Since we've established a method of melting the fuel that we believe is safe and it appears to greatly improve the burn rate we moved on to trying to determine the best way to add it to our process. Also, since we're getting very fast burning we would like to know if it's fast enough – will it produce thrust if packed in an appropriate combustion chamber (one with a choke nozzle and that is NOT a pipe bomb)?

For this test we produce two separate batches of the standard mix (12g KNO₃ + 6g powdered sugar).

We also removed the residue from the casings of two used (bought) rocket motors, and threaded 5 seconds of fuse into the nozzles.

The we set up the oscillating polisher to use as a vibrating table to help clear bubbles when we poured the fuel into the casings if needed.

For the first we dissolved the KNO₃ in 40ml of hot water, then added the sugar (which also dissolved). We let this boil (stirring constantly, etc.) until it formed a paste and then began to brown. Unfortunately, it was very clumpy and the outsides of the clumps were getting dark while the insides presumably remained moist. After a quit discussion we removed it from the heat since this behavior was enough different than our safety test that we weren't sure that we knew where we were going. We pressed the paste into the open end of the first engine, leaving room (~1cm) for a plug, and setting the excess aside. Vibration wasn't needed, since the paste was very doughy and we were able to just tamp it down gently. It did stiffen up / harden as it cooled.

For the second batch we used the spark free hand mill to reduce the KNO₃ grain size before mixing it with the powdered sugar. Then we combined them and melted them to a thin brown liquid. We poured this into the second casing, using the vibrating table to clear bubbles (we hope). We again left room (~1cm) for a plug and set the excess aside. This time the mixture was much stickier and we had to gently scrape it from the inner sides to make space for the plug.

We plugged both engines with Rock Hard Water Putty (introducing some moisture) and set them in our drying room at 30% relative humidity and 22C.

Disposal of residue

We disposed of the excess by (what else) doing an ignition test. We placed both the wet mix and the dry mix residuals on a flameproof concrete board. We lit the wet mix first because we didn't expect it to burn as well (if at all) since (we thought) it still had lots of moisture in it.

The wet mix caught abruptly and set off the dry mix sitting about 10cm away. They both burned very rapidly and dramatically.

So, we clearly should have kept them separate if we expected to get good data on their relative burn rates. This was also a violation of our safety rules–when we were testing the wet mix the dry mix should have been several meters away, not sitting nearby in the test area. "Sufficiently far from all flammable material" also means the material for all subsequent tests.

Still, it was pretty impressive and confirmed that or mix and process are working.

Test firing

While we waited for the engines to dry in our drying room we built a dynamic Newton meter to measure the thrust of the engines.

We tested the wet mix first (expecting it to be weaker than the dry mix) and...wow! A huge jet of smoke shot straight up and the beam of the dynamic Newton meter clearly bent down. It sounded just like a model rocket engine usually sounds. Our initial estimate of the max thrust was 11N, which would be plenty to launch a small model rocket.

The dry mix was also quite impressive, but only produced about 8N of thrust.

The dynamic Newton meter appeared to work great and we were all anxious to analyze the data.

Looking at the data

We went several rounds on analyzing the data. Our initial plan had been to manually step through the video and transcribe the readings. But then we thought "why not try a little image processing first–just to see?" It turned out to be very effective. We got 375 data points on the wet mix (~15 seconds at 24 frames/second) and over 1900 data points on the dry mix (8 seconds, 240 frames per second). We would never have had the patience to transcribe that many points by hand.

We also realized that there was a problem with harmonic oscillations (too much bounciness) with the beam of our dynamic Newton meter) but we managed to work out a good way to compinsate for that (see the dynamic Newton meter construction page for details).

The red line is the raw data and the blue line is the data with the beam harmonic filtered out. X-axis is seconds and Y-axis is Newtons.

We didn't quite get the 11N we thought, but a still-respectable 9.6.

Again, the red line is the raw data and the blue line is the data with the beam harmonic filtered out. X-axis is seconds and Y-axis is Newtons.

This time the peak was only 5N, and the shape of the curve is a little odd. One possibility is that, because it was so sticky and hard to work with, the fuel was not evenly packed in the engine. So the initial hump, fall back, and final thrust might represent an air bubble in the fuel near the nozzle end.

Comparison with published data

For comparison, here is the data from the wet mix plotted against the published NAR certification data for[Estes A8 & B6 engines.