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.

Date: 10th February 2013
Location: Doonside, NSW, Australia
Conditions:  Sunny, light winds < 10km/h early, 28C
Team Members at Event:  GK, PK, Paul K and John K.

Launch Day

Today was a bit of a mixed day again with some good successes, but also a spectacular failure. The launch conditions were great with nice blue skies and fairly calm conditions.

Shadow II

The Shadow II rocket was set up first as normal while conditions were still calm. We decided to keep the camera we used last time as tests showed that it was working properly. We shook the electronics section vigorously dropped it several times to test if anything would interrupt the video, but all was good.

Assembling the payload bay	Filled with 1.7L of water	Loading it onto the launcher
Final preparation		Launch pressure: 440psi

We pressurised the rocket to 440psi (30 bar), the highest pressure this rocket has ever been and launched it. The rocket beautifully CATOed about 8 feet in the air, The payload section and the upper foot of the rocket kept going a little ways up. The parachute was ejected as a result of the explosion and gently guided the whole payload section back down. You can hear the servo motor activate about 1 second before it touches down. The delay was set for 9 seconds.

Here are some frame grabs from the video of the explosion.

The bottom 3 feet of the rocket also survived and landed right next to the launcher. The middle section of the rocket completely disintegrated in a shower of fiberglass confetti. This is why you want to stay completely well away from these kinds of rockets when they are pressurised. The debris field was well within the exclusion zone. The sound was quite impressive and you could hear it echo off the hills.

Top and bottom landed right next to each other	Inspecting the payload

We weren't too disappointed because we knew this day was coming sooner or later. The plan was always to keep pushing this rocket to see how far it will go and is the reason we have been increasing the launch pressures incrementally. Personally I was expecting one of the bulkheads to give way before the pressure chamber.

Lower rocket section in bottom right	Rocket eating trees are starving today	Lower section already on the ground

Okay let's go to the video tape ....

Failure Analysis

We believe this was a very unusual failure as once a rocket launches the internal pressure starts dropping rapidly so there should be less and less stress on the pressure chamber as the rocket climbs. One would expect that either the rocket blows up on the pad or it flies, but not explode mid flight. We believe there was an additional source of stress in the pressure chamber during the flight which caused it to fail. At the time the rocket was 2.4m (8 feet) off the ground and travelling at 42m/s or 151km/h (94mph).

Potential sources of additional stress:

• The large acceleration (~44G at the time) may have added enough compression stress to the pressure chamber that caused it to fail at a weak point. At 44G you have about 15Kg (33 lbs) pushing down on the top of the pressure chamber due to weight of the payload section, parachute, deployment mechanism and nosecone. If you include the weight of the pressure chamber up to the failure point, that's a 29Kg (64lbs) load. The interesting thing is that at the time of the explosion the rocket hadn't hit peak acceleration yet. There was also water left in the lower section as can be seen in the video.
• The rocket may have been undergoing a bending stress which may have put more stress on one side of the pressure chamber than on the other.

A shockwave may have travelled up the rocket as it cleared the launch tube. The reason this may be a real possibility is that the simulation results show there is a large spike in the acceleration just after it clears the launch tube. On the right is an example from Dean Wheeler's simulator. At first we thought this spike was an artefact of how the simulation calculates the acceleration, but from real data captured with an accelerometer we see that this acceleration spike is a real phenomenon. Here are two examples showing this spike from our earlier flights of different rockets: PolaronG2_FC_Acc_Vel_290111.png and aircommand_flt2_boost.png.

Contributing Factors

A contributing factor may have been a weakening of the epoxy due to a combination of air compression heating and the higher ambient temperature.

Here is a post http://ausrocketry.com/forum/viewtopic.php?f=10&t=3841&p=46345#p46345 from Chris about glass transition temperatures he found for epoxies. We use the West Systems 105/206 combination and only do room temperature curing. With a Tg of ~60C means we could be close to having reduced strength in the epoxy when pressurised. Here is more data from West Systems: http://www.westsystem.com/ss/assets/Product-Data-PDFs/105206-Technical-Data.pdf

We really need to do more investigation in this area as to how much the air heats up inside the pressure chamber at these pressures and how that impacts on the strength of the pressure chamber. The rocket was pressurised in about 1 minute for this flight.

Weak Point Location

It is hard to find the exact point of failure, but it looks like it was just below the top coupler. The top coupler was added after the crash we had a year ago. Although we cut away the damaged section there may have been an unseen stress fracture in the rocket below the replaced section. Stress also tends to concentrate at transitions and there is a sharp transition from the main body tube to the coupler. On the right is an example of a pipe and coupler stress analysis. While not dimensionally related to this rocket it does give you an idea where the stresses are concentrated.  (Image source)

Below is a shot from the high speed video that shows the approximate location of the failure. It appears to be right near the top coupler. The distance from the tip of the nosecone to the end of the coupler is 92cm. From the video we measured the distance from the tip of the nosecone to the location of the puff of vapour and got approximately 99cm with a margin of error +/- 10cm. This is pretty close to the coupler.

Moment of failure	Top coupler	Delaminated coupler stuck to body tube.
Lower coupler	Confetti	The Shadow now fits into a suitcase

We had a close look at a couple of the pieces that were in contact with the top coupler. You can see that at least these two pieces had the outer layers of the coupler still stuck to them implying that the bond between the coupler and the tube was good, and it was the coupler that delaminated during the explosion. None of the cracks in the body tube propagated very far along the coupler (as can be seen especially on the lower coupler). If the two couplers hadn't stopped the crack propagation we would have had a lot more confetti.

It is also interesting to note that from the on-board video you can see how much the camera turned while the rocket was pressurised. This seems to imply that the tube must have been stretching considerably and the twist comes from the cloth "unwinding" in the opposite direction it was laid up. This is only a theory and we can't be sure. Was this partially the result of weakening epoxy, or just the normal stretching of the composite under load? I believe we saw a similar rotation in earlier flights. The rotation was perhaps 20 degrees on this flight.

Camera direction before pressurisation	Fully pressurised

Good information did come out of the CATO though. This was the first time we had a fiberglass pressure chamber burst since we started making and testing them so it was good to see what happens when it actually does break. It's also a good wake up call for how much energy is stored in the rocket and why safety is so important. We also learned that a CATO mid flight IS possible and the compression forces need to be taken account during acceleration. A hydrostatic test may not reveal all the potential failure modes when acceleration is not taken into account. Applying an axial load to the pressure chamber under test may be required in future tests.

The launcher also didn't leak after the quick connector was properly tightened and the video camera recorded the flight without problems unlike last time.

We are considering what we'll do next with the Shadow series, but most likely investigate new pressure chamber materials and improving their strength. We are likely to put the Shadow on the back burner for now, and concentrate on some of the other projects we have lined up for this year.

Inverter Rocket

We repaired the rocket during the last two weeks, with most of the repairs only involving splicing new bottles. See the Inverter build log for more details. We also made some modifications to the parachute release mechanism to prevent it from opening early.

After the Shadow debris shower, we swapped the release head again and set up the Inverter rocket on the pad. When we tried locking the nozzle into the release head it would go in fine, but would get wedged and would not release. This is what happened during the last launch as well. We tried a second nozzle with exactly the same problem. So we disassembled the release head and used sandpaper to file down the little plastic tabs by about 1/2mm. On re-assembling the release head we found that it was still sticking. It turned out to be the wider section at the top of the nozzle that was getting wedged not the little tabs. So we used the sandpaper to sand that down and then the nozzle released easily.

We had to sand down this section to make it not stick in the launcher

With field surgery over, we set the rocket up on the pad and filled it with 6.5L of water. We used a little food colouring this time to make it stand out better on video.

Filling with water	Attaching camera	Paul looking forward to launching it

As we approached ~110psi (target pressure was 125psi) something started leaking. Leaving the air on we decided to launch the rocket anyway guessing it should have enough pressure to fly. The rocket took off very slowly again before accelerating nicely. It arced over and deployed the parachute.  Here are a few shots of the launch from different camera angles.

From the ground it looked like the parachute opened late, but the on-board video showed that it deployed right at apogee, and took 2 seconds to fully inflate. The modifications we made to the release mechanism worked well, so no more early deploys. :)

The rocket came in for a gentle landing because the large surface area of the rocket helps to slow it down. The altimeter gave a reading of 315 feet which was not unexpected because of the lower launch pressure and the fact it arced over rather than going straight up. You can see how stable the rocket was on the way up with almost no roll.

Here is a highlights video of the flight:

While swapping the release head to fly one of Sam's rockets we found the cause of the Inverter leak. We forgot to re-tighten the release head on its seal which we loosened while trying to work out why the nozzle was wedging. Doh!

We were going to launch the Inverter again, but by then the wind had picked up to around 30km/h with stronger gusts and so we decided to abort the flight. Better to save it for the next launch. This rocket has a very large surface area and so cross wind can easily affect it.

We were happy that the rocket finally flew well and we're looking forward to flying it again. We only had the two launches on the day, but it was fun watching all the other pyro rockets as there were quite a few new faces again.

Flight Details