Bike Speakers

This spring I went to my first East Bay Bike Party, and it was really neat. There were a thousand people riding bikes together at night, with lights and awesome stereos on their bikes, and people were wearing costumes. We were mostly law abiding, polite, quiet in residential areas- it was much better than Critical Mass! When I say it was better, I mean that it’s the kind of scene that a person could have some pride in talking about in most places

I saw the kinds of things that people were bringing to make music- they were using boom boxes, a few custom things, and portable DJ speakers on trailers. Some of these were powered by heavy batteries that were driving inverters that regular PA amps were plugged into. By the end of the night I heard stories about people having to be pushed up the hills to get their speakers to the next party spot.

I’ve long had a love for designing and building speakers, and I always keep my eyes out for new technology and ways to make things happen, so I got the idea that I could make a totally cool bike stereo that would be light and loud. Light, loud, and good sound! I hope that I can make it look good too, but my priority is on good sound, as it always is when I approach audio, and on weight. I don’t want it to be any heavier than if I’m riding home from Berkeley Bowl with a good pile of produce (and beer).

My brain started chewing on the design problem in the background as life went on, and from time to time I collected another puzzle piece. The first thing I found was that there are a lot of high efficiency, light weight, high quality “class D” amplifiers. These are the new generation of amplifiers that take advantage of the better (and cheaper) electronic chips that are available today, and make sound by switching on and off very quickly to create a pulse stream rather than modulating an analog voltage, and so their efficiency can be in the +90% range. This lets them sip power from a battery, so I don’t have to carry around such a heavy battery. They also have versions available that take DC battery input instead of AC input from a plug into the wall, so there will be no need to use an inverter.

That was the first step to make it possible; it cut the battery requirement at least in half. The next step that made it possible was the Aura NS6, a lightweight paper coned woofer with neodymium magnet, on sale at Parts Express. Even compared to professional audio drivers, I couldn’t find anything that would allow me to get more bass per pound- not unless I wanted to spend BIG money. These cost less than $8 each when I bought eight in one order, and I am certain that for $64 there’s nothing that will be able to play louder.

I’m loading them up into a pile of enclosures like this:

I used cardboard tube that’s normally sold as a form for concrete columns. It’s fairly strong and stiff, and it’s very light compared to most speaker enclosures. They may not seem very substantial, but the round shape is great at resisting pressure. The end caps are laser cut acrylic (gotta love Tech Shop!), which will be transparent once I remove the protective liners. The laser made it fast and easy, since instead of cutting round holes and doing precise drilling, I was able to place the sheet of plastic into the laser cutter and hit “go”. The other cool thing about the acrylic is that it will let me put glowing LEDs inside the tubes. This will make my bike more visible, and it will be fun!

If you know much about the physics of sound, you probably know that making a lot of bass from something small and light isn’t easy, and it’s not efficient. To get efficiency out of this arrangement, I am doing a few things. The final build will include eight bass modules, four on each side of the rear wheel, and it will be electronically limited not to play much below 50hz. I’ll be the first to admit that this isn’t very low, but I did an experiment back in college to see where most of the bass ‘lives’ for the kind of music you hear at clubs and in concerts. It turns out that the ‘pounding’ bass that you can feel in your body is largely between 80 and 160 hz. There’s a solid component of many bass drums where the initial impact is even higher! Yes, a bass guitar’s low string goes down to around 40 hz, and many styles of electronic music contain much much lower notes- but sacrifices had to be made.

These are light, not especially expensive, and although they’re not small, they’re not too big to fit on my rear cargo rack.

That covers the amplifier and the bass, two of the hard parts, but what about the tweeters and crossovers? Tweeters were easy- I got a pro midrange tweeter with a big horn that plays nice and low, and a smaller tweeter that plays as high as you can hear. No big deal, light cheap and sure to work. The really cool thing is the crossover- this is the component that made me look at the project and go from saying “This is too much work!” to “Yeah, I could totally do that!”

The MiniDSP is the most expensive part of the build, but it does the job of something that would have cost three times as much just five years ago, and barely existed ten years ago. In real time it takes the audio and digitizes it at high quality, and mathematically manipulates it to emulate the circuitry normally used to separate sounds into high, medium, and low pitches to send to each of the speakers. This is the kind of thing that could be done on a circuit board with op amps, and people have been doing that for a long time- but to get good sound you need to use a lot of expensive op amps, you have to do a lot of soldering, and you have to change the values of components based on measured and calculated values specific to your speakers. It’s hard to tweak, easy to break, still not cheap, and there are things you might want to do that are very difficult to achieve in an analog circuit. You can also buy one off the shelf, like the Marchand XM9, but that’s really expensive, and still not as flexible as the MiniDSP.

I used WinISD to simulate the woofers boxes, and with eight of them together I should get an efficiency of nearly 98 decibels per watt, and with the 4×100 watt amplifier I should be able to just produce just over 120 decibels. That’s probably louder than your stereo, and it’s more than loud enough to cause hearing damage and get the police involved.

More pictures coming as I build this thing!