Friday - May 25, 2007
Until the bugs get fixed...
Moved to Wordpress.
Though I am trying out RC3, just to see.
See http://dr2chase.wordpress.com/ .
A lot more work to do stuff, but I'm not exactly ignorant of html, and this way, it actually works.
To fix this, use agitprop to convert iBlog to Movable Type format, then edit the file fixing links to images and removing line breaks. You'll need someplace to put your pictures.
Still haven't figured out how to make their RSS work.
Thursday - May 24, 2007
iBlog bug report
Release Candidate 2?
Problem #1: On the front page of the blog,
http://homepage.mac.com/dr2chase/iblog/
at the bottom of each entry, there is a "LINK" that goes nowhere:
http://homepage.mac.com/dr2chase/iblog/C1163187547/CA273233-CB28-45A3-B556-38D5E7831C01/index.html
However, if I remove the funny-looking stuff in the middle
http://homepage.mac.com/dr2chase/iblog/C1163187547/index.html
I get to a page that looks right.
Problem #2:
On the file-sharing page, there is a link to
http://homepage.mac.com/dr2chase/iblog/C1163187547/E20070515081339/Archives/MVI_2134.AVI.zip
but of course (why else would I mention this?) there is no such file there. I mounted my iDisk, and typed
/Volumes/dr2chase/Sites/iblog chase$ find . -name "MVI*" -print
and in fact there is no such file, anywhere.
Problem #3:
The RSS feed has extraneous formatting gunk in it. For example "<p>A Little Pufferfish Video <p>"
instead of "A Little Pufferfish Video".
Problem #4:
The "Help" is largely missing.
I think I should have been more careful before buying.
(When all these things are fixed, it will be obvious to anyone reading this blog.)
Tuesday - May 15, 2007
A Little Pufferfish Video
13 Seconds of 640x480 15 fps fishes in a jar at the beach
We took this at Honeymoon Island, Dunedin, May 13, 2007. Katie and Thomas and Ben captured quite a few little fishies at the beach, then let them go.
(I've never done an entry like this, it will be interesting to see if it works.)
Sunday - April 15, 2007
More fun
I had a recent birthday windfall and I've ended up spending it on (mostly) bike stuff. Since some of this spending is experimental, I thought I should file a (not-entirely objective) report.
Schwalbe Big Apple tires (the biggest ones): Big Fun. Useful fact: 60mm tires on a 26-inch rim, are actually a little larger than 28mm tires on a 700c rim. I liked them enough that I am buying a pair of cheap 26" disk rims for my xtracycle (a 700c bike) to see how they do there. I am sure the rolling resistance will be higher, but Big Apples do very well on rotten roads, which we have in abundance. I also plan, in the name of Science, to do some timed downhill rolls with both wheels, to see how they compare for rolling resistance. I'm sure the high pressure tires are better, but how much better? One downside of Big Apples in the rain -- they pick up and throw an enormous amount of water.
On the advice of icebike.com, I decided to try some tights, socks, and balaclava from Foxwear. The (medium-weight) tights were absolutely everything that the guys at icebike claimed; I took them out in 35-degree weather and had no complaints at all. The balaclava was also excellent, and I was able to move it around on my face in just the right way to avoid fogging my glasses, and it kept my head warm in cold sleet-and-rain.
Monday - April 09, 2007
Longtail fun
I ride an xtracycle, which is a bike plus an
extension, a most-convenient way to carry many things. It looks like long bikes
are becoming more popular.
My xtracycle -enhanced bike (towing another bike),
and some groceries.
Tired driver of little bike not shown, but she rode on the deck in front of her bike.
Spencer Wright's Chupacabra. I think the crazy interim chain wheel is supposed to cut the chain flop, and keep it out of the underbrush, but I don't really know.
Surly's Big Dummy . Quien es mas macho?
A longtail "Vanilla " (via CleverChimp). This bike is just plain beautful.
Clever Chimp's Xtravois
Here's a couple of pretty bikes from Fraser Cycles . These are custom, non-Xtracycle beasts. Notice the chain guide below; this is a good thing, because otherwise the long bouncing chain can (sometimes) shift gears for you.
And this is one heck of a rack.
Clever Chimp deserves all the credit for finding most of these or pointing them out, but he's got them scattered all over his blog. I wanted a bunch of photos so I could point to them and say "these are longtails, aren't they cool?"
Tired driver of little bike not shown, but she rode on the deck in front of her bike.
Spencer Wright's Chupacabra. I think the crazy interim chain wheel is supposed to cut the chain flop, and keep it out of the underbrush, but I don't really know.
Surly's Big Dummy . Quien es mas macho?
A longtail "Vanilla " (via CleverChimp). This bike is just plain beautful.
Clever Chimp's Xtravois
Here's a couple of pretty bikes from Fraser Cycles . These are custom, non-Xtracycle beasts. Notice the chain guide below; this is a good thing, because otherwise the long bouncing chain can (sometimes) shift gears for you.
And this is one heck of a rack.
Clever Chimp deserves all the credit for finding most of these or pointing them out, but he's got them scattered all over his blog. I wanted a bunch of photos so I could point to them and say "these are longtails, aren't they cool?"
Sunday - April 01, 2007
Comcast Digital Voice bait-and-switch
If someone from Comcast makes you a very
attractive offer to upgrade your service to include Digital Voice, they're
lying.
I haven't figured out what the official corporate
excuse was, but we were offered an upgrade to include Digital Voice at a very
good rate. And, indeed, "Digital Voice" is under $10/month. HOWEVER, somehow
the rest of the services now cost enough extra that instead of a $57 bill, I got
a $153 bill. $32 was one-time charges that I would swear I was told would not
be charged. I assure you, if they had told us that these would be the charges,
we would not have done this, and we pressed the sales guy hard on the details
(or so we thought).
I'd like to think that I could convince my family that it would be ok to cancel the whole damn thing for a few months, but I don't think I can pull that off. I'm not in a mood to waste my time on hold or talking to people without the authority to fix the problem. I could try the trick that I used on the phone company once; record my grievance on the computer, and just leave it playing in a continuous loop with the phone handset sitting next to the speaker, while I spend my time usefully.
So, be warned. Comcast salespeople will flat-out LIE TO YOU. I know that salespeople tend to put a happy face on things in general, but this is the first time I've ever been lied to like this.
I'd like to think that I could convince my family that it would be ok to cancel the whole damn thing for a few months, but I don't think I can pull that off. I'm not in a mood to waste my time on hold or talking to people without the authority to fix the problem. I could try the trick that I used on the phone company once; record my grievance on the computer, and just leave it playing in a continuous loop with the phone handset sitting next to the speaker, while I spend my time usefully.
So, be warned. Comcast salespeople will flat-out LIE TO YOU. I know that salespeople tend to put a happy face on things in general, but this is the first time I've ever been lied to like this.
Saturday - March 31, 2007
Muffler-shop handrail
Need a handrail, feeling cheap? Get some help
from a custom muffler shop.
We replaced the concrete pad and rusty steel
rails at our front door with something better -- prettier, less steep stairs,
and bench-style sides to sit on. We had to add a handrail to meet the building
code, and spent some time trying to figure out something that would not mess up
our nice stonework. I had seen mention of using a muffler shop to get metal
bent , so we decided to try
that.
I had a bender for electrical conduit, so I bought some of that (it's cheap), and made a prototype. The goal was something that would attach only at the house and the ground, and would be strong enough to not need any connections to the stonework. My conduit was only ten feet long, which was not quite long enough, but it was enough to include all the bendy parts.
I marked each section and bend with what I intended, which was a help, because even though I didn't quite get the prototype right, the guys at the muffler shop (who are trained professionals) could figure out what I wanted and where the mistakes were, and fixed it. There were three bends -- a 30 degree bend to follow the steps down, then a 10 degree bend to put the base outside of the stone walkway, and then a 60 degree bend down to that base. The 10-60 combination, I did not get right, but apparently this is an elementary mistake addressed in MufflerShop 101.
It turns out that the stock length for muffler pipe is ALSO ten feet, so the muffler guys captured all the bendy parts, but I was left with a little more than a foot to finish. This turned out to be a feature, not a bug, because the last section gave me a way to fine-tune the railing position.
The muffler guys put a 3-bolt plate on the each end. One end is bolted to the house (we need to find prettier monster bolts).
The other end is bolted (nutted?) to three pieces of threaded stock. The three attachment points, and six nuts, are what allows the fine tuning.
The other ends of the threaded stock are attached to a metal tray, which is embedded in a pile of rocks and concrete. The nuts on top of the base are used to clamp down as tight as possible on the concrete and hold it in compression, so that the attachment is rigid and the concrete won't crack. There's also epoxy underneath. It looks flimsy, but it absolutely is not
This is what it looked like when I finished. The height is specified by code..
The railing is parallel to the stone work in both planes.
There is a code-specified clearance between rail and "wall", and when your wall is a piece of rock, it matters.
Here you can see the 10-degree bend and how it is necessary to put the base outside the walkway.
and here, the finished railing from the street. The lower section of the rail is vertical.
The PVC pipe on the ground is a temporary sorta-cosmetic cover, removed to show how the lower attachment works. We need a perhaps a garden gnome, or maybe several seasonal-themed covers.
The paint, is Toyota nondescript brown, the same as our Toyota, which was visiting the body shop to fix a scrape. Why get two bottles of touchup paint?
I had a bender for electrical conduit, so I bought some of that (it's cheap), and made a prototype. The goal was something that would attach only at the house and the ground, and would be strong enough to not need any connections to the stonework. My conduit was only ten feet long, which was not quite long enough, but it was enough to include all the bendy parts.
I marked each section and bend with what I intended, which was a help, because even though I didn't quite get the prototype right, the guys at the muffler shop (who are trained professionals) could figure out what I wanted and where the mistakes were, and fixed it. There were three bends -- a 30 degree bend to follow the steps down, then a 10 degree bend to put the base outside of the stone walkway, and then a 60 degree bend down to that base. The 10-60 combination, I did not get right, but apparently this is an elementary mistake addressed in MufflerShop 101.
It turns out that the stock length for muffler pipe is ALSO ten feet, so the muffler guys captured all the bendy parts, but I was left with a little more than a foot to finish. This turned out to be a feature, not a bug, because the last section gave me a way to fine-tune the railing position.
The muffler guys put a 3-bolt plate on the each end. One end is bolted to the house (we need to find prettier monster bolts).
The other end is bolted (nutted?) to three pieces of threaded stock. The three attachment points, and six nuts, are what allows the fine tuning.
The other ends of the threaded stock are attached to a metal tray, which is embedded in a pile of rocks and concrete. The nuts on top of the base are used to clamp down as tight as possible on the concrete and hold it in compression, so that the attachment is rigid and the concrete won't crack. There's also epoxy underneath. It looks flimsy, but it absolutely is not
This is what it looked like when I finished. The height is specified by code..
The railing is parallel to the stone work in both planes.
There is a code-specified clearance between rail and "wall", and when your wall is a piece of rock, it matters.
Here you can see the 10-degree bend and how it is necessary to put the base outside the walkway.
and here, the finished railing from the street. The lower section of the rail is vertical.
The PVC pipe on the ground is a temporary sorta-cosmetic cover, removed to show how the lower attachment works. We need a perhaps a garden gnome, or maybe several seasonal-themed covers.
The paint, is Toyota nondescript brown, the same as our Toyota, which was visiting the body shop to fix a scrape. Why get two bottles of touchup paint?
Monday - February 12, 2007
Bike path IED
No, really. Near Boston, even.
But fortunately, it was an itty-bitty one. I saw
a dab of trash in the bike path, and (as I usually do, when it isn't glass) rode
directly over it, and BANG! My tires were ok (and it wasn't really as loud as a
blowout, but a very sharp report) so I circled back to see what it was. And
lo:
Looks like the product of a "bomb factory", doesn't it? I did not alert our panicky authorities, and when I returned this evening it looked like a "bomb disposal squad" had detonated the rest of the charges, rendering it harmless.
I imagine our guys in Iraq wish their "devices" were as silly as this one.
Looks like the product of a "bomb factory", doesn't it? I did not alert our panicky authorities, and when I returned this evening it looked like a "bomb disposal squad" had detonated the rest of the charges, rendering it harmless.
I imagine our guys in Iraq wish their "devices" were as silly as this one.
Sunday - February 04, 2007
Stupid resistor tricks
I needed pairs of resistors matched as precisely
as possible
So I take a bunch in a taped reel, taped them to
a sheet of paper, and measure them on a 3-digits digital meter (rechecking to
see if the measurements repeat, and also slapping my hand on a bunch to check
for thermal drift while measuring). Then I type the measurements into Excel,
sort, find the values with more than one hit, and assign colors for those
values. Color the tape, and we're there -- a bunch of small sets of resistors
that are within .1% of each other. (I used color code ordering with gaps, so a
black and a brown differ by at most
0.002).
Why am I doing this? If the "simple" synchronized rectifier doesn't work, I think I can build one using an LM158 (low-voltage op-amp) together with a pair of precision voltage dividers (to divide inputs at or outside the supply voltages, into valid inputs for the op-amp). I've done this trick once before building a light-board, where I needed 24 sets of 3 matched resistors. For that, I just took 100 resistors, measured them, tossed out the outliers, binned and sorted the rest, and then just took them three at a time starting at one end of the sort. The resistors seem to stay similar over time.
Why am I doing this? If the "simple" synchronized rectifier doesn't work, I think I can build one using an LM158 (low-voltage op-amp) together with a pair of precision voltage dividers (to divide inputs at or outside the supply voltages, into valid inputs for the op-amp). I've done this trick once before building a light-board, where I needed 24 sets of 3 matched resistors. For that, I just took 100 resistors, measured them, tossed out the outliers, binned and sorted the rest, and then just took them three at a time starting at one end of the sort. The resistors seem to stay similar over time.
Friday - February 02, 2007
Woo-hoo, a simple low-voltage synchronous rectifier circuit
If I only had a brain (left) I could have done
this myself.
Actually, I am not 100% sure that this actually
works. It did not simulate very well in Qucs. Maybe I made a mistake. It is
not a very expensive thing to test in real life, so I
will.
Here: http://www.solarbotics.net/library/circuits/misc_switching.html
Via here: http://groups.google.com/group/aus.bicycle/msg/c8734fd7ece1c485?&hl=en
Via here http://www.cyclingforums.com/showpost.php?p=2753864&postcount=44
Which I found among the 5 reported pages for googling "buckpuck synchronous".
Why is this important? At low (bicycle generator, solar cell) voltages, a normal diode drop (0.7V) is a big hunk of the total voltage available. Step one in understanding the circuit is to understand that the MOSFET is run "backwards", and its body diode is used as a "real diode" to kickstart the circuit.
I still need to check that this will work right when translated into its P-channel equivalent.
Here: http://www.solarbotics.net/library/circuits/misc_switching.html
Via here: http://groups.google.com/group/aus.bicycle/msg/c8734fd7ece1c485?&hl=en
Via here http://www.cyclingforums.com/showpost.php?p=2753864&postcount=44
Which I found among the 5 reported pages for googling "buckpuck synchronous".
Why is this important? At low (bicycle generator, solar cell) voltages, a normal diode drop (0.7V) is a big hunk of the total voltage available. Step one in understanding the circuit is to understand that the MOSFET is run "backwards", and its body diode is used as a "real diode" to kickstart the circuit.
I still need to check that this will work right when translated into its P-channel equivalent.
Thursday - February 01, 2007
Update on bike light design
It's always better to use a canned solution,
assuming you can find it.
It turns out that there is a company LEDdynamics designing and selling power
supplies customized for driving LEDs. The BuckPuck, BoostPuck, and BuckToot all
look useful. Their products are generally aimed at the Philips Lumileds
"Luxeon" LEDs. These are high-current LEDs, starting at 350mA and going up.
Definitely, this is the way to go for the power supply, and probably for the
taillight (they sell a very efficient orange-red light with 110-degree spread --
perfect for a taillight). One minor difficulty is deciding whether to go boost
or buck, and also to figure out the best way to convert generator AC to DC. The
BuckPuck claims to run on AC, but it appears that it uses conventional (0.7v
forward drop) diodes to rectify, and this is not a good match for low bicycle
generator voltages.
The choice is not quite so clear for the headlight; other people propose to use a cyan-green color (good for the low-light rods and middle-color cones in our eyeballs) but it seems to me that this would destroy night vision (it IS night vision, but if you got a pop in the eye from this thing, good luck seeing anything else in low light for a few minutes), plus the rods don't give a sharp focus. There's also the issue of GREEN=GO reactions from car drivers; these guys were talking about mountain bikes, not commuter bikes. I'm going to have a look at how the whites I already bought do (they've got efficiency comparable to the Luxeon whites, and a slightly better color mix) but it seems that an amber light might be a much better choice. Amber (590nm) might win because: (1) lower input power, (2) better transmission through the eye itself, (3) YELLOW=CAUTION, (4) it seems to have little overlap with the rod wavelengths, but still overlaps the the "red" cone wavelengths. I can always recycle the whites for indoor lighting; it's pretty easy to wire a boatload of them in series to get to a line voltage, and the result is both efficient and dimmable.
Another useful thing about newer LEDs, especially these high-powered ones, is that that for as much as a 50% current reduction, they still produce light, and it is reduced only linearly -- that is, given a 60mA drive, 4 white LEDs and 3 white LEDs produce just about the same amount of light. This means that it should be possible to fine-tune the voltage of a parallel array, if necessary, and also to provide some headroom for power supply glitches. Under-driving a parallel array of lights also provides protection against runaway failure, which can occur if one LED fails and the current is concentrated onto the remaining LEDs. The amber/orange/red Luxeon III LEDs can be driven through a large range of currents; anywhere from .14A to 1.4A (continuous) produces light, and pulses of 2.2A can be tolerated.
Someone else has looked at this issue, though I'm not sure where they went with it.
The choice is not quite so clear for the headlight; other people propose to use a cyan-green color (good for the low-light rods and middle-color cones in our eyeballs) but it seems to me that this would destroy night vision (it IS night vision, but if you got a pop in the eye from this thing, good luck seeing anything else in low light for a few minutes), plus the rods don't give a sharp focus. There's also the issue of GREEN=GO reactions from car drivers; these guys were talking about mountain bikes, not commuter bikes. I'm going to have a look at how the whites I already bought do (they've got efficiency comparable to the Luxeon whites, and a slightly better color mix) but it seems that an amber light might be a much better choice. Amber (590nm) might win because: (1) lower input power, (2) better transmission through the eye itself, (3) YELLOW=CAUTION, (4) it seems to have little overlap with the rod wavelengths, but still overlaps the the "red" cone wavelengths. I can always recycle the whites for indoor lighting; it's pretty easy to wire a boatload of them in series to get to a line voltage, and the result is both efficient and dimmable.
Another useful thing about newer LEDs, especially these high-powered ones, is that that for as much as a 50% current reduction, they still produce light, and it is reduced only linearly -- that is, given a 60mA drive, 4 white LEDs and 3 white LEDs produce just about the same amount of light. This means that it should be possible to fine-tune the voltage of a parallel array, if necessary, and also to provide some headroom for power supply glitches. Under-driving a parallel array of lights also provides protection against runaway failure, which can occur if one LED fails and the current is concentrated onto the remaining LEDs. The amber/orange/red Luxeon III LEDs can be driven through a large range of currents; anywhere from .14A to 1.4A (continuous) produces light, and pulses of 2.2A can be tolerated.
Someone else has looked at this issue, though I'm not sure where they went with it.
Sunday - January 07, 2007
Car with Mass plates 4057 RO is a menace.
This afternoon, around 3:30, I was bike-riding
with my kids. To cross one road, we pushed the button at a signaled pedestrian
crossing (Concord Ave and Belmont High School), waited for the light to turn,
got our walk signal, and watched a woman, talking on a cell phone, Massachusetts
plates 4057 RO, blow through that red light without even slowing
down.
So, lady, imagine the things that I am thinking. There's words I wouldn't use in front of anyone. That's what I was thinking. I wish I'd had a rock to toss into the crosswalk, so as to crack your expensive windshield and maybe get your attention.
Nice lesson for my kids, though. People in cars can be stupid irresponsible shitheads, who would kill you because they're too busy to notice. When you learn to drive, don't be a shithead.
Fucking unbelievable, Ms. 4057 RO. I wish I could get your license yanked. My kids are more valuable than your time, you can take the bus. It'd give you time to think about how lucky you were that you didn't kill anyone with your careless driving.
So, lady, imagine the things that I am thinking. There's words I wouldn't use in front of anyone. That's what I was thinking. I wish I'd had a rock to toss into the crosswalk, so as to crack your expensive windshield and maybe get your attention.
Nice lesson for my kids, though. People in cars can be stupid irresponsible shitheads, who would kill you because they're too busy to notice. When you learn to drive, don't be a shithead.
Fucking unbelievable, Ms. 4057 RO. I wish I could get your license yanked. My kids are more valuable than your time, you can take the bus. It'd give you time to think about how lucky you were that you didn't kill anyone with your careless driving.
Friday - December 22, 2006
Ever wonder what happens when your hood latch comes undone at 70 miles per hour?
Your hood flaps back into your windshield, like a
giant steel lasagna noodle.
This happened about a year ago. I was in the
"slow" lane on 128, and WHAM! my hood is in my windshield. I decided that going
straight had worked ok for the last few seconds, maybe I would keep doing that,
edge into the breakdown lane, and slow down fast. I peeled the hood back, tied
it down, and drove slowly home. Didn't even wet my pants, but apparently I was
a little hyper when I got home.
The windshield was a total loss, but I managed to bend the hood mostly back into shape, and it was too expensive to replace it with a nice new one (maybe when I've spent $700 on a lot of more important things, I could do that, but really, it's just a car).
The windshield was a total loss, but I managed to bend the hood mostly back into shape, and it was too expensive to replace it with a nice new one (maybe when I've spent $700 on a lot of more important things, I could do that, but really, it's just a car).
Thursday - December 21, 2006
Unsafe storm sewer grates in Lexington, Mass.
On my bicycle commuting route there's some old-style
tire-grabbing storm grates, oriented in the worst possible way. I pinged the
Lexington selectmen, in hopes of getting the grates fixed, or, failing that, to
make it easier to sue, win, and collect the largest possible judgement when
someone is harmed by these gratuitous
hazards.
On the off chance that anyone reading this doesn't know what I'm talking about, here's what the bike tire does when you ride into one of these. Most likely, the tire will come to an instant stop, perhaps collapsing in the process, and you will continue forward into the handlebars and over the front of the bike. It's hard to imagine a way that this will not hurt, a lot. Broken wrists, smashed groin, and a face-plant in the pavement.
The first grate, not shown, is in front of the Mobil station at the corner of North Street and the Lowell Street (aka Middlesex Turnpike) on Lowell Street.
The next grate is in front of Locke Village, as you turn in from Lowell Street.
The third grate is a little further along, almost to the Burlington town line.
Here we are in the next millennium, and not only do we lack flying cars, world peace, and a cure for cancer, we still do the same stupid things we knew were wrong over 30 years ago. It doesn't even cost any extra money to do it right (in fact I think that the grate can be turned 90 degrees for a quick fix) but we still do it wrong. And remember, this is Lexington, with the highest per-capita concentration of Nobel Prize winners on the planet. Odds are pretty good that your town is stupider.
On the off chance that anyone reading this doesn't know what I'm talking about, here's what the bike tire does when you ride into one of these. Most likely, the tire will come to an instant stop, perhaps collapsing in the process, and you will continue forward into the handlebars and over the front of the bike. It's hard to imagine a way that this will not hurt, a lot. Broken wrists, smashed groin, and a face-plant in the pavement.
The first grate, not shown, is in front of the Mobil station at the corner of North Street and the Lowell Street (aka Middlesex Turnpike) on Lowell Street.
The next grate is in front of Locke Village, as you turn in from Lowell Street.
The third grate is a little further along, almost to the Burlington town line.
Here we are in the next millennium, and not only do we lack flying cars, world peace, and a cure for cancer, we still do the same stupid things we knew were wrong over 30 years ago. It doesn't even cost any extra money to do it right (in fact I think that the grate can be turned 90 degrees for a quick fix) but we still do it wrong. And remember, this is Lexington, with the highest per-capita concentration of Nobel Prize winners on the planet. Odds are pretty good that your town is stupider.
Sunday - November 12, 2006
Trying to design a bike light
Once upon a time, we were happy to power a 3 watt
generator; now, we make do with 80 milliwatt LEDs.
I've got some old leftover bicycle alternators
("generators") that can crank out about a half amp at 6 volts, alternating.
I've got an expensive LED bike light, but for all its brightness, it doesn't
give me enough light to see road hazards. LEDs are only about 60-80 milliwatts,
depending on color. Seems to me that with an old generator, a little EE work,
and a lot of LEDs, I could put together a vastly brighter bike
light.
I've been reading a lot about LEDs, power supplies, components, and the response of the human eye to light. For my purposes, the following things seem to be true about LEDs:
I've been reading a lot about LEDs, power supplies, components, and the response of the human eye to light. For my purposes, the following things seem to be true about LEDs:
• blue, or blueish, light is not the best
choice. This is because we don't focus well on blue light.
• red light is not the best choice. This
is because it is intended to be a front light, and because (while commuting)
streetlights and oncoming traffic will ensure that my night vision is not well
adapted, so I'll get no benefit from the preserved color vision.
• white LEDs vary, a lot. Some of them
have a spectrum with blue spikes in it, some do not. I am very interested in
the 3000K color temperature "warm white" LEDs from
LEDTronics , because they have only a tiny blue spike, peak at 565nm,
and deliver the bulk of their light as "yellow".
• white LEDs have one disadvantage; they do
fade a bit after 1000 hours of use. However, in practice it appears that I
would only want to ride in the dark four months per year, or 126 days, or 18
weeks, two hours per week. That's 36 hours per year, which gives a more than
adequate efficient bulb life (the bulbs also last longer when run in the cold,
which correlates nicely with the darker months).
• green LEDs are another choice. For
reasons that are unclear, to me, some of these LEDs are incredibly efficient,
and also put all their energy into spectrum that we can easily see and focus on.
For example, 25000 mcd @ 3V , versus 14000 @3.2V (both with a 16-degree beam).
However, I am a little worried about confusing drivers with a green light, and
the green LEDs are 40% more
expensive.
I am stuck on the power supply issue. It seems to be a standard that the lights should come on at 5.5mph. I need to calibrate the power output of my little generators to see how much they produce at that speed. My old generators are clearly unregulated; I cooked a few headlights coming down hills years ago. LEDs are much less tolerant of over-voltage. So I need a regulator. There are several designs that I am considering.
One is a straight current-source regulator, out of EE textbooks, probably built using a power MOSFET so that I can make the turn-on voltage as low as possible. This has the disadvantage of pulling a constant current from the generator, no matter what the voltage. At high speeds, I would not only be dumping some number of watts into the MOSFET, I would be taxing my own self to do it.
A second choice is a switching power supply. The National Semiconductor LM3478 looks like a good choice, and they have lots of design notes at their site. I'm particularly interested in the SEPIC designs, because they can take a high or low input voltage and regulate it to a constant output. There are four sources of power loss here. One is in the switching diode, which will consume 0.5 volts. The second is in the current sense resistor, which will probably consume 0.1 volts. Switching losses in the MOSFET are at high frequencies. Resistive losses in the power supply during the charging duty cycle also contribute.
A third choice is a cheesy variant on a switching power supply. This would simply use the input voltage to charge a capacitor, and use a MOSFET to gate current into the capacitor, based on current feedback. The inefficiencies here are resistive losses in the MOSFET, resistive losses in the current sense resistor, and resistive losses in the MOSFET during switching.
A fourth choice is not to conserve excess power, but instead to dump it into additional lights. When riding fast, I want as much light as possible, and that could well be more than the light supplied at lower speeds. One easy way to do this to set up multiple simple-switching power supplies for larger and larger stacks of LEDs, run in parallel, and to gate the MOSFETS from the overall input voltage.
One thing I did learn -- Schottky diodes are probably my friend. I can save 0.2 V, or 28%, on my forward voltage drop over silicon (if I could find a power Ge diode, that might be even better).
The fourth choice is looking pretty good, except that my transistor design skills are stale, and my FET design skills are really stale.
I am stuck on the power supply issue. It seems to be a standard that the lights should come on at 5.5mph. I need to calibrate the power output of my little generators to see how much they produce at that speed. My old generators are clearly unregulated; I cooked a few headlights coming down hills years ago. LEDs are much less tolerant of over-voltage. So I need a regulator. There are several designs that I am considering.
One is a straight current-source regulator, out of EE textbooks, probably built using a power MOSFET so that I can make the turn-on voltage as low as possible. This has the disadvantage of pulling a constant current from the generator, no matter what the voltage. At high speeds, I would not only be dumping some number of watts into the MOSFET, I would be taxing my own self to do it.
A second choice is a switching power supply. The National Semiconductor LM3478 looks like a good choice, and they have lots of design notes at their site. I'm particularly interested in the SEPIC designs, because they can take a high or low input voltage and regulate it to a constant output. There are four sources of power loss here. One is in the switching diode, which will consume 0.5 volts. The second is in the current sense resistor, which will probably consume 0.1 volts. Switching losses in the MOSFET are at high frequencies. Resistive losses in the power supply during the charging duty cycle also contribute.
A third choice is a cheesy variant on a switching power supply. This would simply use the input voltage to charge a capacitor, and use a MOSFET to gate current into the capacitor, based on current feedback. The inefficiencies here are resistive losses in the MOSFET, resistive losses in the current sense resistor, and resistive losses in the MOSFET during switching.
A fourth choice is not to conserve excess power, but instead to dump it into additional lights. When riding fast, I want as much light as possible, and that could well be more than the light supplied at lower speeds. One easy way to do this to set up multiple simple-switching power supplies for larger and larger stacks of LEDs, run in parallel, and to gate the MOSFETS from the overall input voltage.
One thing I did learn -- Schottky diodes are probably my friend. I can save 0.2 V, or 28%, on my forward voltage drop over silicon (if I could find a power Ge diode, that might be even better).
The fourth choice is looking pretty good, except that my transistor design skills are stale, and my FET design skills are really stale.