After the static tests and flight trials of the dual thrust rocket on day 138, we believed there was some room for improvement with the performance. One of the major problems was the blow-through effect allowing pressurised air to escape from the boost chamber before all the water had run out. (Here is an example of the loss of thrust due to blow-through from previous experiments.) The sustain air pulse thrust also didn't look like it was efficient due to the internal nozzle being so far into the boost chamber. Kind of like blowing into your own sail. Some water was being held up against the sides during the sustain air pulse as well. Another problem was the vacuum created in the boost chamber during the sustain air pulse causing the bottle to collapse inward.

To address these issues we extended a narrow thin-walled tube down through the boost nozzle. This tube weighs 9 grams.

Original configuration with Jet Foaming spacer

New improved configuration with Non-return valve and extension tube

Launched with the same Clark cable-tie launcher

Non-return valve

Extending the tube all the way down to the nozzle meant that we would lose the water self-levelling feature of the jet foaming spacer. So we built a non-return valve into the bottom of the spacer so that water and air could flow up into the sustain chamber during pressurisation. 

How it works

The non-return valve is built into the adaptor that holds the extension tube inside the spacer. We cut a "V" shaped groove around the outside of the adaptor to house an o-ring. Small inlet holes were drilled from the bottom into the base of the V-groove.

During pressurisation the higher pressure in the boost chamber causes flow through the inlet holes and because the o-ring is in a V-groove it pushes the o-ring out of the way breaking any seal allowing the air and water to flow out from under the o-ring. The top of the adaptor is narrower than the spacer with less than 1mm gap all around. This allows the water and air to flow upwards into the spacer without affecting the flow efficiency back down the extension tube during the sustain phase.

Shortly after launch, the pressure differential between the sustain and boost chambers switches and the water pushes the o-ring back into the V-groove blocking the inlet holes. This ensures that all the flow from the sustain chamber only goes through the extension tube.

Non-return valve and extension tube adaptor

Extension tube fitted into the non-return valve

All the components including the existing spacer

Fully assembled ready to be fitted into the boost chamber

The extension tube extends past the end of the nozzle

Thrust Measurement

In order to do some quantitative comparisons we ran a series of tests of the rocket on the load cell to measure the amount of thrust generated. We used our load cell setup for these experiments. Because this dual thrust rocket produces such a wide difference in peak thrust between the two phases we had to reduce the gain on the load cell amplifier so that we wouldn't clip the top of the boost thrust phase but on that setting the amplitude resolution for the sustain phase wasn't very much. At 100psi the boost peak thrust was expected to be around 450 N (101 lbf) and during the sustain phase the peak thrust was expected to be around 20 N (4.5 lbf). So we ran some tests where the boost phase was completely visible and other tests where we turned up the gain to get better amplitude resolution in the sustain phase, but clipped the boost phase.

Experiment Setup

The extension tube had a 7mm internal diameter, which was a little on the larger side, so we made a small plastic insert that was glued into the bottom of the tube with a 5mm hole. This way we could compare the performance of the previous spacer design with the new design.

Small 5mm ID insert for the extension tube to reduce it from 7mm to 5mm

Extension tube fitted with reducer.

The rocket under test had the following parameters:

• Diameter: 90mm
• Capacity: 6085mL
• Boost nozzle: 22mm (20.86mm effective when we remove the cross sectional area of the extension tube),
• Sustain nozzle: 5mm
• Configuration: 3 x spliced pairs joined with a full-bore tornado tube on the bottom, and 18mm tornado tube at the top.

We used 3 video cameras to record each test. The HD camera for real-time view, the GoPro at 240fps to film the entire rocket, and the Exilim FC100 at 210fps to film close-ups of the boost chamber.

The release mechanism was a standard Clark Cable-tie launcher in order to use the full 22mm nozzle. It was attached by a string to prevent it from impacting the ground during launch.

Results

Preliminary Tests - 22/12/2013

The first tests with the new configuration were carried out just mounted to the test stand (no load cell). These tests were used to check whether the system even worked at all. Two test firings were made with a full 7mm sustain nozzle. These tests were carried out at 100psi (6.9 bar) and 1700mL of water.

Test #1

Test #2

Measurement Tests - 23/12/2013

The following tests were carried out the next day mounted to the load cell.

Test #	Config	Water Amount (mL)	Split Boost / Sustain (mL / mL)	Pressure (psi / bar)	Nozzle Boost / Sustain (mm)	Thrust curves
1	Extension	1800	1150 / 650	100 / 6.9	22 / 5	Load cell amp set to low gain to show all of the boost phase. Second view is the zoomed in view onto the boost phase.
2	Extension	1700	1045 / 655	100 / 6.9	22 / 5
3	Extension	1700	1100 / 600	100 / 6.9	22 / 5	Load cell amp set to high gain to show details of the sustain phase. The boost phase is clipped.
4	Extension	1800	-	~30 / 2	22 / 5	Self launch - Invalid test
5	Extension	1800	1300 / 500	100 / 6.9	22 / 5	No o-ring test to try to help self-levelling.
6	Spacer	1800	800 / 1000	100 / 6.9	22 / 5	Looking at the sustain phase of the spacer, with dramatic drop off at the start of the air pulse.
7	Extension	1800	1100 / 700	100 / 6.9	22 / 5	No o-ring - split water (poured half into sustain chamber)
8	Spacer	1800	800 / 1000	100 / 6.9	22 / 5	Low gain to inspect boost phase of the spacer. Second curve is a zoomed in view.

Analysis

Comparison of test #6 and #7

Measuring the thrust difference in the boost phase with and without blow-through is difficult to do with our current setup due to how short the phase is. The boost phase water and air last about ~0.25 seconds. We sample the load cell at only 240Hz and due to the initial launch upward kick, the rocket briefly oscillates on the load cell so we get quite noisy data during the first 0.1seconds of the launch. How effective the no blow-trough boost actually is will be done during flight trials.

Here are some video frame captures from the tests:

Looking at the boost phase

There is no blow-through during the boost phase

With the o-ring removed you can see some spray from the inlet holes.

Looking at the spacer sustain phase with collapsed bottle

Another view of the collapsed bottle

Start of boost air pulse

Dramatic bottle collapse during the spacer sustain phase

Video

Here is a highlights video from the tests:

Water Self-Levelling

The one problem encountered during the load cell tests was that the water self-levelling didn't quite work as well as expected. This was traced to pressurised air entering the extension tube directly from the release head, preventing a large enough pressure differential between the sustain chamber and the boost chamber. The pressure differential is important in order to force the water into the sustain chamber. In test #7 we poured some of the water into the sustain chamber beforehand to help distribute the water better.

Self-Levelling Tests - 1/1/2014

After a short break for Christmas and New Year we followed up with the self-levelling tests.

The first attempt at fixing this issue was to add a small flexible tube to the release head that bypassed the bottom of the extension tube so that all the air would go into the boost chamber. The first trial showed that only very little water and air went through the non-return valve mostly because the hole in the extension tube offered much less resistance than the non-return valve. This makes sense because the holes in the non-return valve are small and the o-ring provides some amount of resistance. With most of the water going through the extension tube this meant that most of the water was pushed up into the sustain chamber.

We then took a skewer stick and wrapped some electrical tape around it to fit into the extension tube. This acted as a plug to prevent water being forced through it and allow it to flow through the non-return valve. The plug just gets ejected during launch. It could be permanently attached to the inside of the release head. This did the trick and allowed the non-return valve to do its work and correctly self-level the water in the two pressure chambers.

Too much air entering the extension tube directly prevents correct water self-levelling

Using a thin bypass tube to prevent air entering the extension tube.   The bypass tube allowed too much water to flow up through extension tube.

Plugging the extension tube prevented water flowing up through it allowing it to flow through the non-return valve. This achieved the correct self-levelling operation.

Other Observations

• During test #4 when we reached about 30psi the rocket self released. Review of the video showed that the released head collar wasn't pushed all the way up and the tops of the cable ties were showing.
• During test #5 the release head impacted with the deflection brick and snapped the Gardena adaptor. So we switched to our second identical Clark cable-tie launcher for the remainder of the tests.
• During the boost phase it was evident in the slow motion video that the extension tube moved around quite a bit from side to side, but during the sustain phase it was nice and steady. It should be possible to put little flat brackets on the end of the tube that sit just above the release head to help steady it during flight, but this may not be necessary. The brackets would be directly in the water flow and cause extra resistance.
• During correct self-levelling the water always settles a little bit below the non-return valve. This is because the bubbles coming up through the water increase its volume raising the water level. When the bubbles stop, the water level drops.

Conclusions

The extension tube prevents the blow-through effect and allows the water to be ejected cleanly from the boost chamber. The extension tube also prevents the vacuum forming in the boost chamber during the sustain air pulse. The sustain air pulse also generated an additional 8Ns of thrust in these tests. With the extension tube plugged in the launcher, the non-return valve works well to self-level the water in the boost and sustain chambers.

Next will come a series of flights with this new arrangement to see how it performs.