[mercuryrising]

This thread needs more hardware.

I'm making bike lights. First a little story. I was late to class one day in March, pothole season. I was cruising in the drops, when I came up to a T intersection. A van pulls up, and stops at the stop sign. As I approach the intersection, I see a pothole, swerve to avoid the hole, and the van pulls forward. I have about 1 second before I hit the van, land with my back on the hood, slide onto the ground. I'm lying there for a moment, trying to figure out what just happened, wondering whether or not anything is broken. I wiggle my fingers, wiggle my toes, don't feel any pain. I stand up, the guy gets out of his car "Sorry man! I thought you were turning!" I say I think I'm fine... I go look at my bike - it's still upright, the front tire got wedged in a rust spot. I grab it out, hop on, and ride. I couldn't help but laugh the rest of the way to class.

That's the day I decided that bikers and automobiles need better communication. So I made a bike light that's easy to use, has front & back blinkers, brake lights, turns on when you ride (so you can't forget), a bright front light, USB rechargeable, 3400 mAh LiIon battery. I made a 3D printer to print the parts, I did the PCB design, physical design, and software. It's pretty damn cool. I ride pretty regularly and recharge it about once every three weeks. It's surprisingly water resistance - I am a little scared of selling this to people as I didn't know how it would handle water, but I've taken it through two large storms, one with huge puddles splashing all over the cases, and the light handled it all in stride.

Here's the images - http://imgur.com/a/EUzXm

I'm stalling on it a little bit right now - there's not a very good way of bootstrapping into manufacturing. I could 3D print the cases and assemble the boards, but each case takes about 20 hours to print (on my fiddly printer). There's ~100 components too, which is a pain for manual placing and reflowing. I have a couple other designs brewing that are simpler and lower cost, but without all the fancy features. Speaking of which... I have to go get some interrupts working.

[DigitalJack]

It is hard to tell from the photo, but it looks like the the turn signal indication is the standard "blink on the side" style. When the lights are close together like they are with this design (as is practical for bikes), it's very hard to distinguish which side is blinking from even a short distance.

Could I suggest that instead of blinking on one side or the other, you turn the lights on in sequence heading toward the direction of the turn? For example, for a turn, you could illuminate light #1, then #1 and #2, then #1 and #2 and #3, then off and repeat.

That seems to me a much clearer indication of intention from the point of view of a vehicle some distance away looking through less than clean glass.

[mercuryrising]

Definitely possible, thanks for the suggestion. I'll consider it and do a visibility test.

There's 5 LEDs on the back - three reds and two amber turn signals. When blinking, the center red LED stays solid to provide a reference for the turn signal.

[sycren]

Why not try and make a kickstarter campaign?

- You have a working prototype. - You have a problem where there is a real need for a solution. - People on this thread are interested.

Now you just need to interview a few people who cycle around, show off your gizmo and make it part of the video on your kickstarter page ;)

[mercuryrising]

I've thought about Kickstarting a bit. I guess my hesitation is that I'm terrified of pulling the trigger to get something injection molded. I don't trust my 3D modeling abilities enough to know that when I get something injection molded it will work. And injection molding is a lot money, so if I mess it up, I might not have enough money for a redo. That scares me, a lot. I'm not sure, I guess I'm scared of actually making something people would use - I want it to be perfect, and I know it isn't yet.

[sycren]

Have you considered asking for help?

I'm willing to bet that there might be other hardware startups in your area, on HN or in your university that would be willing to give you a hand.

[mercuryrising]

I actually went to the local university - they said that you can't use university resources for personal gain (I was trying to print a case on their 3D printer). I'm sure there would be some people in the engineering department willing to give me some pointers, there's also a local prototyping company (I've emailed them a few times). Nothing too serious though - I guess I don't really have much of a plan yet, and didn't want to rush into getting roped into sticky situations.

Thanks for asking the questions - I just reached out to a friend who said he'd help me before, maybe this thing will start rolling :).

[sycren]

Where are you based?

[mercuryrising]

Minnesota, USA.

[kordless]

Maybe render it so you can do a nice photo. Wouldn't be anything wrong if you put up a video of the prototype and you telling your story...

[Dwolb]

Hey I'm working on the same thing. Right now I only have two rear blinkers hooked up to an arduino and a couple of sensors but I'm slowly building and planning more functionality.

It's a bit sad the lack of attention the rear of the bike receives with current product marketing.

Let me know if you want to develop together (totally understand this is a solo project and you're much further ahead!)

[karussell]

You should have a look into these devices here:

http://www.engadget.com/2013/05/20/helios-handlebars/

http://www.engadget.com/2013/09/12/hammerhead-navigation/

[pedalpete]

Can you explain a bit more about how this is improving communication between bikes and cars? Auto-on is nice, but we have lights that last a long time already. You mention brake lights. Are you actually riging this up to the bikes breaks? How about turn signalling? Stuff like that?

Clearly, it appears you've hit a thread here, maybe you can share more.

[mercuryrising]

So auto on / off is engaged with a reed switch and spoke mounted magnet. Auto on is more of a user convenience than a battery saver - you never need to think about turning your back lights on again. You will be seen when you ride. The reed switch doubles as a way of determining velocity. When the wheel velocity changes by a certain set amount, the brake lights will engage (three red backlights flash solid, then stay solid). It waits for either an increase in velocity or a time duration to turn the running lights back on. Turning is signaled with front mounted switches (teal switches in the pictures). One on either side of the front light. Currently, they turn on for about 5 seconds then shut off. There's a couple improvements that could be done with it (if stopped, keep blinking until a speed increase), but I haven't done that yet. When blinking is engaged the back lights stop running, the center light stays solid, and the amber LEDs blink at a frequency similar to that of cars. The front light has three modes - on, solid, flashing. Holding down one turn switch will engage the front light, holding down the other turn switch will cycle modes (flashing, solid).

Originally - I wanted it to be accelerometer based, but as you may know, the roads are not very smooth, so any bumps will affect the measurements significantly. The reed switch offers silky smooth data - as the data gets integrated over one wheel rotation, it's extremely noise resistant.

Ah and your main point - communication between cars and cyclists - turn signals allow cars to know where I'm going, the brakes will let them know when I'm slowing, and the automatic lights will make sure they can see me. Be seen, be safe, be smart. It's not a perfect solution, but I think it's a step in the right direction. There's a lot of aggression between cyclists and drivers, this might be something that can allow the driver to anticipate the actions of the cyclist.

[pedalpete]

Very cool. I figured you'd be integrating those sorts of things.

[graue]

Cool! How do you plan to deal with theft? Most bike lights are trivial to steal. I tend to avoid the fancy ones since in my experience lights disappear immediately as soon as the bike's parked (unless I remove them and put them in a bag every time I park, but that's a pain). Curious how you're addressing that.

[mercuryrising]

That's something I only put a little bit of thought into, as I'm not very worried about my bike getting stolen.

It would be non-functional if someone stole it without stealing the magnet and reed switch as well, but they wouldn't know that.

There's a bit of a trade off, as you're going to want to remove the light to charge it up, in that if you want it to remove easy, it will be easy to steal, and if you want it to be hard to steal, it will be a pain to charge. I don't have any solutions for that yet.

[theon144]

http://i.imgur.com/KeX2Fb3h.jpg

That's one hell of a print. Maybe try lowering the temperature? What are you slicing it with?

[mercuryrising]

That wasn't intended to be the final one, but I made it and it worked. It was printed with support material that peeled off a little too strongly, exfoliating the back surface. The orientation is kind of tricky as there's a lip/groove interface between the front and back of the case which needs to be cosmetically clean to ensure a snug fit, so the only way to get that to work well is to print it with the back down, but there's a 'hood' over the ethernet connector that disallows printing directly onto the back. Slicing with slic3r.

[amwelles]

Whoa, this sounds really awesome! Are you keeping updates on your progress anywhere? I commute every day, and I'd love to have something like this.

[mercuryrising]

Not at the moment. I'll shoot you a PM or something if I get something together.

Or maybe you'll see it on kickstarter one day :)

[mynegation]

Absolutely amazing. Bikes desperately need proper signalling and lightning and I would go as far as to say it must be mandatory.

[readme]

The idea just hit you one day.

[michaelmior]

This is awesome! Hope you figure out a way to manufacture it :)