Are you breathing? Do you have a gadget that can be charged via a USB 
port? Well if you answered yes to both, then you are in luck. This 
instructable shows how to make a device that will charge your 
USB-capable devices while you do what you do best. Breathe. Using some 
parts scavenged from an old CD-ROM drive, a simple electronic circuit, 
and a few rubber bands you will soon be huffing and puffing your way to 
fully-charged pseudo-useful electronic gadget nirvana.
Introduction and Step 1
This project requires a wide range of "maker" skills, such as PCB board 
manufacture, dismantling of electronics, cutting and drilling plastics, 
mixing epoxy, designing a gear train, kludging together a bunch of 
parts, bending paper clips, and risking the well being of your ever so 
expensive phone, camera, or PDA.  All in all, good fun.
Since 
everyone will have a different collection of junk parts to build this 
from, I will just give you a detailed overview of how I went about it 
and you can apply these ramblings to your own project.  Which will 
consist loosely of four steps.
1. Scrounge up some suitable parts for the generator
2. Build the charger circuit
3. Assemble the generator, thorax coupler, and mechanical return
4. Connect the charger circuit, and test
Step 1:
I
 had about four old CDROM drives hanging around and took apart a few of 
them to see what cool parts were inside.  Turns out there are lots of 
cool motors, gears, and other parts inside that fully validate my 
insistence of keeping such crap laying around.  Seeing the gear trains 
inside these units used for opening the tray gave me the idea for this 
project.  The small low-torque, high-RPM motor is linked to the tray via
 a gear train that has a final ratio of about 20:1  Previously I had 
been using a parallel array of tiny pager motors to generate electricity
 from breathing (see below) but the linear travel from your chest 
expansion is not that great (around an inch) so in order to generate 
useful voltages you had to really huff and puff. 
Anyway, tear 
into those CDROM drives, which you can find at pretty much any garage 
sale, thrift store, or landfill.  The pic below shows the results.  Lots
 of potential projects in there.  For now, we are only interested in the
 plastic gears and the motors for opening the tray and/or moving the 
laser carriage. 
Look over the various gears and drives and try 
to visualize a way to add additional gears to increase the gear ratio, 
or how to add another motor in series.  You want to minimize the changes
 to the gear train.  Alternatively you can just scavenge all the gears 
and build your gearbox up from scratch.  
You are also going to 
need at least one motor with a small gear or pulley on it so that you 
can connect it to the gear train.  The motors in the CDROM drive are 
typically simple permanent magnet DC motors designed to run on 5V, 
except for the spindle motor, which you don't want to use anyway.
At
 this point you also want to think about what you are going to use for a
 strap to go around your chest.  An old belt, some webbing, an old 
shoelace, a name badge strap, or anything that will fit around you 
comfortably without any stretch to it.  You want all the expansion to 
take place in your linear generator.  Any stretch that occurs in your 
thorax coupler will be wasted energy.
Step 2 - Build the charger circuit
The charger circuit is pretty simple.  It consists of:
1. A diode bridge to turn the AC voltage from the generator into rectified DC.
2.
 A rechargeable battery to level out the voltage and hold excess 
generated power when nothing is hooked to the USB port.  You could use a
 big  capacitor too, but batteries offer a more predictable voltage 
level.
3. A boost converter to bring the low voltage up to 5VDC for USB charging
4. A USB plug.
I've drawn up the circuit in EAGLE, a program that I highly recommend. You can download it for free from 
cadsoft.de.
  The schematic and single layer board layout are attached.  The actual 
use of EAGLE and the board manufacture are beyond the scope of this 
instructable.  Many great instructables are out there to cover these 
topics.
The parts list for the charger circuit (quantities in bold):
1x   L6920 Adjustable output step up DC converter (1V minimum input, 
Datasheet here)
Digikey# 497-4593-1-ND
4x   1N4148 switching diodes (I used tiny SOD523 smds, but you can sub in what you have handy)
Digikey# 1N4148WTDICT-ND
2x  10uF ceramic or other low ESR capacitors (I used 1206 smds)
Digikey# 39901299-1-ND
2x  100k thin film resistors
Digikey# P100kFCT-ND
1x  10uH wirewound inductor
Digikey# 490-2519-1-ND
1x  USB female Type A smd connector
Digikey# AE9924-ND
Above you can see the schematic and board files, and jpegs of them as well.  
The tough part is making a good PCB in your kitchen that has traces 
small enough for the TSSOP package of the L6920.   As you can see in the
 pic, I made 4 boards at once since each is so small.  
The trick
 to putting it together is to start in the middle and move your way out,
 begin with the L6920, and add the SMD discretes as you go.  A pair of 
tweezers is essential, along with good eyes or a magnifying glass, 
bright light, and a steady hand.  Don't worry about getting too much 
solder in there, use your solder wick to clean up any accidents, and 
check your work with a multimeter after every step. Practice makes 
perfect.
Step 3 - Build the generator
Now you need to make the generator.  You should play around with the 
gears and motors until you get a satisfactory arrangement.  You will 
want to use a multimeter on the motor while turning the gears to see how
 much voltage you are getting.  You want to get in the 2-3 volt range 
while moving the linear gear slowly about an inch in travel.  When 
setting up the gears, you want to use the ones that have a large gear 
molded with a smaller gear.  Stacked in series these will give you a 
good gear ratio as shown in the drawing.  (ignore the fact that the 
teeth are the wrong size in the drawing, I was too lazy to redraw with 
matching tooth pitch)  You should shoot for somewhere in the 25-50:1 
range.  More is better but eventually the losses in the gear train pile 
up and it get too hard to turn the motor and the gears will strip.
One of the keys is to find a way to use the linear gears on the CD 
tray or other piece to turn your breathing motion into rotation of the 
DC motor.  I included a pic of another prototype version of the CD drive
 generator where you can see the linear tray gear clearly.  Also visible
 are the cut marks in the plastic.  This prototype was also capable of 
lighting the LED array pictured.  Don't be afraid to chop this thing up 
to suit your needs. 
In the other pic the DC motor is mounted in place in the plastic of 
the drive I cannibalized.  Near this was a linear slider that I used to 
couple the breathing motion to the gear train. I also added another gear
 (see pic) to the drive train in order to increase the ratio and to 
allow mounting another motor in the future to increase output.  The main
 challenge is to effectively get the breathing effort translated into 
rotation of the motor efficiently. 
Step 4 - Put it all together and test it out
Once you have a satisfactory generator setup, then you want to connect 
the generator to the charge circuit, insert the battery, and use your 
multimeter to test the output voltage at the USB port.  If you don't see
 5V then there is a problem.  Fix it before plugging your pricey gadget 
into the USB port.
Below you can see my assembled breath powered 
USB generator in all its glory, top and bottom. You can see the rubber 
band used for return, along with the linear gear carriage, the strap and
 the paper clip I used to connect the linear gear to the strap.  The key
 here is to have all motion transferred to the linear gear so you want 
the strap and connection method to be stiff with no give.  The strength 
of the rubber band or spring return is up to you. My half-assed 
experiments indicate that you can pretty much handle a 1N force without 
feeling too labored in your breathing.  Ideally you want as small a 
rubber band as will return the linear gear to the starting position when
 you exhale.  If you get enough generating capacity either through high 
gear ratio, extra motors, or a bigger motor, then you will need a bigger
 spring return.  Essentially you are storing mechanical energy during 
your inhalation that is used to turn the generator on the exhalation so 
that you can generate on both push and pull.  You need the diode bridge 
to successfully take advantage.
So I strapped on this monstrosity and hooked it up to my trusty data acquisition box from 
DataQ.
  Attached is the voltage plot output of the generator before step-up 
conversion to the 5V USB.  Basically the battery runs the step up 
converter and the breath generator charges the battery.  In the plot you
 can see the leveling effect of the battery, with the voltage spikes 
when I was breathing.  Actually I was approaching hyperventilation, but 
in the name of science.  The results can be seen in the photo of the 
phone charging.  One thing to mention is that I had to modify a USB 
cable to get the RAZR to charge as 
detailed on this website.  I don't have any solid numbers on the power I was generating, I haven't come up with a good way to measure that yet.
Typical
 resting metabolism is on the order of 50-75W of which a substantial 
portion is due to breathing effort (I have seen north of 50%).  So if we
 assume 25W continuous energy used for breathing, it seems reasonable 
that we could increase that 4% to harvest 1W for charging a cell phone. 
 Based on my cell phone, and these assumptions it would take about 3 
hours to charge the 3.7V 800mAh battery.  Assuming 100% efficiency.
Sadly,
 based on the few measurements I was able to make, the breathing 
generator I built is putting out more like 50mW.  Way to breathe no 
breath.  It would charge the phone, but the NiMH battery would be doing 
most of the work until it was drained.  Then you would have to breath 
for a day or so to recharge the NiMH battery.  You were planning on 
doing it anyway right?  So there is room for improvement.  One area I am
 looking into is using carbon nanotubes and polyurethane to make an 
electroactive polymer generator.  This is the type of technology that is being used to make 
boot-strike generators for the military.
Future
 improvements could get this device into the 1W range.  Specifically, 
using a better DC motor (higher voltage per rev) and custom building the
 drivetrain to be more comfortable and better coupling to breathing 
motion.