Scope

Overview

This document will analyze the flight of Hermes 1 at Friends of Amateur Rocketry in July of 2018. The document will outline the flight performance of the vehicle, calling specific attention to any deviances. These flight events will be discussed in roughly chronological order, by subsystem. Under-characterized or poorly understood behaviors will be noted for future efforts. A section at the end of this document will discuss lessons learned.

Methodologies

Times are indicated as T+ or T- from the first vertical motion on the rail (liftoff). Times are indexed off of the Telemetrum data.

Timeline

Flight Statistics

Simulation Accuracy

The last simulation before flight predicted an apogee of 31600', which proved to be very similar to the true performance of the rocket. At 2.5% error, it seems our model is accurate enough to predict future flights of similar rockets.

Avionics

Payload

Propulsion

Thrust Curve

See the motor data page.

Long Startup Time

The motor took significantly longer to light than previous tests, including the second P motor test that used an identical igniter with significantly more propellant surface area. During static fires the igniter was retained by gravity until the motor had built up enough pressure to eject it. Video evidence shows that the igniter fell out of the motor within a second of firing.

Future efforts could faster-burning igniters (pyrogen or BKNO3), higher surface area propellants (pixie dust), or additional mechanical retention. While chuffing represents a minor issue on this flight, successful, rapid, and complete ignition is a necessary technical milestone for multistage flight. Care must still be taken to not over-pressurize the case or clog the nozzle with a large igniter.

Recovery

Piston Deployment

Firing Force

Comparison with Ground Test

Parachute Drag Coefficients (Effective) and Velocities

It is not possible to do perfect analysis of the Parachute Drag coefficients because we cannot back out the drag on the mission package and booster sections. However, we can determine the effective coefficients (including the drag on these bodies).

Drogue

We trimmed the data so that it didn't include opening shock loads or main opening. We then performed analysis at two different altitudes (~7k m, ?, and ?) in order to see the difference between different altitudes.

7k-8k meters

After parsing the data between 7000 and 8000 meters and eliminating outliers, this is a scatter plot of velocity vs. altitude. Velocity was determined using backwards differentiation on a centered-moving average of 201 data points of height and time.

A normal probability plot of the velocities shows that the distribution of mission package velocity under drogue at this altitude range was approximately gaussian (similar results were shown, but not included here, for velocity squared, which is proportional to drag):

Main

Parachute Shock Loads

Drogue

Main

The Telemetrum switches the state to "main" at 298.22 seconds. Shown by the red line in the graph below, this indicates that the Raven fired the Tender Descenders first. Also shown by this graph is that main opening appears to take approximately seconds and be an approximately constant opening force (main opening is shown approximately by the dashed vertical green lines).

This is not exactly what I would expect–I would expect that as the canopy changes shape during the filling process, the opening force changes somewhat.

As you can see here, main opening took approximately 11 seconds. We can compare this with a theoretical estimate derived from the following formula (from NASA TM X-1786):

To determine the shock force, we need to know the upwards acceleration applied by the parachute during opening (subtracting off the effect of gravity). Then we can calculate the opening shock factor using the opening shock force formula.

To do this, we need to smooth the data even more than originally as the double differentiation to find vertical acceleration introduces a good bit of noise.

E-match voltage anomaly

Post flight data review notes abnormal voltage readings on Telemetrum apogee channel. Despite commanding an apogee event, the Telemetrum continued reading 4.2V across the channel for the remainder of flight. The e-match should register a significant increase in resistance within 10 mS of exceeding it's no-fire current of no fire current here. This suggests that the initiator misfired.  A fishbone analysis was conducted to classify the nature of the issue. Evidence collection is ongoing, however it appear probable that the Telemetrum e-match is internally shorted, either autogenously or to the piston.

We checked continuity post-flight and neither e-match had continuity. This suggests that the e-matches were not internally shorted to the piston. This is not a conclusive test because the wires could have been jostled during landing or travel.

Structures

Integration

Integration began at roughly 1:30 pm PST and the rocket arrived on the pad at roughly 3:00 pm PST.

Lessons Learned

• Having a good checklist would have been beneficial, even though we likely wouldn't have been able to integrate the rocket in the amount of time that we were given before the end of the launch window if we had been following a strict checklist.
• Regular batteries are NOT enough to fire 2 e-matches in parallel. We ran into this problem during the ground test, as well as a previous static fire.