TrackPoint a.k.a pointing stick 2-axis force sensor

Question

I have an application that would benefit from a sensitive two-axis force or bending torque measurement very much like what a pointing stick (a.k.a. TrackPoint) used to do in a ThinkPad, apparently using resistive strain gauges. I have been unable to find any commercial sensor with similar functionality at any price; all online references seem to be to salvaged pointing sticks from old laptops and ThinkPads, and the interface is not ideal on these anyway.

My application is not for human/machine I/O but rather to measure the horizontal force (in two dimensions) acting on an object at the top of a ~30 cm tall vertical rigid rod (the rod would transmit the force as a torque acting on the sensor at the bottom). I would consider constructing the sensor myself from two pairs of piezoresistive strain gages, but there too I have been unable identify parts with the appropriate size, low cost, and sensitivity.

Since pointing sticks were mass-produced at low cost for many years, I have to imagine that force sensors of comparable size and cost are still being manufactured somewhere.

Any leads or advice would be greatly appreciate

Answer 1

I'm collecting info about the history of trackpoint to avoid re-inventing wheels (and put into web resource). It appears that some mil-spec and industrial applications have put the tech to use, here's a group that may be available; http://www.ultra-msi.com/miniature-joystick-models.html#force. I may have useful info for you in future. I've worked with strain gauges and MAY reconstruct the crime (after understanding the patent trail better).

Answer 2

piezoresistive strain gages

Foil gages (metal pattern on plastic) may be even easier to use.

I have been unable identify parts with the appropriate [...] sensitivity

Strain gage sensitivity doesn't vary all that much within a given strain gage material, e.g. all uniaxial normal load (tension/compression) strain gages made of a given resistive material have similar ballpark strain to resistance change sensitivity.

I have been unable identify parts with the appropriate [...] size

Foil gages can be fairly microscopic, so no problem there.

I have been unable identify parts with the appropriate [...] cost

Their cost is nothing very much compared to the cost of labor of installing them on a transducer body. And if you've never done it before, then your time as you learn strain gage application techniques (you have to do it, repeatedly, and measure results!) will make strain gage cost utterly irrelevant.

I have to imagine that force sensors of comparable size and cost are still being manufactured somewhere.

Unit cost goes approximately inversely to volume. So if there's no ThinkPad-sized market, it won't be nowhere near as cheap, simple as that.

Any competent load cell manufacturer can make tiny bi-axial load cells for you, for a price.

rather to measure the horizontal force (in two dimensions) acting on an object at the top of a ~30 cm tall vertical rigid rod. The rod would transmit the force as a torque acting on the sensor at the bottom.

Glue four foil strain gages 0,90,180,270 degrees around a circumference of the rod. The thinner the rod, the harder it is to glue them accurately while the sensitivity is higher.

To calibrate, apply biaxial loads of various directions and magnitudes. Measure each. The calibration matrix that goes from 4 raw ADC signals to 2 calibrated loads will be a 4x2 matrix. Collect raw signals (time averages) for every applied load condition into a matrix, take pseudoinverse of that, multiply by the matrix of calibrated loads you've applied (in N, for example), there's your calibration matrix. Output will be calibrated in N force, the stick length will be "calibrated out".