Report on the H5a rocket flight

1.

Construction of the rocket [Top] [Contents]

• battery pack,
• connection board (a sort of switchboard),
• timer unit,
• instrumentation unit,
• transmitter.

A dipole antenna is attached to the lower part of the rocket body.

The upper compartment contains the parachute system. This consists of a small drogue parachute, a main parachute and two pyrotechnical devices, one for ejection of the parachute compartment hatch and one to release the main parachute. And finally, the nose cone, which is designed (von Karmann shape) to minimise drag in the trans-sonic range (between Mach 0.8 and 1.2), on top of the rocket.

2.

The preceding events [Top] [Contents]

Soon, this proved to be a very difficult task, especially since the mass of the motor and the minimum required components, excluding tail fins and rocket body, already totaled 6.5 kg. When the complete rocket was weighed for the first time, about a week before the launch, it turned out to be on less than 9.5 kg! And this didn't even included some of the smaller parts, which had still to be incorporated.

The goal of being the first NERO department to fly a complete rocket supersonic thus proved to be unreachable. Nevertheless, in trying to achieve this goal, we managed to get a lot of experience in designing and building light-weight constructions. For example, a de-blocking device of 100 gram, the instrumentation container of just 190 gram, the compartment joint rings of 290 gram, etc. However, looking back the so-called "penny-wise, pound-foolishness" management has been followed!

An example to illustrate this: suppose you decide to keep a rocket very simple and that thus no separation system can be used to disconnect the upper and lower parts of the rocket in order to deploy the parachute. And furthermore, the nose section of the rocket is not available for the parachute since it is in use for measuring the temperature of the nose during flight. This decision has then the following consequences:

In this manner, the mass of the complete rocket will grow exponentially. It is therefore necessary to stop this downward spiral in the design. The problem with the development of the H5 rocket that there were a lot of uncertain factors (e.g. strength and mass of the carbon fibre tube) which meant that the mass distribution was unknown for a long time. Because of this and other things, the final mass of the rocket was 1.5 kg more than originally calculated.

The inefficient construction was by the way not the cause of the flight failure. The rocket appeared to be very stable, both during the powered flight as during the ballistic flight.

In contrast to the technical difficulties which had be overcome on the mechanical side, the construction of the electrical part of the rocket progressed quite rapidly. Despite the usual design problems, everything went very smooth. The majority of the electrical units were based on 'qualified' designs and we never doubted the timely operation of the electrical systems.

The different measurements were recorded during the flight in a NOVRAM (non-volatile RAM, with back-up lithium cell). After the flight, this measurement data could be read by connecting the NOVRAM to a specially constructed interface to a C64 computer. This kind of rapid flight data processing, just after recovery, would have also meant a first for NERO. Comparable and sometimes more advanced systems had always failed prematurely in the past.

In hindsight, a clearly weak point in the design of the H5 construction could be pointed out: the interface between the mechanical and electrical systems. On the one hand we had a good (though somewhat inefficient) mechanical construction, having a well-working parachute hatch ejection system and a parachute de-blocking device. On the other hand all the electrical systems were in perfect working condition: the complete electrical lay-out had been design, built and tested weeks before the launch.

However, as a result of the enormous time pressure, certain electrical/mechanical interface issues had not been dealt with and constantly postponed. As a result of this, all the existing problems had to be resolved at the final moments before launch, which had a negative impact on the quality of the overall system. Which interface problems did actually surface? They were numerous:

• An amplification Printed Circuit Board (PCB) had to be placed in the nose cone, close to a thermocouple. It was not clear how this PCB had to be mounted, so eventually it was decided to just bond it onto the 'ceiling'.
• An acoustical alarm (buzzer) had to be placed in the nose cone, with the problem and solution as in the previous item.
• After testing the pyrotechnical bolt, the main switch of the electrical system proved to be damaged as a result of the gas pressure and the shock wave associated with the detonation.
• The sensors in the parachute system had to be tied down using ordinary tape because the details of the required mounting construction were not worked out..
• A lot of miscellaneous electrical components had to be fitted somewhere in the rocket because they didn't fit into the containers designed for this purpose. In the end, these components were mounted on two make-shift platforms, which were given the label 'connection-board'.
• The whip aerial had to be mounted into the motor section, which actually proved to be a good solution. However, the aerial could not be fixed to a support plate without completely disassembling the motor section. Since there was not enough time to do this, it was decided to fix it with a piece of tape.

The above list illustrate the fact that, although the rocket contained lots of appropriately functioning systems, the bad workmanship resulted in safety risks for the entire rocket. From the history of rocket flight at NERO we now know that workmanship and success go hand in hand for rocket technology. As Murphy's law goes: everything that can go wrong during a flight will eventually go wrong. The only question that remains is: at what moment?

From the above we can only conclude that the H5 flight just had to go wrong.

3.

Reconstruction of the flight [Top] [Contents]

In these type of conditions, nothing goes right. To make things even worse, the H5 rocket proved to be too large to fit into the wooden launch tower: the fins could come into contact with some of the frames when the rocket would rotate more than about 15°. The launch tower had to be taken down in order to remove the rocket. Fortunately, the 4 meter tall mobile launch tower from the Drechtsteden department can be erected very quickly. This meant that, with about half an hour delay, the H5 could still be launched.

After ignition, the rocket took off quite stable and almost immediately disappeared into the low cloud cover, thus preventing visual flight tracking. Through the telemetry instruments it was at that moment already possible to detect that something had gone terribly wrong, but because of the excitement of the moment and the corresponding noises, nobody had noticed yet.

The expected detonation sound from the pyrotechnical device about 30 seconds into the flight did not come. A short while later, the eerie sound of a rocket coming down ballistically could be heard across the launch area. When this abruptly stopped and the telemetry data stopped coming in, everybody realised that the H5 rocket was no longer.

The consternation was immense: six months hard work, in the last few days even 24 hours a day, all blown to smithereens in just a couple of seconds. This makes you wonder what for god's sake do you do this for?

Although nobody had seen the rocket impacting, the recovery team was able to detect the wreckage in a short time, following instructions from battery sergeant-major Teela and people at the observation site Barbara (the bunker). It appeared that the crash site was about 600 meters to the east of the launch tower. The main parachute was still in its folded state inside the wreckage. The cords of the drogue parachute were completely severed. The pyrotechnical device of the parachute hatch had functioned; the hatch itself and the drogue parachute were gone.

It was clear that a premature parachute deployment had occurred at (very) high velocity, destroying the drogue parachute, which made it therefore impossible for the main parachute to be deployed. The intriguing question was whether the parachute hatch had been ejected because of a premature pyro activation or because of mechanical causes. To resolve this question, it was necessary to find the parachute hatch, which was accomplished after an extensive search. The hatch did show burning marks, which meant that premature activation of the pyrotechnical device had caused the early deployment of the drogue parachute and therefore failure of the parachute system. This could only be caused by a failure of the electrical system (in particular the timers), the connection board or the wiring.

Hopeful that the measurement data in the NOVRAM could lead to more precise conclusions, it was carefully removed from the wreckage using a pair of tweezers and a metal saw. Several pins appeared to have been broke off. Nevertheless, after an hour long 'surgery', the NOVRAM could be connected to the C-64 computer. The chip proved to contain non-reproducible data unfortunately, and had apparently been damaged upon impact as well. Detailed research later on showed that the internal lithium battery was still full and that the connections to the chip were not severed.

After removing the, only slightly damaged main parachute from the wreckage it was found that the de-blocking device had worked properly. This must have happened during the flight since the pyrotechnical charge does not ignite spontaneously at impact. It is also highly unlikely that the piston could penetrate the shock absorbing plate in the rearward direction (i.e opposite to the direction of the loads at impact). The timer of the de-blocking device was set at 100 seconds into the flight, which is well after the moment of impact.

Now, it was time to study the radio data in more detail.

About 2.2 seconds after lift-off, just after the motor had stopped burning, all systems appeared to have activated simultaneously, This situation remained unchanged, right until the moment of impact, after a flight duration of 36.7 + 0.5 seconds. Given the very different construction of the de-blocking device and the pyrotechnical device, a direct mechanical cause, such as vibrational loads acting on these systems, could never have induced the activation of these systems. In that case, they would not have been activated at the same moment, but one after the other, or just one of the two would have been activated. This points to a common cause, lying in the electrical system.

A comparative argument can be given for the timers and the wiring of the igniters; nearly all systems were built separate from eachother. Only a few places in the electrical system can be pointed out where a short-circuit of dislocation of a single component could have resulted in a premature activation of all systems. These places are on the connection board (the switch and distribution centre of the rocket), in particular at the power stabiliser.

Unfortunately, it is no longer possible to determine the exact events leading to the failure with certainty since several pieces of evidence have been destroyed in the impact. However, with high probability it can be claimed that the cause of the failure of the H5 rocket lies in the electrical system. The most likely direct cause can then only be the occurrence of resonance in the Printed Circuit Boards (PCBs) of the connection board, which led to contact between the PCBs and thus a short circuit. Fact is that the PCBs were indeed mounted very close together, lacking any insulation.

4.

Conclusion [Top] [Contents]