[sdiy] optical retrofit for minis?

[Richard Wentk]

At 17:38 22/03/2005, Kevin Lightner wrote:

>>I wouldn't do this. I don't see how laser modulation would tell you which 
>>key was pressed unless you had some kind of optomechanical scanner, and 
>>if you go down that route you might as well just use individual LEDs.
>
>I'm not planning on doing this, but if a string of pulses was modulated on 
>the laser, a keypress could indicate at which point, relative to the 
>frequency and count of the pulses, which key it is.
>A missing pulse set up.

No, still can't see how this would work in practice. If you're basically 
measuring left - right transit time, then you don't need pulses, just a 
*very* fast counter. If you're using pulses you need some way to synch them 
with 'keypresses' - as you would with a laser harp.

You could certainly build such a thing but it probably wouldn't be the most 
reliable of all possible solutions, and getting the scanner inside a 
keyboard isn't something I'd want to try.

>>There are also issues with diode life (5000 hours, or 10,000 if you're 
>>lucky) and also possible problems with using the system on a smoky stage.
>
>I've tried putting an incense right next to it and it still works fine.
>Can't imagine a club THAT smoky. The audience wouldn't find the stage ;-)

The club probably wouldn't, but the stage might. I agree it's unlikely to 
be a problem in practice, although this would depend on available power and 
how marginal the sensing levels were.

>But the life of the diode is a point well brought up.  Thanks.
>
>>
>>A more conventional multi-LED based system makes a lot more sense to me.
>
>Are there photointerruptors available with standard LEDs?
>Making them would be a real task.

There are some fairly standard proximity sensor or interrupter units 
available which could work. I don't think they're cheap though.

Another option would be to use electroluminescent wire as the light source 
'bus bars'. Then you'd only need one (or two) photodiodes for each key. 
With the right shutter arrangement, you could even abstract velocity 
information from the light curve on a single diode. If you sample at say 
1MHz then the last ten samples would easily tell you what you need to know.

>>How about bundling up a collection of projects - I'm sure you have more 
>>;) - and going in search of venture capital?
>
>I have no experience at obtaining investors.
>I know lots of rich folks, but it's not the same thing.
>Any advice would be appreciated though.

I'm hardly an expert, but as I understand it you're unlikely to get funding 
unless you go for the 7-8 figure slot. A one-off like this probably counts 
as too small.

Maybe you could put it to one of your rich folks directly. If they're a 
vintage fan and want to do this thing they might pay up front for the R&D 
and not be in too much of a hurry to collect a profit.

How much do you need, anyway? I'd guess $10k would probably be generous, 
and $20k would be extravagant.

To some people that's pocket money.

Richard