Back to Dave's Planet

My design goal is to build a tensiometer that can be mounted at the glider end of a winch launch and supply telemetry back to the winch operator at a minimal cost and within FCC regulations.

The first thing I noticed is that I don't need to measure the tension on the line directly, I can use a variation on a running line tensiometer. Actually it's a little simpler than a running line because the line is fixed at that end. I use a spring attached to the end of a lever to measure how far I can deflect a cable under tension. The lever is used to keep the size of the spring to a minimum and also magnify the deflection for easier measurement. The spring and lever would be designed together to achieve the desired swings in deflection at various tensions. Attach the spring end of the lever to a sliding resistor and use the value of that resistor to modulate the signal sent back to the winch. See Fig A.

At this point we need to communicate back to the winch. We can use a radio link for this but it would have to be reliable at distances of a mile. This would be a challenge to stay under FCC power restrictions. Communication only has to be one direction, and it only has to communicate the amount of tension in the line. A sufficiently powerful modulated laser diode could easiy be visible by a ground unit with a fixed cylindrical fresnel lens focused on a phototransistor. A fixed carrier frequency is matched to a filter at the receiver end to filter out the signal.

This dove tails nicely with the design requirements of the tensiometer since we can built the unit into a triangular "wind vane" and keep a laser diode facing in the correct direction of the winch.

On the winch we have a cylindrical fresnel lens focusing on a phototransistor, possibly using an infrared filter if we pair it with an infrared diode to protect it from focusing direct sunlight. The circuitry attached to the phototransistor could most easily control an LCD character display (with a super bright backlight). It would be simple enough for the operator to use a dial to set the desired tension and have various color LEDs light up to tell him if he is too high or too low, as well as displaying the exact tension on the LCD screen. I have source code to accomplish most of these tasks in a circuit built around a $2 Microchip brand microcontroller.

Getting technical

Let us arbitrarily assume a 4:1 second class lever (pivot at one end and opposition in the middle) and design from there.

Let us select an arbitrary deflection range of 1" at the spring, or 0.25" at the cable.

At zero tension F will be zero, we won't stretch the spring at all and the deflection D will be at maximum, let's arbitrarily design for this zero tension deflection D to be 0.5".

At 3000 lbs tension F will be at maximum and deflection D will be 0.25" less, 0.5" - 0.25" = 0.25". The angle theta will be 2.5278 degrees. The amount of force F necessary to deflect a 3000 lb tension cable 2.5278 degrees will be 2 * 3000 * Sin(2.5278) = 265 lbs. Dividing by 4 for the lever, this equates to 66 lbs at the spring and a deflection of 1" for the spring, so in this specific design we need a spring with constant K = 66 lbs/inch that can tollerate a 1" deflection.

The scale of this design will be fairly straight but not perfectly linear. A spring scale is a linear device, but the tension angles applied by the cable being measured are somewhat greater at lesser tensions, this will add a shallow curve to our scale which we will adjust for in the programming of the receiver. We will have slightly less movement per unit change in tension at the larger measurements.

A real world design

Now let's try to design a more realistic version, we are trying to fit it into the 3" high 6" wide package of Fig B. The most base length B of cable we can offer is 1.5". We will design for deflection at 3000 lbs to be 1/16" and for deflection at zero lbs to be 5/16". The green/gray part is a slider, green areas are 3/8" thick while gray are 1/8" thick, leaving 1/4" clearance for the cable and for the lever. The angle of deflection of 1.5" of cable deflected 1/16" is 2.38 degrees. The force necessary to cause this deflection is 2 * 3000 * sin(2.38) = 250 lbs. We are still able to fit a 4:1 lever into this design, and that makes the force at the spring 62 lbs when displaced by 1". In this design I'm envisioning the cover piece pivoting around the right hand pivot point where the spring is anchored, allowing the operator to thread the cable and access the battery compartment. To keep the slide potentiometer out of the way of the spring I moved it to the point where the lever throw is 1/2".


A fantastically sensitive laser receiver circuit suitable for daylight use has been developed by K3PGP. In addition, K3PGP has also developed the best and simplest laser diode driver circuit with modulation support that I have yet found. This laser link is so sensitive it was designed to detect indirect scattering of light in the atmosphere and can detect an infrare LED that's pointed striaght up into the night sky (not even pointed at the receiver) at an astounding 5 miles! Way to go! We will enhance this circuits daytime performance by shielding the receiver with bandpass optical filtering matched to the wavelength of our laser diode. I am really hoping to get away with wrapping a cylindrical fresnel lens around the receiver circuit and not using an active tracking system. Worst case is that I end up needing to track the glider along a single axis using two photo diodes, the main detector and a trailing detector - if the trailing circuit receives too much signal then we use a simple $20 hobby servo and a microcontroller to drive the tracking up till the main detector is getting all the signal, then stop until the signal moves some more. More complexity = more points of failure, simpler = more reliable, I hope we don't have to track the glider.

I'll post source code for the microcontrollers and additional circuit diagrams as they become available.

Later Enhancements

Using the laser diode offers one additional advantage over a radio link, the ground based unit can be designed to pinpoint the exact location of the signal. We can in fact come up with a very nice estimate for how fast the glider is moving by how fast the glider progresses. Measuring velocity at the glider end is problematic in the amount of instrumentation required, but measuring how quickly a modulated light source moves across the sky at a distance that can be estimated by the program in the PDA, we can have a gauge of how fast the glider is moving. It's not really that difficult to use a simple $20 hobby servo to track a light source moving along a single axis, it's mostly a programming problem.

Back to Dave's Planet