Criteria

• Now that we're starting to make engines that could (at least in theory) produce thrust, we should have a way to measure the thrust they produce.
• If (big if) we get something as good as the smaller engines you buy at the store, we'll be producing on the order of 10N initially. If (as is more likely) we don't even produce enough thrust to lift the weight of the engine at 1G that's about 20g/1000*9.8 ≈ 0.2N. So we'd like our meter to measure forces between 0 and 20N in 0.2N increments (at least near the low end).
• It also needs to be continuously readable from several meters away (or record multiple readings per second in some way).
• And the sample pan needs to be heat/flame proof and NOT a pipe bomb.
• Plus if possible we don't want to spend any money on it.

Basically what we want is a low cost, remote reading (or even better, recording) dynamic Newton meter.

Design ideas

With those constraints in mind, we considered a number of possible designs. At first we were trying to find a way to indicate the force clearly enough that we could read it from a safe distance (through binoculars?) or record the force electronically (digital accelerometers on an arduino?). Then we hit on the idea of transmitting the data to where we were...with a cat laser.

The idea is pretty simple (once you think of it) – make the force we're trying to measure slightly change the angle of the beam of the laser, and then measure how the spot moves.

We then considered various variations on this idea, such as bouncing the laser off a mirror attached along the bottom of a beam that would be deflected by the force to be measured, and so on.

What we ended up with was pretty simple: a long board, with one end firmly clamped to a sturdy base (an old book case), with the cat laser attached to the free end and pointing away from the base.

The beam of the laser was aimed at a wall in the shade where we taped up a number of meter sticks end to end. Small changes in the force pressing downwards on the board moved the red dot on the meter sticks a measurable amount. By placing known weights at a designated spot we were even able to calibrate it!

Testing & design review

After initial calibration we tested it by using it. To our surprise we got amazingly good data. Instead of having to stop frame through the video we were able to pull it out with a little python program.

The bad news was that it bounces pretty bad. And even worse, the oscillations are only a little higher frequency than the changes we're trying to measure–about 4Hz according to a music nerd.

We considered a number of possible solutions:

• Use a stiffer beam. This should make the oscillations smaller, but it would also make the readings smaller, meaning we'd have to amplify them more and bam! we're back where we started. We tried it anyway, but it didn't help.
• Use a lighter beam, so the oscillations are higher frequency and dampen faster.
• Use padding to dampen the oscillations. We tried this and it may have helped a little (unclear).
• Attach the rocket engine and cat laser to a wheel. Max B
• Put the engine facing downwards in a long upright tube with a slit for the cat laser (on the engine) to shine through. Or just use it to launch a rocket. This was ruled out for safety reasons.
• Put the engine on a digital scale, and take a movie of the reading. Except that the scale would be in the smoke and so we might not be able to read it, scales bounce too, we only have one good one at present (a chemical spill spoiled the other), and it's sort of cheating since the goal is to make a Newton meter, not find or buy one.
• Put a scale under the plank thing, that's attached to the ground and touching the plank. (Note, this idea won't help either)
• Measure the exhaust velocity (either with the anemometer or a camera) and calculate the thrust. After thinking for a moment we were pretty sure that the exhaust would melt the anemometer before we could get data. Also, we've no way to read it from far away, and it would be in the smoke cloud so using a camera is doubtful. It might work, except we have no good way of measuring the volume or density of the exhaust, and without that the raw speed doesn't help us much.
• Ignore the problem
• Keep the design as-is and "fix" the data. We had a number of concerns about this approach so we tested it. It gave better results than the skeptics had expected

It turned out that only one of these (data cleaning) stood a decent chance of working in the short run so we kept the basic design as is, with the following modifications:

• On sunny days, we may use shades to further darken the area where the cat laser will be read.
• We're going to add some padding to dampen the oscillations if we can demonstrate that it doesn't compromise the linearity.
• We've built a laser pointer pointer to simplify the alignment of the dynamic Newton meter with the scale.
• We're going to put registration marks on all the parts to increase reproducibility & simplify calibration.