Spoke tensiometer

A tensiometer is simply an instrument for measuring the magnitude of some physical tension. In our case we specifically want to quantify the spoke tension in bicycle wheels.

For the purpose of building strong and durable wheels such a tool is by no means necessary. A good "feel" for what constitutes the optimal, or at least adequate, tension will be aquired with experience. And for the less experienced, direct comparison with a suitable example wheel helps in determining when a sufficient tension is reached. Nevertheless, a tension meter can be a very useful tool, as it more or less eliminates the guess-work and simplifies quality control.

There are different types of spoke tensiometers. Some are shaped symmetrically, either scissors-like or with a central piston, others are implemented in a geometrically more complex manner. The more expensive ones use precision dials while the simpler have cruder scales. The basic principle of operation is common to all, though: In order to calculate the actual axial force in the tensioned spoke, its resistance to a deflection caused by a controlled and perpendicularly applied (normal) external force is measured. The result is then checked against a look-up table which translates the input values to corresponding tension values, based on previously calibrated measurements.

I had a spare digital linear measurement tool left over from the truing stand project, and figured it would be easy to adapt it for use as the indicator part in a tensiometer. As I also had a collection of scrap aluminium extrusions, I decided to build a tool using those parts.

For an instrument like this, which must consistently measure very small deflection deltas, it is of the greatest importance that it is built as stiff and invariable as practically possible. So there is no point in trying to reduce weight and materials just to make it more elegant and, possibly, more ergonomic. On the contrary, the more robust the tool the more reliable the results!

The build

There is not much to tell about the build process. Only basic metalworking tools like a hacksaw, files, pop-rivet gun, threading taps and a pillar drill were required.

Operation

The mechanical action is defined by a spring-loaded, compression resisting central piston, which pushes on the spoke. The spoke reaction, in turn, is concentrated on a pair of "anvils" which have ballbearings at their fulcrum points too minimize friction. Attached to the rear end of the piston rod is a small platform with a low-friction plastic surface. The digital indicator's slide rests against this via a ball point and register the displacement. This minimizes friction in the contact point and allows the joint to tolerate small axial misalignments. In order to apply the full deflective load onto a spoke the hinged front part (with its spoke supporting ballbearings) is pulled back with the forefinger as far as it goes.

My tensiometer is designed to be operated by the right hand, which means that the digital indicator's display must be read upside down. Annoying, but not a big problem.

Feature creep... So hard to resist!

The inverted LCD solution soon felt too clunky. Seeing the linear indicator's unconnected digital output port, it was obvious that a more legible backlit display could easily be attached, even though it would require a separate micro controller to work. But then with a capable µC in place it should be trivial to perform the deflection-to-tension calculations in real-time, as it were, instead of forcing the operator to refer to a separate conversion chart. And hence we arrive at version 2 of our tensiometer (below).

To avoid having to deal with batteries it is now powered via a USB cable. Data is only transferred when a new spoke definition file is uploaded.

On the left of the display are the characteristic measures of the type of spoke loaded in the definition file, for reference. The threaded end is assumed to always be 2.0 mm.

On the right is the raw displacement value, which, as such and without calibration, has no particular significance. (This is the same value shown on the upsidedown display.) Below that is the smoothed output converted to force, expressed as decanewtons.

Another thing that turned out to be unsatisfactory in version 1 was the choice of surface material at the contact point between piston and caliper slide. The delrin type of plastic used there was too soft to resist being indented by the concentrated pressure from the ball point. It developed tiny surface pits, which meant it could not always move sideways freely. This part is now replaced by a polished, very flat steel disc (cut out of a 0.4 mm thick springy steel putty knife blade) and backed by a reinforced phenolic sheet base.

The schematic

Calibration

To begin using the tool in a meaningful way you need to know what the display readings actually stand for, in terms of tension force. A manufacturer supplied calibration chart would be really handy now, wouldn't it! Sooner or later, it will also be necessary to recalibrate the instrument, for instance if it has been dropped to the floor, or just suspected to have lost accuracy. However, in order to compile your own tables of values you obviously first need to have access to a range of reference spokes of different gauges under varying, verified tensions.

To that end a simple test jig was made. It consists of a 300 kg capacity hanging scale, a spoke and nipple connected to two aluminium tangs (representing the hub flange and rim respectively), and finally a large hook for attaching various dead weights. By hanging a short swing here I use my own body mass for a start, and then incrementally increase the load using combinations of heavy objects like dumbbells, 25 liter water cans, etc. It is then quite easy to read the total load off the scale's display while simultaneously operating the tensiometer on the stressed spoke. It turns out that in the load range that we're interested in (say 70 -140 kg or 700-1400 N) each 0.01 mm of spoke displacement corresponds very approximately to a tension change of 1 kgf (10 N).