay more than anyone by a large margin, measured either in absolute dollars, or as percentage of our GDP, and among “modern” countries (OECD, more or less) we die sooner, we live sicker, and more of our babies die in infancy.

If you hear it would be expensive, remember that what we have now, is already expensive — it’s just, instead of “taxes”, you pay “insurance premiums” (or your employer pays “insurance premiums”), or you do without because you cannot afford it, (and/)or you go bankrupt. Wheher it is “premiums” or “taxes”, it is money out of my pocket.

The people advertising against this do not have your interests in mind; they have their own financial interests in mind. It takes a lot of parasites in the system for us to spend as much as we do, yet deliver such incredibly inferior results. The parasites fund propaganda.

There are a few people who have this crazy idea that market healthcare solutions would work better. They are airy-fairy theoreticians who ignore the experimental data — 20+ countries with cheaper, better, universal, health care.

When I finally tested this circuit on a real bike connected to a real hub, it didn’t work right. I got light, but it was flickery when slow, and I seemed to notice a little more drag (this had to be in my head) than the old 350mA system, and more important, the system voltage never made it above 9 volts; no chance of charging standlight batteries using this design as a starting point. The discrepanies from the plan (too much power to lights at lower speeds) pulled too much current from the hub, so it did not run in the more efficient mode.

I figured out the difference between the model and reality, tweaked the model, and then it tracked observations. With that, I could test a fix before soldering it in, and it worked just fine.

First in-the-field update — vibration cracked off the lead to one of the LEDs, which allowed a test of the voltage shunt. It worked — 300+ lumens from the overvoltage shunt, POW. So the shunt works, too. I haven’t yet managed to activate it with mere speed, up to about 25mph (and now my speedometer is busted, so further measurements need to wait on that).

The fix was to realize that I had not included the resistors used to set the precision voltage sources in my model — they’re voltage sources, PLUS a resistor.

This yields this graph of circuit behavior, which checks out against the actual board. It’s not exactly what I want because the current rises above 200 milliAmps before the cutoff knee, but it’s not far off, and it tests ok in real life.

On the actual bicycle, it has the following desirable attributes:

• Flicker stops between 3 and 3.5mph (5-5.5 kph)
• Lights are on “pretty bright” at 5mph (8kph)
• The system voltage is around 20 volts at 10mph, rising to about 24 volts at 20mph
• The shunt has not yet proved necessary. It trickles about a half a milliamp through the dump, which is more than planned, but not enough to care. This is enough to put the barest light into a power LED.

I have not yet run it down a big straight hill to see if I can force enough current through the shunt to get serious light out of the shunt power LED. The circuit below is the old one, with the obvious repair of the wrong choice of op-amp pinout. For grins, if you view the image and remove the width limit (“?w=700″), you can check my solder work. It’s uneven in two places — the diodes have big wires, in the middle of big pads of copper, and the replaced resistors were done outdoors in the breeze, with a new (and somewhat unfamiliar) soldering iron.

It could still be improved, of course; the board is relatively big and non-compact (then again, that made it easy to work on, so no complaints) and a standlight function would be ducky. I’ll work on that next, in my copious free time.

If you want one of your own, you can get the parts from Digikey, and the PCB from BatchPCB.com. You’ll need to refer to the schematic and board pictures below to get the component values right. Be sure to get an LM258 (not LM358) if you intend to use this in freezing weather.

New board, with fixed opamp selection, and fixed resistor values.

Schematic.

If you want to play with this, here’s all the files.

LICENSE: Creative Commons Attribution-Share Alike 3.0

Gerber files.

Eagle files (board, schematic, library, project).

Qucs simulations.

Once upon a time, Apple wrote this great book on human interface guidelines. It seems nobody reads it any more.

Here’s an example of an insultingly bad error message. A “build” failed (it means, there was a bug in someone else’s program). It pops a little error message window, with the option to show the “details” of the failure. Curious guy that I am, I ask to see them. I get this:

The additional information in the “details” is “assertion failed”. No hint of a log file to look at, no stack trace, no clue as to the cause.

The best part is not quite clear just looking at this picture. Clearly, when I made this image, I had selected the text of the “details”, so I could copy it into a bug report to tell them about how the message was really quite uninformative, and needed to be better (I have my suspicions about the cause of the build failure — however, with no details, who knows?). HOWEVER, this was pointless, because though the text was selectable, it was not copyable. I took this picture so that I could capture the text to file the bug report.

https://bugs.eclipse.org/bugs/show_bug.cgi?id=272012

I am aware that my bug report lacked diplomacy. I hope that you would recognize that Eclipse has serious deficiencies along the path to making a good bug report. If it is not important enough to announce the location of the log files in a dialog from which I can select the name, are those log files that important? (and the crash put them in an unexpected place!) If it is not important enough to put the plugins in a form that allows me to quickly summarize which ones I’ve got, is that information that important?

In short, the Eclipse UI does not look like it was designed to generate good bug reports — it put several STUPID extra steps in my way, and left it up to ME to find the information that might be relevant, or not.

This bug is, for me, 99% reproducible, and has been for weeks, through at least the 3.4.1 and 3.4.2 releases. I suggest you try installing the plugins that I installed (you want a concise summary, you make Eclipse generate it more easily — I gave you an ls -l, you ought to be able to grep out the irrelevant ones), wait 24 hours, and get the latest-and-greatest Scala plugin and watch it crash.

I think it is highly unlikely that you actually tested this, since you need to wait for a plugin to get updated. A link to a “how to politely report bugs” site is not evidence that you tried diddly squat.

Furthermore, the bugzilla UI leaves plenty to be desired. It should let me know that there will be a chance to add attachments in a later step, it should make it clear ahead of time that there is a 65536 character limit BEFORE I generate 65k characters. This lack of design induces pilot error and leads to pissed-off pilots, who will then not really give a rat’s ass about your snotty instructions about how to file bugs. You think you’re an expert, yet present me with this crap UI (unselectable text in important messages)? Puh-leeze.

And I think this matters. I am a long-time Eclipse user, and my enthusiasm about recommending Eclipse has gone down in recent months. After all, I’m going to look like an idiot if I recommend a piece of software that crashes on every update, aren’t I?

What we spend, what we get. Small words, for small brains. We’re not best in the world. Clear?

All data from nationmaster.com, who in turn got it from reputable sources. Add to this, the economic inefficiency of having insurance tied to your job, of all the small businessmen that never got the chance because they could not hack the lack of health insurance, of the tremendous penalty paid by mature industries with mature workers that they insure.

Last I heard, everyone in the OECD except the US, Turkey, and Mexico, has universal health care. Nobody spends more. Most countries have (much) healthier citizes. To suppose that if we tried universal care, we would spend even more than we do now, and be unable to match any of our 20+ peers, is difficult to believe.

And yes, of course lots of people in nice suits will very nicely explain how none of this could ever work, and why we cannot change the system we have now. Those people in nice suits are well paid to say that — all that extra money that we spend ends up in someone’s pocket, and that someone would very much prefer that we do not economize them out of existence.

	y2000 OECD per capita spending (OECD)
United States:  	$4,631.00
Switzerland: 	$3,222.00
Germany: 	$2,748.00
Iceland: 	$2,608.00
Canada: 	$2,535.00
Denmark: 	$2,420.00
France: 	$2,349.00
Belgium: 	$2,268.00
Norway: 	$2,268.00
Netherlands: 	$2,246.00
Australia: 	$2,211.00
Austria: 	$2,162.00
Italy: 	$2,032.00
Japan: 	$2,011.00
Ireland: 	$1,953.00
United Kingdom: 	$1,764.00
Finland: 	$1,664.00
New Zealand: 	$1,623.00
Spain: 	$1,556.00
Portugal: 	$1,439.00
Greece: 	$1,399.00
Czech Republic: 	$1,031.00
Hungary: 	$842.00
Slovakia: 	$690.00
Mexico: 	$491.00

	y2004 % of GDP spending
United States:  	15.40%
Marshall Islands: 	15.20%
Kiribati: 	13.70%
Malawi: 	12.90%
Lebanon: 	11.60%
São Tomé and Príncipe: 	11.50%
Switzerland: 	11.50%
East Timor: 	11.20%
Germany: 	10.60%
France: 	10.50%
Austria: 	10.28%
Serbia and Montenegro: 	10.10%
Iceland: 	9.90%
Monaco: 	9.90%
Canada: 	9.80%
Jordan: 	9.80%
Portugal: 	9.80%
Palau: 	9.70%
Norway: 	9.70%
Belgium: 	9.70%
Argentina: 	9.60%
Australia: 	9.60%
Malta: 	9.20%
Netherlands: 	9.20%
Sweden: 	9.10%
Brazil: 	8.80%
Italy: 	8.70%
Israel: 	8.70%
Slovenia: 	8.70%
South Africa: 	8.60%
Denmark: 	8.60%
New Zealand: 	8.40%
Bosnia and Herzegovina: 	8.30%
Uruguay: 	8.20%
Nicaragua: 	8.20%
United Kingdom: 	8.10%
Spain: 	8.10%
Croatia: 	8.00%
Luxembourg: 	8.00%
Macedonia, Republic of: 	8.00%
Bulgaria: 	8.00%
Hungary: 	7.90%
Greece: 	7.90%
El Salvador: 	7.90%
Colombia: 	7.80%
Japan: 	7.80%
Suriname: 	7.80%
Turkey: 	7.72%

	y2008 life expectancy (years) at birth
Macau:  	84.33
Andorra: 	82.67
Japan: 	82.07
Singapore: 	81.89
San Marino: 	81.88
Hong Kong: 	81.77
Australia: 	81.53
Canada: 	81.16
France: 	80.87
Switzerland: 	80.74
Sweden: 	80.74
Guernsey: 	80.65
Israel: 	80.61
Iceland: 	80.55
Anguilla: 	80.53
Cayman Islands: 	80.32
New Zealand: 	80.24
Italy: 	80.07
Gibraltar: 	80.06
Monaco: 	79.96
Liechtenstein: 	79.95
Spain: 	79.92
Norway: 	79.81
Jersey: 	79.65
Greece: 	79.52
Austria: 	79.36
Malta: 	79.3
Faroe Islands: 	79.29
Netherlands: 	79.25
Luxembourg: 	79.18
Martinique: 	79.18
Germany: 	79.1
Belgium: 	79.07
Guam: 	78.93
Virgin Islands: 	78.92
Saint Pierre and Miquelon: 	78.91
United Kingdom: 	78.85
Finland: 	78.82
Man, Isle of: 	78.8
Jordan: 	78.71
Korea, South: 	78.64
Puerto Rico: 	78.58
Bosnia and Herzegovina: 	78.33
Bermuda: 	78.3
Saint Helena: 	78.27
Cyprus: 	78.15
United States: 	78.14
Denmark: 	78.13
Ireland: 	78.07
Guadeloupe: 	78.06
Portugal: 	78.04
Albania: 	77.78

	y2001 expected healthy years (OECD)
Japan:  	73.6
Switzerland: 	72.8
Sweden: 	71.8
Australia: 	71.6
France: 	71.3
Iceland: 	71.1
Austria: 	71
Italy: 	71
Spain: 	70.9
Norway: 	70.8
Luxembourg: 	70.6
Greece: 	70.4
New Zealand: 	70.3
Germany: 	70.2
Finland: 	70.1
Denmark: 	70.1
Canada: 	69.9
Netherlands: 	69.9
Belgium: 	69.7
United Kingdom: 	69.6
Ireland: 	69
United States: 	67.6
Portugal: 	66.8
Czech Republic: 	66.6
Poland: 	64.3
Slovakia: 	64.1
Mexico: 	63.8
Hungary: 	61.8
Turkey: 	59.8

	infant deaths per 1000 live births
Singapore: 	2.3
Sweden: 	2.75
Japan: 	2.8
Hong Kong: 	2.93
Macau: 	3.23
Iceland: 	3.25
France: 	3.36
Finland: 	3.5
Anguilla: 	3.54
Norway: 	3.61
Andorra: 	3.68
Malta: 	3.79
Czech Republic: 	3.83
Germany: 	4.03
Switzerland: 	4.23
Spain: 	4.26
Israel: 	4.28
Korea, South: 	4.29
Slovenia: 	4.3
Denmark: 	4.4
Austria: 	4.48
Belgium: 	4.5
Liechtenstein: 	4.52
Guernsey: 	4.53
Luxembourg: 	4.62
Netherlands: 	4.81
Australia: 	4.82
Portugal: 	4.85
Gibraltar: 	4.91
United Kingdom: 	4.93
New Zealand: 	4.99
Jersey: 	5.01
Canada: 	5.08
Ireland: 	5.14
Monaco: 	5.18
Greece: 	5.25
San Marino: 	5.44
Taiwan: 	5.45
Italy: 	5.61
Man, Isle of: 	5.62
Cuba: 	5.93
United States: 	6.3
Faroe Islands: 	6.46
Croatia: 	6.49
Belarus: 	6.53
Guam: 	6.55
Lithuania: 	6.57
Northern Mariana Islands: 	6.72
Cyprus: 	6.75
Poland:  	6.93

I decided, that the next circuit I build, will do one thing well, and I’ll worry about the rest later.My brother and I have been over the issue of “what do you want in a bicycle light”, and what we both basically want (and both currently compromise on) is something that is bright as all get out, silent, requires no fiddling to keep it running, and also stays on for a bit when stopped.

Crudely, a “dynamo plus a battery plus LEDs”, but the problem is that you can overcharge your batteries, and the battery control chips that I can find seem to assume conditions that do not hold on a bicycle, like civilized temperatures that don’t change much. So, instead, I plan only to either “do no harm”, which means that the standlight eventually starts to dim and the batteries need charging (that might take months), or perhaps to be far less ambitious in my charging plans — stop at 80%, instead of aiming for 100%. It looks like voltage changes suffice for Sanyo Eneloop, if I stop early (and I want to use Eneloop because they have low self-discharge).

It turns out that 4 lithium batteries would probably be just right. Compact, light, and if you stay below a critical voltage you won’t overcharge them. 3 volts per cell is the minimum output, 4.1 volts per cell is the maximum charge, call it 4 to be safe, and the operating voltage will normally be in the 3.5-3.7 range. Times 4 gives a minimum of 12, usually 14, which is plenty to run the lights at a medium level. I still need to scope out the temperature effects, but wonder of wonders, the ATTiny chips have an on-board temperature sensor.

I also decided that the previous circuit was just too spread out; real estate costs. It should be possible to stand things up, perhaps with some tape and/or glue to help lash them together and cut down on vibration. I also simplified the 3-part ramp to the BuckPuck control, and settled on a 2-part circuit instead. The shunt is not quite as sharp, but is still better than what I am using now.

The trickier part (for me, now merely a hobbyist EE) was to figure out a circuit that could be powered from the battery itself, that would allow current to flow in either direction when switched on, and that could be switched with TTL voltages. Given that, a microprocessor can sit and watch the wheels go round, or not, and watch the system voltage, and when things look slow, it can turn on the batteries. Optionally, it can also turn off the lights, so that they can be blinked or varied while stopped, and it can get limited inputs (inconveniently, using analog-to-digital conversion) on its reset pin.

A bicycle light system does get to play one trick that is not available to (say) a cell phone or a personal computer; it can turn the lights on or off for a few milliseconds while the batteries are charging, and see how the battery voltage responds to the change in the load, and nobody’s going to notice. This might be a good indicator, I don’t know yet. It is, of course, a Simple Matter of Programming.

The software does NOT get to control the shunt that clamps the system voltage below part-frying levels.

The other tricky part is not knowing exactly how much noise there will be in this system.
The BuckPuck is a literal black box, but is presumed to contain a switching power supply. The capacitors feeding it are big. The microprocessor is alleged to need noise reduction, so I took the recommended precautions and followed them as well as I could (so, the ground fed to the voltage dividers for the analog signals, is at the other end of the circuit from the ground feeding the processor, and I decorated the processor with capacitors. Is this enough?

There’s a learning curve. I think I’m past it.
I repaired the busted board. It looks like this:

The so-called second-second-system, using an Atmel168 chip, intended to use Arduino software (minor mods required for lower clock rate), with the option of controlling the LED power profile in software, and also of controlling external electronics.

The pins in a fence around the microprocessor are intended to allow a piggy-back board with additional electronics to be loaded on top. For example, the lights might be controlled independently, or perhaps some rechargeable batteries could be used to provide a standlight, or maybe a USB charger for cellphones and cameras.
Here’s little board to allow a 20+ volt battery pack to be connected or disconnected, under the control of pin B0. Do note, the ability to do battery management is pretty limited — NiMH batteries are finicky, it’s hard to get a good handle on the current going in or out of the battery (10-bit DAC ranging to 5 volts, at best is measuring current through a 1 ohm resistor, with a 5mA resolution. Typical currents will be in the 200mA range, or less, so one part in 40.

I thought about the costs a bit, and had some ideas about simplifying the first board, and arrived at this smaller, simpler board. The shunt comes on a little earlier, but it uses the same ramp-up trick to direct power to the LEDs as the system voltage increases.

Every system is a second system, or so it seems.

Update: it really, really helps to use the right chip. After being incredibly careful about every orientation of every weird transistor, I designed the board around a “LM387N”, which is not the chip I intended or the one that I had ordered — LM258 (usually referred to as LM358, but the 258 works in the cold). I was testing the board before putting on the big components, and it was just not working, and in the end, it seemed like the op-amp was on drugs. In the end, I got out a magnifying glass to check the label. DOH!

All is not lost, but the board needs to be rerouted. I’ll sort out the pins and test again.

Up-update: After installing a pin-scrambler, it works roughly as predicted. The most important part of the circuit, the high-end shunt control, delivers 3.85 volts to the base of the shunt transistor at 31.1 volts. This is a perhaps a half-volt earlier than planned, but it’s a half-volt to the safe side. In the high 20s, the shunt is not activated, as planned.

Anomalies:

(1) the LED control sends a little more power to the lights than I had planned. Not lots, and it’s mostly a low-end effect. At 31.1, the control voltage is 1.7, which is exactly on target. At 22.5 it is 2.6 (plan, 3.1) and at 12.9 it is 3.24 (plan, 3.6) and at 9.6 it is 3.4 (plan, 3.8).

(2) the off-level shunt base voltage is 0.6, where the plan was 0.2. This should not be an issue with the particular Darlington shunt transistor; it claims to have a Vbe of 2.8 volts.

I’ve decided to take the plunge and learn how to get PCBs made.

The new hub (Shimano 3D71) produces too much power, at too high a voltage if your LED circuit does not use it all. The new plan uses a piecewise linear function of the input voltage to drive the voltage control circuit on a BuckPuck, plus a shunt to guard against overvoltage. The idea is that at lower power, enough light is produced, but at high power, the excess is dumped into the lights, with the intent that somewhere between 30 and 31 volts, the LEDs are receiving near-maximum power (9-10 watts), and the shunt is activated to drain up to an amp (30 watts) to keep the voltage from going any higher. The shunt circuit allows an external device (as many as 7 LEDs!) to dissipate the power instead of the transistor. My plan is to use an Endor Star (3 surface mount Luxeon Rebels on a single puck) that I bought but could not find a use for (could not heat sink effectively enough) to dissipate up to 10 watts externally. The goal is to ensure that the supply voltage is limited to 32 volts, no matter what, because the op-amp and the BuckPuck are rated for 32, the capacitors are rated for 35, and the diodes are rated for 40.

The combined circuit is just complex enough that it should be on a PCB, which should also make the result more durable. Vibration on a bicycle is a serious problem.

Resources:

• Digikey

  I buy parts here. I can make sense of their search, they generally have links to the documents that are inevitably necessary to figure things out.

• Sparkfun

  More parts, more discussion of parts, more tutorials. A good place to find answers to Eagle questions.

• Instructables

  This is where I learned how to make autorouting parameters more fumble-fingers-friendly.

• Qucs

  Qucs is a circuit simulator, a little quirky, but it gets the job done, and gives me some confidence that my Great Designs will (or won’t, as the case may be) work.

• Eagle

  CadSoft sells Eagle, software for designing printed circuit boards. The full version does way more than I need; the free version is enough for me, at least so far. The autorouter is mostly my friend, especially once I got used to its quirky habits.

Qucs tips:

• Consider device specs

  Quite a few zener diodes don’t reach their rated voltage until they have quite a few milliamps flowing through them. If those milliamps are also flowing into the base of a transistor then the shunt will have a long, wasteful “on ramp”.

• Consider temperatures

  Diodes, including LEDs, have a forward voltage that increases significantly when the temperature falls from 35 degrees to -5 degrees.

• Consider device precision

  If you want a ramp that comes on at 30.5 plus-or-minus .3 volts, that is a 1% device. That suggests that the resistors had better be at least that good.

• Tinker and fiddle

  Just because it looks like it works, how will it behave in the real world with slightly different parts? Tinker with element values, tinker with resistor gain, try monkeying with the forward voltage on transistors. Push things around till the simulation goes wonky, so that you can know how large your sensible-operation margins are.

Eagle tips:

• Instructables design rule changes
• Pouring polygons, joining nets
• Removing thermals

  In the board editor, click rats-nest to get the polygons poured, then click info on a polygon. One of the info options is a “Thermals” checkbox. You might want thermals; they are intended to make it easier to solder onto a ground plane, which can sink a lot of heat from a soldering iron tip.

• Using restricted areas to make the autorouter behave

  I had a heck of a time get the autorouter to behave, and not rip up polygons. One way to help it is to remove choices; by putting a long skinny polygon of tRestrict or bRestrict, I force it to not cut across some polygons (polygons are optional, tRestrict and bRestrict are not).

• Must have a 3-button mouse

  Working on a Mac laptop, pretty quickly you hit one of those “hit control-right-button for next choice” items in the Eagle UI. Unfortunately, by default you hit control with the trackpad to get it to do “right button”, so you this doesn’t work. A 3-button mouse is a big help. (I went rummaging in System Preferences, and just now discovered the two-fingered “secondary click”, not sure that works with control also).

• Checking design rules
• Important layers
• Checking signals

Schematic

Oh yeah, about the copyright. I’d like to open source it, but cannot figure out how to put the whole GPL on one PCB. I’m a bit ticked off at all the bicycle lighting companies; everything costs too much, way too much. Bought retail, the cost of the components on this board is in the 15-20 dollar range, plus the buckpuck, which is another $15. LEDs+lenses cost about $10 apiece, you need at least 3 for a bike (4 is better), unless you want to also use an LED to help burn off excess voltage. The board itself, quantity 1, costs $45, with shipping and handling. SO, all the parts, except for the hub, cost under $100, and direct 200+ lumens out the front of the bike, and boatloads to the back. Dinotte is probably the best deal out there, and their 200 lumen headlight is $160 all by itself. Add a taillight, another $160. Add a fog/running light, another $160 — here, a fog light costs you an SPDT plus two more LEDs (i.e., $25 more).

This graph shows the intended operation of the varying power control.
At low speeds, the buckpuck is not really on, but if the drive current is low it will start to come on at around 10 volts. 350 mA is the goal; that corresponds to 16 volts, which the hub seems to cheerfully generate if not all its current is sunk. Beyond that, nothing exceeds like excess, we’ve got to do something with the power, might as well turn it into light.

The “efficiency” line is an approximation of the Buckpuck’s predicted efficiency, to allow a rough extrapolation of current draw from the hub/rectifier. The design goal was 200 mA, which correponds to greatest efficiency, according to the data at pilom.com. Higher voltages and lower currents are good; they correspond to lower drag on the bike, and less heat in the hub itself.

The clamp was much harder to get right. “Zener” diodes are not as stable as you might like, and plain old diodes can vary from about 0.6 to 0.8 volts in their forward bias, depending on temperature. It was very helpful to learn that Qucs includes temperature data for many of their devices, so that the shunt could be simulated at various inhumane temperatures. I see this temperature dependence already in the shunt that I use; it comes on more easily when it is cold. So, don’t use Zeners, do use voltage reference devices.

Note: Pr2.I is the current through the shunt. VREFmilliAmps is the current through the voltage reference that controls the shunt. Pr1.V is the drive voltage from the op-amp to the base of the shunt transistor (which is a Darlington, beta of about 20000).

I had to add three devices (two voltage references, and a buckpuck) to the Eagle library. Here they are. The Buckpuck includes a hole drilled underneath its trim pot, if you care to use it. I hope it’s in the right place.

Additional eagle devices

Eagle schematic

Eagle board

I modeled the two circuits separately in Qucs.

Qucs clamp circuit

Qucs varying power circuit

It definitely occurred to me that I could probably program this up with a PIC microcontroller instead (if I could program a PIC microcontroller), and use that to get special modes, allow secondary operation from or charging of batteries, etc. I would not trust the PIC to run the shunt, though.

The general plan would be to feed AC1 and AC2 to a comparator to keep track of wheel rotation, monitor V+ with an A/D to determine the BuckPuck control voltage, and include a current sense resistor on the battery stack to keep track of the cumulative (dis)charge. Outputs would be a relay or MOSFET control voltage on the battery, plus several bits (4?) fed to a D/A (resistors plus an op-amp) to derive a control voltage. So it could definitely be done — but not today.

Turns out I could not help myself; after looking at microprocessor choices, the Atmel chips look like the best bet, mostly because of a better set of tools and community. I planned something that is Arduino-like; it has the same connector, and mostly the same electronics, but some of the pins are different,
and the clock is internal, both for reduced footprint and reduced power consumption. It’s probably best to start with the analog version of this controller first, but this would be a fun project.

I’m still tinkering (I really need to stop), with the current plan being to line up all three headers on the same 1/10 inch grid, so that a daughter board (carrying switching and sensors for battery charging, etc) could be added. One reason for a 3-header daughter board is to make it completely impossible to plug it in wrong, because one of the pins is now the unregulated up-to-32V DC high-power supply. That could cook some parts.

I looked further into battery charging; “optimal” charging is hard, and I think the best bet would be to aim for casual battery management. Since a stack of 20 AA batteries (20-24 volts) could run the lights for perhaps 10 hours at full charge, it is not necessary to aim for 100% charge, and stopping short would be safer. Maybe later.

This is here only so search engines can find it and other people can avoid this mistake.I bought a PIC ICD2-Pocket from SparkFun, intending to program it from my MacBook.

Long story short, there appears to be no Mac OS drivers for the USB chip. I can find lots of advice online that tells me about PC drivers, but I have a Mac, so that is not useful.

This guy seemed to be making a stab at it (in particular, his vendor ID matches), but his page is two years stale.

Here’s drivers for handheld devices, but those seem slightly non-serial (IP?)

This looks better yet. Allegedly a serial driver, for a variety of devices. Perhaps it will suffice to edit the kext properties.

I’ll update this if I succeed. If YOU succeed (with a Mac, not a PC), please let me know in the comments. Otherwise, this is a dead end.

This also suggests that communication with between the USB on a PIC18F2550/4550 will not work so well. That seems implausible, but lack of a driver, is lack of a driver.

Update: libusb (from MacPorts, for example) can see devices without kext drivers. I bought a Pickit2; that seems to be the better choice, and both Pk2cmd and piklab can see it (note that piklab from MacPorts does not build out of the box because it depends on arts, which is broken). I have not yet successfully programmed the PIC, but I have not yet given it a serious try with the new programmer.

Update 2: The new version of Piklab at MacPorts recognizes the Pickit, but is surprised by the firmware version. Pk2cmd seems to work, once the firmware upgrade is applied to the Pickit. To do that, you must install pk2cmd, and then type “pk2cmd -DPK2V023200.hex” to download the new firmware.

MacPorts has support for sdcc — no, not really, things fail later. Better to follow these instructions instead, which first require installation of gputils.

Gputils you can download from sourceforge. Be sure to choose a directory that has no spaces in its path name. Then configure: “./configure –prefix=/opt/local”, and “make”, and “make install”. This installs it into the MacPorts /opt/local area, which might be a mistake, but that’s a problem for another day. It seems to build and install ok.

Using the instructions above (which seem a little dated, I sent email to the author, but no reply yet) I got a good deal further and installed sdcc, but assembly of part of the test program failed. It could be for some stupid reason, but I ran out of steam. Mail to author:

Your instructions are more helpful than anything else I have come across, but they seem to be getting maybe a little dated, and there's at least one step missing.

Dated:

3.5. Compiling/Installing the Remaining Device-specific PIC Libraries

this part doesn't match what is there any more; perhaps it is not necessary, I cannot tell for sure.

Missing:

How to get crt018.o ?

"Obviously", do this, but then it fails:

gpasm -c crt018.asm
crt018.asm:5:Warning [205] Found directive in column 1.
crt018.asm:5:Error [176] incorrect syntax. use `CONFIG KEY = VALUE'
crt018.asm:6:Error [152] Executable code and data must be defined in an appropriate section.

(Source code follows  -- perhaps there was a mistake in the cut/paste, or the html?)
----------------
   processor  18F2550
   radix  DEC
   include "p18f2550.inc"

config CODE
   DATA 3F32H

IDLOCS CODE
   DATA 1234H

etc.

Today’s lesson: Hub generators are current sources.The sidewall generator was starting to tick me off. Fiddly, the mount cracked once, a little noisy, and too slippy in the winter. So, I took my Christmas money, and bought a custom wheel from a local bike shop, with a Shimano 3D71 in the middle.

Hooking this up to the regulator for the sidewall dynamo worked great, till it failed. Simulation of the circuit revealed two unexpected things; the two small capacitors were probably swinging well past their voltage limits, and they were also getting reverse-biased in the process. However, the part that actually failed, was the power transistor in the voltage clamp; it had gotten so hot that the epoxy on it was cooked to a dark brown. Replacing that transistor restored the clamp to its former function, but this time I added an indicator LED (a power indicator LED) to let me know whether the clamp was running part time, or often. Turns out, the light went on solid at about 10mph, and got plenty bright at 15mph.

Given that the clamp is supposed to only turn on at 23.5 volts, and sinks 350mA, that the hub was delivering somewhere in the neighborhood of 10 watts of power (it’s rated for 3).

It seems a shame to convert all that useful power into heat and transistor flambé. Plan B, is to string all the power LEDs I’ve got kicking around into one big load, and either switch that on with the clamp, or maybe just use that for the clamp. Eight LEDs would be a little pricey, but wow, think of the light.

Or maybe I could run a stereo.