s4ndp4pper Posted October 22, 2003 Report Share Posted October 22, 2003 Hi! Here is a little something-something that i've been thinking of.I thought about making a breath controller for my upcoming midibox, and looked at Matteo Bosi's design at the MIDIbox Extensions page. I love his concept8), but it uses a pressure sensor that costs about 40 Euro in Sweden. And that was a little too much for a poor student like myself.However, i thought about making a breath controller that uses a light sensor instead, like this one: http://www.elfa.se/elfa/produkter/en/2016264.htmThe concept looks like this:The idea is that a valve (made of i.e. some thin, flexible, non-transparent plastic or metal film) shuts between two chambers (one with a LED and one with the sensor) and then bends open when air presses it, thereby letting the LED illuminate the sensor, in relation to how much air is blown into the device.The main advantage of this device is that the cost of the optical sensor is a fraction of the price of a pressure sensor. As i understand, the sensor also has it's own amplifying circutry.I havent built this or anything like it. I do not have much electronic knowledge, so i don't know about how the circuit diagram should look like (or if this solution is even possible at all???)What do you guys think?/J Quote Link to comment Share on other sites More sharing options...
Wilba Posted October 22, 2003 Report Share Posted October 22, 2003 Nice diagrams!I've experimented with infrared LEDs and photodiodes...In my experiments, I had them about 7 cm apart and was able to detect the blocking of the beam with a finger, with a fairly good range of output voltages over about 7mm of finger movement.You could use the sensor you suggested, but if you want full control over sensitivity etc. you really should build it using a photodiode and resistor.What worked easiest/best for me was using the photodiode as a basic current-to-voltage converter - the photodiode is reverse biased, anode to ground, cathode to a resistor to positive rail, voltage at cathode will then range between ground and positive rail (eg. 5v) which you can then plug into an AIN input. You'll need to tweak the resistor value to get the sensitivity you want. NB: voltage will be inversely proportional - ie. more light = closer to 0v, less light = closer to +5v. You could either invert the value in firmware or even redesign the layout - ie. put photodiode above the pipe, LED below the pipe, valve between them and the mouthpiece so the valve opens and blocks the ligh path. Just a thought...The other way of using a photodiode involves wiring it up with an opamp so that it gives a better linear response - ie. rather than outputing a voltage relative to ground/+5V, it outputs an absolute voltage relative to the amound of light. But I found that it wasn't worth the extra effort. If your +5v reference voltage isn't too stable, then the first way (a pseudo-potentiometer in effect) works better.I'm interested in building something similar... I'm not sure what to make the valve out of... if it's too flexible it might flutter like a flag in the wind.... which I guess could be a feature!Will post circuit diagrams soon - they're all scribbles on notepaper atm.Wilba Quote Link to comment Share on other sites More sharing options...
goyousalukis Posted October 22, 2003 Report Share Posted October 22, 2003 That's a great idea and diagram. I might try this also if I get a chance.Justin Quote Link to comment Share on other sites More sharing options...
Duggle Posted October 22, 2003 Report Share Posted October 22, 2003 Another idea is to have an LDR (light dependant resitor) connected to +5V with a (say) 10k pulling the other side of the LDR to 0V. The analog voltage is at the junction. Choose an LDR with a dark resistance of several Mohm and a light resitance of a few hundred ohm or (even a few kohm). Use a high intensity LED, and make the arrangement small so that good transfer is achieved.I imagine that the valve is a rubber flap against a round or slotted apperture that is completely closed when there is no air velocity. As air passes through the apperture the flow rate exerts a proportionally larger force, creating a larger opening allowing accordingly more light onto the sensor. The shape of the apperture should have an effect on the linearity of the output.cheers Quote Link to comment Share on other sites More sharing options...
s4ndp4pper Posted October 22, 2003 Author Report Share Posted October 22, 2003 Percisely my thought (the valve)! Also, the way the inside of the controller is painted can also play a difference.Thanks for the advices - i never thought about reversing the lightflow like Wilba suggested! I think i'll go for the LDR though.Btw, another aspect is saliva :P ;D and condense water from ones breath. The electronics must be sealed well, maybe behind transparent plastic or something./J Quote Link to comment Share on other sites More sharing options...
pay_c Posted October 22, 2003 Report Share Posted October 22, 2003 Hmmmm, as a nearly finished Physical Technician ( ;D ) I just have to mention that:That little blockade (thin plastic or so?) will definitely start to swing around like wild. It will not go to a certain point and stay there. You´ll have something like a trumpet. ;)Workaround: Use a lowpass filter (about 20 Hz or less) which will have a bad effect on the reaction time of that thing (although not to bad I think).But anyhow a good looking idea!!! :oGood luck! Quote Link to comment Share on other sites More sharing options...
s4ndp4pper Posted October 23, 2003 Author Report Share Posted October 23, 2003 Ok, thanks for mentioning. It would be a nice "physical" touch, though ;) Maybe one could make a rigid, square-shaped flap and attach one side of it with flexible material, so that the flap itself stays rigid? Quote Link to comment Share on other sites More sharing options...
arumblack Posted October 23, 2003 Report Share Posted October 23, 2003 or maybe some sort of spring behind a ridgid flap? Quote Link to comment Share on other sites More sharing options...
Duggle Posted October 23, 2003 Report Share Posted October 23, 2003 Try experiment with a capacitor of say (10.. 100uF) across the 10k pulldown. This will smooth any AC component of the vibrating flap. Will need to experiment with the values in the cap and pulldown if there is too long "attack" and "decay" introduced. I think without the cap there maybe a a lot of jitter on the midi CC output.Also consider that LDRs have various response times depending on their design. You may be able to use this fact to your advantage! Quote Link to comment Share on other sites More sharing options...
Wilba Posted October 23, 2003 Report Share Posted October 23, 2003 I'm using photodiodes instead of LDRs... they have fast response times and they are usually spectrally matched to infrared light, so you can use infrared LEDs and ambient light is less of an influence. You can effectively treat them as LDRs... more light = more current flow = effectively lower resistance. Don't know much more about the electrical differences.What I do know is: a reverse-biased photodiode gives high-speed response and a wide propertional range of output - ie. good sensitivity to light levels.The best advice I can give is: get one of those experimenter boards and experiment with different parts and circuits. The real trick will be getting the vane right... I think reducing the amount of movement between no blowing and full blowing, and then mounting the light source and sensor close together will give the best results... but that's just a guess...Some links that I found while researching MIDI controllers and optical sensors:http://tomscarff.tripod.com/The Theramidi uses two LDRs and a PIC, includes circuit diagram.http://www.soton.ac.uk/~rmc1/robotics/artactile.htmInteresting theory regarding tactile/force/pressure sensors, including using optical techniques.Wilba Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.