Actions
Forum Thread Details
  • State Not Answered
  • Replies 8 replies
  • Subscribers 304 subscribers
  • Views 1483 views
  • Users 0 members are here
Related
See a helpful answer?

Be sure to click 'more' and select 'suggest as answer'!

If you're the thread creator, be sure to click 'more' then 'Verify as Answer'!

Electric Trombone: Choosing a suitable slide sensor

Former Member

I've had an idea in my head for a few years now - I'd like to build an electric trombone.

 

This is what I've got in mind:

http://i.imgur.com/YWB22.png

 

The above does look a little ambitious, but perhaps I might get there in the third or fifteenth prototype...

 

 

My main question would be: what is the best sensor to use for the slide?

 

 

The slide has to glide smoothly like a knife through butter, and have as little momentum as possible - so you can do fast passages. The "rod" it moves up and down on should be as light as possible, otherwise there will be quite the moment when both handles are together (a counter balance might be needed...  image).

 

If you hold your left hand to your lips, and position you right hand as far away as possible, then that should be the maximum distance the sensor should have to sense.

 

If the above wasn't limiting enough, then the sensor should also have a good sample rate (in the milliseconds?), and a good sample depth (in the millimeters or fractions of a millimeter?).

 

From time to time, I've looked around for a suitable sensor, but I've never yet found one I've liked.  The closest I've come accross would be to have a softpot with a wiper, or to use graycode on the slde - but I've never really felt satisfied with these solutions.

 

Does anybody know of any sensors that would be suitible for this?

 

Dave

 

Edit:

The MIDI specification for pitch wheels, is to use 16,384 possible variation values over 4 semi tones.  The trombone slide covers 6 semitones (with 7 positions) over aproximately 600mm (about 100mm per semitone).  So to use the same accuracy as a MIDI pitch wheel, the resolution would have to be 16384/4 = 4096 per 100mm, or 40.96 per millimeter.  Obviously, this is a bit of overkill!  I think the best sample deapth would be in the millimeters or fractions of a millimeter.

Parents Reply
  • 0 in reply to michaelkellett

    Ok, don't over think it.. simple, pure your picture is of the slide, but that of a "real" side, not an electric slide.

    simplest an cheapest route is to use capacitance image. Just like we measure fuel in a plane. But we see the wet area or lack of it, between a rod inside a pipe. what you need is a pipe and a rod and you move the rod in and out that will give you your change in capacitance. And all you need is the delta of the length of the rod. Then by picking the right ADC (ie, how many bits wide, and how fast) your about done. 

    image enjoy ~

Children
  • 0 in reply to phoenixcomm

    I'd be concerned that the human presence might affect capacitance.

     

    Another approach would be to use an inductive loop, like traffic signals.  Have an inner plastic tube with a loop of wire running the length of the tube, wrapped around the outside.  Then have an outer iron or other ferromagnetic tube that slides on the inner tube.  Inductance is approximately proportional to how much loop is covered with iron.

  • 0 in reply to johnbeetem

    The problem with simple capacitave sensors like Christina suggests or inductive like John suggests is that the total sensor length is at least double the working range - this may be OK because the same is true of a real trombone.

     

    The capacitave caliper technique I suggested is widely used but perhaps a bit complicated to build.

     

    You will find similar capacitave sensors documented in  various microprocessor manufacturers app notes about touch sensing but I'm not sure that the touch sense technolgy can be stetched to such a long device.

     

    Annother approach you could consider is an unrolled shaft encoder, machine or paint stripes along the tube and use an optical sensor (needs two sensors 1/2 pitch apart) to detect the movement of the slider over the stripes. The resolution is only one half of the stripe pitch and since it's a relative sensor you need to count stripes passed up and down to keep track of absolute position. The slider is active.

     

    My favorite idea so far is more complex but does allow absolute detection, a passive slider and a reasonable sensor length:

     

    Wind several overlapping coils along the length of the tube, single layer of turns for each coil but they probably need to overlap by 50% so you end up with two layers. The slider needs to be magnetically permeable (ferrite or iron dust) and probably 1 coil length long. It will work much better inside the tube but might just about work outside. You will need to energise the coils in turn to measure the inductance and will be able to estimate the position of the slider quite accurately.

     

    With any of the techniques suggested you will need quite  abit of signal processing and compensation to get a useable digital signal.

     

    MK