Oct 10 2019

Progeny Mk7-A Flight 2 Analysis

After the failure of the first flight earlier this year, the Mk7 project was mostly shelved as a stand-alone endeavor to develop an orbital small-sat launcher and had to wait until it could provide more use to future Ascension designs before enough reason existed to attempt a second mission. In addition to making a second attempt at testing the steerable guidance fins, gimbaling engine and reaction wheel control system from the first mission, a new payload fairing system was introduced along with a mock-up of the radioisotope thermoelectric generator that we intended to crash into Kerbin so its casing could be tested.

The Flight

Although everything went smoothly during pre-launch operations a storm system out to sea made the weather over KSC inhospitable to rocket flight even though it was expected to improve by launch time. Forecasting is still a bit of a dark art and so we were forced to hold just outside of the final countdown at L-30 minutes and wait to see if conditions would become more receptive to launch before the end of the day cycle. Due to not needing to immediately recover the payload from the water we had no problems launching at night except for the fact that this rocket design is still new and we wanted good visual tracking conditions which meant at least some daylight. If we couldn’t launch before sunset we would have been forced to scrub for the day cycle.

Thankfully we got the GO from the weather desk to proceed forward with the count to our latest-possible launch time, the rocket remained healthy and properly ignited its lower-stage booster at 15:50:00.06 local time to begin the ascent. Still using the gentler 0.625m SRB first introduced with the Progeny Mk6, initial thrust off the pad was 67kN, which gradually tapered off to just 17kN over the nearly 34 seconds of burn time. Once clear of the launch base the 4 lower-stage fins pitched the rocket over to remove any chance of gimbal lock and then all 8 fins were activated  just 1 second after launch to steer on course of 45.5° while continuing to pitch over downrange.

During the first stage boost the rocket passed through a Max Q of 49.049kPa @ 3.755km and climbed to 10km traveling 464m/s, where the first stage was decoupled 1 second after burnout at L+35s and its fins shredded a second later to prevent it from flying back towards KSC or up into the rocket. This almost didn’t work as intended however and a collision of the booster and rocket was only narrowly avoided as the rocket continued to coast upwards to an altitude of 15km, where it was to ignite the second stage motor. The 4 remaining fins continued to guide the rocket along its programmed ascent profile.

At L+48s the rocket successfully ignited its second stage motor, thrusting at 15kN for the duration of its 13 second burn, not quite recouping the velocity loss from the coast phase and only traveling at 452m/s while passing through 19km ASL. Once the engine burnt out the second stage remained attached so that the guidance fins could continue to properly steer the rocket while it coasted up towards the next staging goal of 25km. Despite the rapidly thinning atmosphere, the large blunt nose was creating more drag than anticipated and as the rocket approached 25km velocity was dropping dangerously low and the fins were losing the authority needed to keep the rocket on course. High angle of attack warnings began to show as the rocket started to drift sideways.

Thankfully the second stage was decoupled and the third stage Ospray LF/O engine was activated without issue in time for the engine gimbal to steer the rocket back on course. As it continued towards space and passed through 45km the payload fairings were detached to test their release mechanisms as well as for a small amount of mass savings. The rocket continued to climb and accelerate under 13.5kN of thrust until the engine burned out just before 50km while traveling at 1.4km/s. This was 20km earlier than expected, leaving the rocket within the atmosphere without any steering before entering space at L+2m52s.

Coasting up towards an apokee of only 117km instead of the planned 223km meant the rocket would fall well short of its target landing area in the mountains east of Sheltered Rock. It was questionable in fact whether it would even make it to the opposite shore of the Kerblantic. There was a fail safe option however and ground controllers began to implement it by first activating the reaction wheel system to attempt to bring the rocket’s orientation back to prograde. With some struggle, the rocket was able to point and hold prograde at which time the RTG was shot off via springs in the base of the fairing to send it several kilometers further downrange in the hope that it would make land. Again, the RTG was not a production model capable of producing power so its loss did not affect the rest of the mission.

Once the RTG had been sent on its way ground controllers used what time remained in space to experiment more with the abilities of the reaction wheel system. They attempted to re-orient the rocket to a radial-out orientation (pointing away from the planet) but after almost two minutes of watching the navball wander around past the target and passing through apokee to begin the trip back down, the decision was made to detach the engine and fuel tanks so there was less mass to deal with. This worked and allowed the remainder of the payload section to run through several more pointing tests before it reached the atmosphere at L+7m9s.

Shortly after re-entry communication was lost due to plasma blackout but was regained for 19 seconds and the telemetry during blackout was downloaded just before the probe fell below the horizon and shortly afterward impacted the water only 1-2 kilometers from the shoreline. It was not meant to survive, but the RTG was and although we had its initial trajectory from separation we could not track it through re-entry. An airship was dispatched from Sheltered Rock to hunt it down in the probable landing area – the radio beacon inside was only good for 2km so they had to get close and it took nearly 2 hours of searching before it was located and recovered after impacting ground and not water. With this, the mission was marked a success.

Flight Analysis

Ascent Design

This was the first mission where we really placed all our faith in the Launch Vehicle Designer tool in KSPTOT, because we had to plan for the rocket to coast between staging while still being guided along a programmed ascent path. We still were not sure whether it was aerodynamic pressure due to the high velocity or simply excessive drag due to the thicker atmosphere that caused the RUD on the first flight. It may have been a combination of both, so we set out to reduce both. The ascent was designed to fly the rocket up to 25km before igniting the third stage and the only major mistake made was in how much drag was simulated. We attempted to use what data was collected from the first launch but still set the drag coefficient too low – this is was led the rocket to almost losing too much velocity before the 3rd stage ignition as the modeling showed it traveling faster by then. Now that we have data from a complete ascent we can better refine the drag model of the rocket.

One aspect of the ascent design we did improve upon to help with accuracy (when we get everything else right at least) was to properly align the LVD ascent to the profile that would be calculated and flown by the rocket using a quadratic fit equation. This is sometimes tricky to do since the pitch-over calculated by LVD is done with a linear profile and the quadratic fit allows for a more gradual curve in some places. This is most obvious in the early stages of ascent, something we’ve also displayed in analysis reports for previous Ascension flights:

You can see two places where the pitch profile of LVD’s linear adjustments do not match the calculated non-linear profile. This is fixed by adding more linear pitch points to LVD:

This is the first launch where we had an LVD pitch profile so closely match the calculated profile the rocket would be following during its ascent. Unfortunately in addition to the poor drag modeling the rocket also suffered –

Ascent Guidance Issues

The rocket had trouble holding the proper pitch as it ascended under the first stage, things smooth out past 10km once the lower stage is detached and the second stage ignited but then you have the big offset prior to third stage ignition and the fall off at the end is due to the early MECO that removed all attitude control capability.

We are fairly certain the early issues lie in the high dynamic pressure exerted on the rocket, and thus the control surfaces, during the first stage boost but we won’t know for sure until we can test the larger fins on an Ascension rocket that will travel much slower through the lower atmosphere. We do have some evidence in the fact that during the previous flight’s second stage boost the higher-velocity rocket was still way off its pitch target and in this flight the second stage boost was much better aligned with its planned pitch profile.

Reaction Wheel System

The issue that caused the reaction wheels to be unable to properly steer the rocket to a certain orientation was found to be power draw in excess of what the batteries could discharge due to the amount of mass that was being moved. So the system would shut down, the rocket would drift, the system would power back up and attempt to re-orient but then shut down and reboot over and over – constantly drifting off target and attempting to re-orient but not before shutting down again.

Once the excess mass of the fuel tank and engine were removed, enough power was being supplied to keep the system online as it re-oriented the remainder of the payload section to the desired position and held it there. Holding the desired orientation is the real purpose of the reaction wheels and doing this requires much less energy, with RCS thrusters for larger movements.

RTG Casing

After close examination of the recovered casing the level of damage was deemed acceptable and had there been actual radioactive material inside it would not have leaked out to contaminate the surrounding environment. Unfortunately while this is a good first step it did not impact anywhere near the velocity we had hoped for or that it could when falling out of a service orbit so more testing will be required before the casing can be certified to carry radioactive materials into space.

While we do have the ability to accelerate the casing to high velocities here on the ground, it must also be subjected to the stress and heat of a re-entry prior to impact to fully ensure the casing would remain intact.

Future Plans

This was the final flight of the Mk7-A and plans are already being worked on for the Mk7-B, which will primarily test a new 0.625m segmented SRB that could be used for future Ascension designs as well. We will keep the 0.35m profile for the third stage now that we know the Ospray engine can handle itself properly although we would still like to move on to a complete 0.625m stack by the Mk7-C. Whether this would involve scaling up the Ospray or designing a new engine has yet to be determined.

The Mk7-B will continue to use the same guidance fins, possibly slightly larger in size. The lower stage fins will continue to be shredded but shall do so prior to separation as the extra length of the SRB means no other flight hardware will be in close proximity to be damaged by any debris.

We hope to have more information on the Mk7-B to release in the coming weeks, and debut its first launch in early 2020.