Track pressure position on a string

Question

I worked a little on electronics at school but it's a long time since then so I need some advices.

I want to track a finger pressure position on a string on the X axis (see the figure below). I think that I can deduce it from the difference of forces on the 2 sensors, but I'm not sure this will work.

So, is this method correct and what kind of sensor can do the job ? Or do you have any other method that'll work ?

Thanks.

Answer 1

From your drawing, it looks like the string deflects relatively little relative to its length. It also seems there may be accidental sideways friction force on the string from the finger. Both these together mean that you do not want to measure the X component of the force at the anchors.

There are two reasons why measuring the X force won't give you useful information:

1. It's a string, so the force along is axis is the same everywhere (assuming no finger friction for now). The X component of that force at each anchor will be the cosine of the angle at which the string leaves the anchor. We'll define zero angle to be when the string leaves horizontally, which is essentially the unloaded position. This angle is expected to be small, so the cosine will be nearly 1 for all cases. Not useful.
2. Accidental sideways force from the finger on the string will show up predominantly in the X forces at the anchors.

What you want to measure are the Y forces at each anchor. To calibrate, measure them with no load, then subtract that off from subsequent readings. All the vertical force on the string must ultimately be ballanced by the sum of the vertical forces holding up the anchors. Sideways friction forces only effect the vertical force by the sine of the angle, which is near 0.

Strain guages or capacitive scales are probably a reasonable way to measure the vertical force. The rest is a little circuitry to convert the signal appropriately for a microcontroller A/D input, and the rest is firmware.

Answer 2

No, that won't work. If the pressure is vertical, as drawn, then the X-components of the force on either side will be equal. If you draw a vector diagram of the forces this will be clear. If the vertical movement of the finger is limited then both change in stress in the string, and Y-components of the force will be too small to make a proper measurement (if you can measure the latter in the first place).

Answer 3

Description was not too bad but, as ever, a betterdescription of what you are really trying to do will help. eg is this a guitar string or ... ?

1. Olin's idea looks goodish.

2. Consider also (trying to) measure the deflection angles.

X1/X2 = Angle2/Angle1

Xx = distances along X axis. Angle = deflection angle of string from horizontal. ie actual angle values in radians or degrees.

If you connect the strings to stiff pivoted rods at the connection points or eg take strings through holes in pivoted capstans, you can measure either the angles directly or if you apply rotational restoring force proportional to rotation, the rotational restoring force can be measured.

IF the string is above a surface then a detection of the finger may be better. This could be optical or capacitive or ... .

Answer 4

One possibility is to measure the acoustic waves that travel down the string when it is touched. Clamp each end to a vibration sensor, like a piezo, and they will produce a pulse when the wave hits them. If the finger is exactly in the middle, the pulses will occur at the same time. If it's 1/3 of the way, one pulse will occur at time T, and the second will occur at time 2T. etc. I believe this is how some touch-screens work.

This will only detect taps or changes in where the finger is touching. For continuous touch, you can assume that the previous touch hasn't changed, since changing the position or removing the finger would (I think) generate new pulses.

Alternately, you could get fancy and do time-domain reflectometry on a conductive string. Send high-frequency pulses into the string from one end and measure the delay until you get a reflection from the finger. If you only get an echo from the other end of the string, then you know it's not being touched.

Answer 5

Yes, you should be able to measure the x-position based on the force on each side of the string. I have used strain gauge sensors to measure force. They are pretty cheap and relative easy to use. I would recommend placing the two strain gauges in a H-bridge and use a instrumentation amplifier to get the best result. (noise and temperature immunity)

Answer 6

The method is correct. The sensor mechanics must be designed to measure the tention force only exactly to current direction of the rope. For example with some recalculation for motion of real sensor, like fisherman weight scale, you can derive exact total length of rope including 2 deltas from 2 reading of springs.

The geometrical rules for X, Y vectors:

1. Y1 + Y2 = Total pressure
2. X1 -X2 = 0
3. F1 = \$\sqrt{X1^2 + Y1^2}\$
4. F2 = \$\sqrt{X2^2 + Y2^2}\$

etc.

The major mechanical requirement is to have sensors directionality to comply with vector. I.e. Sensor must read only tension on string, not the flex of the joint, or torque of string rotation.

The circuitry choice will be next step after you choose the sensor (completly impossible to choose before you decide what kind of sensor, cost, accuracy, noise immunity, reliability you need.