Fixie Brake Lights

[Phil]

Hey Sprout,

I'm working with a friend on a small project to setup dynamic brake
lights on my fixed gear bicycle. After visiting Sprout a couple times
this summer I got the impression that it would be the right place to
get some input on this project. A quick outline can be found on this
page: http://pamolloy.dyndns.org/project/brake/

I would love some input on speed sensors. As I explained on the
aforementioned page, I would like to mount a gear-tooth speed sensor
(e.g. Cherry GS102301) to a seat stay.

Is there another type of speed sensor I should consider using? Does
anyone have experience with a particular manufacturer or model of
speed sensor? Are there other forums and/or resources on the Internets
that you might recommend?

Otherwise, I would love any other input on the project. We already
have a couple TI boards to test with, but we haven't decided on what
board to use in the end product, so any opinions on boards would be
greatly appreciated.

I have some programming experience and my friend is knowledgeable
about hardware/electronics, but neither of us has done a project like
this before, so the more help the better!

[Brandon Stafford]

Hi Phil,

I'd be happy to talk over your bike light project, maybe at Project Night on Thursday? There are usually a bunch of electronics zealots, and a few bike zealots as well, at Project Night.

It's officially 8-10 PM, but people show up early and stay late.

One quick thought-- I'd say the gear tooth sensor is a decent idea. I think its main weakness is that it will update slowly at low speeds. It might be good enough, though.

You might also consider a small circuit board with an accelerometer, maybe like this one: http://www.sparkfun.com/products/9267

Brandon

[Ben Polito]

Gear tooth could work if you just wanted to demo it - but if you had
any notion of making a product out of it, the setup for the geartooth
thing would be way too fussy for your average customer - worth trying
to get an accelerometer to work. e.g.
http://search.digikey.com/us/en/products/ADXL103CE/ADXL103CE-ND/643236
is the type I have used before - I wasn't actually putting it in a
circuit; I just soldered tiny wires to the appropriate leads and glued
it on a machine to suss out vibrations. Not cheap at ~$10 in
quantity, but given that people will pay $200 for a headlight...

Trick is going to be filtering the signal to get rid of the z-axis
jouncing that gets coupled in, while still being decently responsive
to the relatively slow sub-1g signal you're looking for.

> --
> This is a message from sprout-discuss: a high-traffic discussion mailing
> list for http://www.thesprouts.org . You can add yourself or unsubscribe at
> http://groups.google.com/group/sprout-discuss?hl=en
>

[Steve Grandpre]

Have you considered just attaching a switch to the brake levers?  

[Michael Shonle]

I think the point is that there is no brake lever for the rear wheel, which is frequently the only thing used to slow/stop, and for that really clean look, some even remove the front brake entirely.

I think the accelerometer is a great idea, just add a Kalman filter and you're all set.
I see there's some lower-cost options as well, e.g. $1.41 qty 1 for this, if it will work:

http://www.mouser.com/ProductDetail/Freescale-Semiconductor/MMA7660FCR1/?qs=sGAEpiMZZMu5vlrqIFXt5Z8SMEL0Irt9T39wQoNyxP8%3d

-Michael

[Ben Polito]

some folks feel that the use of brakes enables cyclists to take
unsporting advantage of the roads a la Reinhold Messner:

http://books.google.com/books?id=E4Tf5tL407EC&pg=PA60&lpg=PA60&dq=messner+oxygen+unsporting&source=bl&ots=qcmuVspdTP&sig=8fXOytkQ4AMZQ9reMHXEsrk9NHA&hl=en&ei=PGDBToj9I6GX0QH0yaDNBA&sa=X&oi=book_result&ct=result&resnum=5&ved=0CDwQ6AEwBA#v=onepage&q&f=false

On Mon, Nov 14, 2011 at 1:25 PM, Steve Grandpre <steveg...@gmail.com> wrote:

[Steve Grandpre]

Right, I forgot it was a fixie for a minute.  Accelerometer sounds like the way to go then.    

[Chris Connors]

From what I understand, most bike computers use hall sensors to
determine wheel speed. There is a magnet mounted on one of the spokes,
and the sensor mounts on the frame or the fork. A wire then goes up to
the circuit. You might be able to repurpose a bike computer, but
probably not easily.

If you use a microcontroller to identify the timing of the RPMs,
compare each and sense when the wheel is going slower each revolution,
then you should be able to send a signal to light up the LED array.

Chris

[Phil]

Thanks for the input Brandon. If it isn't raining too heavily I'll
bike over to Sprout for the Project Night tomorrow. Using an
accelerometer seems like an excellent solution.

As you mentioned, the sample interval is very important, particularly
at slow speeds. I have a bicycle computer that I believe uses a hall-
effect sensor, as Chris mentioned. In addition to being difficult to
repurpose, the reason why I initially wanted to use a gear-tooth
sensor was the fact that it would allow me to decrease the sample
interval. In my original plan the gear-tooth speed sensor would have a
sample interval of about 0.5 seconds when the bike is traveling at 1km/
h, unlike a hall-effect sensor, which would have a nearly 8 second
interval at the same speed. I would need 16 magnets evenly spaced on
the spokes of the wheel to produce the same effect with a hall-effect
sensor. Of course, an accelerometer would solve this problem. The
accelerometer that Michael suggested can be set at one of eight
different sample rates, up to 120 samples per second.

Anyway, although I'm not interested in creating a product, as Ben
mentioned, I do like the idea of using a accelerometer. However that
brings up a number of other questions, like whether to purchase an
accelerometer already attached to a board, like the one you suggested,
or to buy one independently, like those Ben and Michael suggested.
Anyway, that is just one of many questions I have. Not to mention that
I'd love to understand the physics behind proper acceleration better.

So I'll try to make it to Sprout tomorrow around 8:00pm, and if anyone
would like to stop by to help out I'd really appreciate it.

Thanks everyone!
~ Phil

On Nov 13, 11:49 pm, Brandon Stafford <brandon.staff...@gmail.com>
wrote:

[Brandon Stafford]

Hi Phil,

For experimenting with an accelerometer, a breakout board is the way to go. Otherwise, you have to add your soldering to the list of possible problems when you're debugging.

Here's a slightly cheaper accelerometer that looks decent: http://www.sparkfun.com/products/9652

Also, this tutorial: http://www.sparkfun.com/tutorials/167

I have some parts I want to print on the 3D printer, so I'll try to show up before 8 and get those started, but who knows.

Brandon