Each flight log entry usually represents a launch or test day, and describes the events that took place.
Click on an image to view a larger image, and click the browser's BACK button to return back to the page.

Team Members at Event: PK, GK, AK, John K and Paul K.
Number of Firings: 10

We have finished constructing the thrust measurement rig, and after a couple of modifications to it we managed to obtain some good quality data for both water and foam. For details of the rig see the section below. This is a bit of a long write up so you may want to grab a cup or skip ahead to the sections you are interested in.

Day's Events

• When we initially set up the test rig we ran a 4 tests to see how effective it was and if our measurements were within range of the scale. We first fired the rocket 2 times, once with water and once with foam. These were uncalibrated tests and were meant to give us an idea how well the set up was working. One the most obvious issues was that the scale needled bounced around so much due to the scale's spring mechanism that reliable readings could not be made with any accuracy.
• During online forum discussions it was suggested by Clifford Heath to use a dampner made from a disk on a stick sitting inside a bucket of water. Anti-gravity Research had used a similar setup for their thrust measurement. We placed a small disk ~100mm in diameter in a bucket and it certainly reduced the wild oscillations but it was still bouncing around too much. We then switched to a much larger disk in more water and that settled the needle down really well on the subsequent 2 test firings.
  
  On review of the data we found that the large disk had induced a fairly large wave in the bucket which then reflected off the sides and gave a great big sine wave in the data. We were getting close but we needed to get rid of the wave effects inside the bucket.
• We then built a much smaller dampner from a 1.25L PET bottle but this time the disk had holes in it and fit fairly snugly in the bottle - a lot like a coffe plunger. This allowed water to pass through the holes and around the edges but no big waves were created and that smoothed up the data really well.

Data

The thrust curve data shown below was for a rocket with the following parameters:

Experiment setup and Explanatory Notes

We chose a restricted nozzle to help stretch out the data so it could be more accurately measured and visualised.

The thrust measurement rig was balanced so that it read zero Newtons with the rocket empty. This was done so that the data obtained would reflect the thrust (force) that contributes only to rocket acceleration, as it already incorporates the force needed to overcome the weight of the water in the rocket. Hence the thrust measurement does not represent the absolute thrust. Ultimately we are only interested in how much force is contributing to the acceleration of the rocket upwards anyway.

If we had balanced the rig while the rocket is loaded with water, by the time the thrust ended and all the water was out, the scale would read the weight of the water since it would now be unbalanced by that amount. This would have introduced a bias to the data. Since we do not know at what rate the water empties from the rocket, we later cannot remove this bias from the thrust data.

Three firings were made with water only, and three firings were made with foam so that reproducibility and accuracy of data could be verified. All the parameters for the 6 firings were identical except in the foam tests, 50 ml of water was replaced by 50 ml of normal bubble bath solution. For the foam tests the water was also redistributed in the rocket so that Jet Foaming could be used. See previous tests for more details.

The total volume and weight of water was the same for both tests.

Analysis

It should be first noted that these are static tests and in real life an accelerating rocket may behave differently, however, these tests are a good approximation to what really happens.

We obtained the following thrust values by video taping the kitchen scale on the rig at 60 frames per second and then painstakingly (and I mean painstakingly) stepping through frame by frame in the video and recording the values in a spreadsheet. Digital logging is definitely needed. Much thanks goes to my wife for helping to enter the data.

Water

NOTE: Click on the graphs for more detail

Water only thrust curves

From the curves you can clearly see that the data is quite clean and test 1 and 3 were almost identical. Test 2 shows a slightly lower thrust at first and remaining lower, however, it is compensated by a longer burn. Why this happened we are not sure but perhaps something to do with the turbulence around the seal in the nozzle.

The area under the curve represents the total impulse generated by the rocket and this was:

Which is within ~3% of each other giving a fairly accurate reading for reproducing the data.

Explanation of different parts of the thrust curve for water only

The initial spike is caused by the release of the rocket and can be ignored, or averaged out.

There is a clear almost flat line for the thrust in the water phase, and then there is a clear transition to air-pulse which remains constant for a short while before tapering off to zero. The shape of this curve was identical in all three tests.

Foam

Foam thrust curves

This is where it gets a little more interesting. All three foam thrust curves show an almost identical shape although the thrust curve is a lot more rough than for water only. This is most likely caused by the foam density varying throughout the firing. If the foam density was homogeneous then this curve would be a lot smoother.

Foam thrust curve explained

Again the curves show the initial launch spike, but also show an unusual effect where the thrust varies markedly for about 500 ms approximately 400 ms after the start. This isn't an anomaly of a single firing as all three firings showed this. This is most likely related to the Jet Foaming technique or the blow-through effect and we will explore this region further in future tests. At the moment we are referring to it as the Foam Thrust Anomaly (FTA) until we can identify the cause. It also doesn't appear to be a resonance in the measurement system as the data does not show this effect in any of the other tests or any other portions of the thrust curves.

The area under the curve represents the total impulse generated by the rocket and this was:

Which is within ~3.5% of each other giving a fairly accurate reading for reproducing the data.

The curve when averaged is a fairly slowly decaying curve with no transition to air-pulse.

The maximum thrust generated during the flight is about 50% of the maximum for water only. The average thrust was 9.98N for foam compared to 20.4N for water only.

Now when we compare the total impulse generated by water only and foam it is evident that foam results in lower total thrust. Foam produced only 86% of the total impulse of water.

Thrust curves combined for foam and water

However, the foam generated thrust for almost twice as long as the water only rocket! In one test the thrust lasted for 8.5 seconds!

Total impulse isn't the the whole story when calculating the apogee of a rocket. When you take drag into account and the fact that it is proportional to the square of the velocity consider this: Is a faster rocket burning for shorter amount of time going to fly higher than a slower rocket burning for a longer time?

We pose this as a challenge to the reader to use the data provided to calculate the rocket apogee. You can consider the dry mass of the rocket to be 650 grams.

The data is available as a spreadsheet here: Thrust_Curves_080407.xls

We would love to hear your answers - contact us here

Water Rocket Thrust Measurement

(If the video does not play, try the latest Flash player from Macromedia)

Thrust Measurement Rig

The thrust measurement rig consists of a balance with the rocket mounted on one side with a kitchen scale on the other. The kitchen scale measures 0 - 500grams. Attached to the lever arm is a dampner that prevents the spring in the scale from oscillating as the rocket induces vibration into the rig.

The scale cost about $6 and will be eventually replaced with an electronic load cell and data acquisition system to make measurements faster.

The scale dial is video taped at 60 frames per second which gives us a stable time reference for the measurement samples. These are then transferred to a spreadsheet by replaying the video frame by frame.

The balance can swing all the way around on the pivot allowing us to permanently mount the rocket on it, while giving us the ability to easily fill it with water.

The nozzle is 85mm from the pivot point, while the scale is at 1415mm from the pivot point.

Refer to the photos at left and the video for further details on the operation of the rig.

Lessons Learned

• Foam produces a lower total impulse compared to water in this test.
• When a water only rocket fired, there was a brief period during the air pulse where constant thrust was produced for about 300ms. This was unexpected as you would expect there to be constant decay in the pressure. This may be due to the bottle's elastic property. The bottle contracts as the air exits keeping the thrust constant until they stop contracting and you get the expected decay in thrust.
• There is no clear transition to air pulse with a foam rocket and the thrust is a lot rougher most likely due to non-homogeneous foam density.
• The foam thrust lasted about twice as long as the water only thrust.

What's Next

We will perform more tests with different sized nozzles, various pressures, different amounts of foaming agent, and different foaming agents.

The big test will be the construction of a number of CD nozzles to see if the total impulse can be lifted above that of a water only rocket. We are hoping that with the lower drag and an improved nozzle efficiency we will be able to improve on water only water rockets.

Update: The CD nozzle results are now available here: Day 32

Acknowledgements

Thanks also to Jim Fackert for suggesting the camera pointed at the scale to take readings. We should also thank Gordon M, Richard W, David L, Trevor, Brian and many others for useful suggestions on how to set up the thrust measurement rig.

Test Record