Nov 09 2017

Progeny Mk5 Block I Flight 2 Analysis

Last week’s launch was the first complete flight to space and back of our new Block I design, which primarily features more powerful first and second stage solid-fuel booster engines. Despite the incredible record-setting apokee of 493km and being able to recover the payload afterwards (barely), many problems became apparent that have needed to be addressed. In this report we will first cover the details of the flight, then look into solutions for the problems that occurred and finally talk about how we plan to move forward.

The Flight

After delays and a scrub of the initial launch time due to weather issues, the rocket was finally launched off the pad at 01:58:00.03 UTC under command of the Automated Flight Control System. The first stage solid fuel booster kicked in at 67.226kN of thrust to propel the rocket at an initial rate of 4Gs off the pad in order to put enough aerodynamic force into effect to keep the rocket’s nose from lifting too high. Beginning at 85° the nose of the rocket reached a maximum pitch of 86.935° at 2 seconds after launch, well-within limits. Burning fuel at a rate of 39.089kg per second, the 0.625m booster propelled the rocket up to 788.124m/s over the course of its 20.42 second burn, topping out at 76.422kN of thrust. The dynamic pressure at flame-out was 139.299kPa, by far the highest sustained so far by a complete stack of the Mk5. The booster was decoupled as planned 1 second after flame-out was detected, which is when the first flight anomaly occurred.

The speed at which the first stage was decoupled was roughly 2x greater than any previous ascent of a Progeny rocket and the force imparted on the booster by the decoupler was not enough to push it away cleanly. Looking at the telemetry data you can see a noticeable increase in Angle of Attack at T+23 seconds that represents the wobble of the rocket after the first stage’s momentum flew it back into the decelerating second stage, bumping it slightly off its prograde vector. This allowed the nose pitch to also wobble enough to pass the 1.5° change threshold and trigger an extremely early start to the second stage boost phase, kicking in at an initial thrust of 18.267kN while burning 11.588kg of propellant per second. It pushed the rocket up to 992.042m/s before burning out with a final thrust of 18.660kN after 10.36 seconds. During a portion of this time we actually lost telemetry signal from the rocket as it was traveling fast enough to heat up the air around it similar to re-entry and interfere with radio communication until after flame-out and slowing down.

Although the second stage decoupled cleanly and a good 16.72 second coast period ensued, the AoA of the rocket never fully stabilized while under boost from the second stage. The third stage liquid-fuel booster engine kicked off at an initial thrust of 11.995kN to maintain a TWR of 2 and being at 34.143km was immediately able to throttle up to full thrust as dynamic pressure continued to fall from 0.945kPa. The rocket ascended under thrust all the way up into space, with the engine running out of fuel past 78.133km after 35 seconds. The precession grew during the entire boost until the rocket was practically laid out in a flat spin ascending towards apokee. The fins were also shredded as normal at 60km, which was still under thrust and potentially had an effect on the rocket’s stability.

The period during which the rocket rose to its apokee and fell back towards Kerbin was obviously much longer than anticipated but otherwise uneventful. Battery levels were holding up very well mainly because the AFCS had crashed shortly after the rocket reached space and electrical consumption was lower than normal. It was later discovered that during this time over the top of its parabolic arc the rocket flew through a region of radiation up to 10x what we would normally expect based on previous space flights. This ended up causing damage to many of the onboard instruments that made recovering data difficult or in some minor cases impossible after recovery.

The return to Kerbin was extremely violent, as the rocket was still flat out in a spin so 18 minutes after launch it slammed broadside into the atmosphere traveling at 2.007km/s. It was no longer rolling, which meant just one side of the rocket got completely scorched. The large amount of surface area hitting the atmosphere helped to spread out the heating and the rocket sustained a max load of 15.93Gs while decelerating – but at this point much of it was fused together so solidly it was able to stay in one piece. Aerodynamic forces also stood the rocket back up eventually to allow the engine to properly take up much of the heating. Although we didn’t anticipate a return at velocities this high, we had still planned for it to be greater than previous flights and lowered our initial chute deployment altitude from 4km to 2.5km. The AFCS was no longer in command but had properly returned manual control and the parachute also contained a pressure sensor as a final backup. The command to deploy was sent, but the rocket dropped behind the mountain of a nearby island before we could determine if deployment was successful.

After a tense wait of 26 minutes for our range monitoring vessel to sail from its station around the island to check, we learned that the chute had not been ripped to pieces upon deployment and the rocket was intact floating on the surface of the water only 46.342km downrange. After recovery, examination of the rocket showed it almost didn’t make it through re-entry. Most of the downward-facing fairings had melted onto the payload truss, and heat had caused pressures inside to build enough to partially blow off a fairing piece on the other side. The probe core casing was also fried but thankfully did not cook the internal electronics. Other than a new casing for the probe core, the nose cone and payload trusses were complete write-offs. Instruments were all damaged but able to be repaired and thankfully none of the batteries exploded. We still consider ourselves extremely lucky to have recovered this rocket!

The Problems

Extreme Dynamic Pressure

Even though it all held together at 139kPa, we’ve already proven in previous flight analysis that putting so much stress on the rocket isn’t worth it. Ideally we would not like to break 100kPa and the range we are looking to aim for is 60-80kPa. We plan to achieve this by having Umbra Space Industries modify the shape of the booster’s solid-fuel core so that it burns without a linear thrust curve. The modified booster will have a strong initial burn of about 3-5 seconds to still launch the rocket at 4Gs and maintain pitch but will then reach a larger bore section that will reduce and taper off thrust.

We will keep the second stage booster with its current linear thrust profile but reduce its overall thrust output further as it will ignite at a much higher altitude than originally estimated. We plan to have it push at an initial TWR of 2 at 15km.

Booster Impact

While the measures taken to reduce dynamic pressure (and thus speed) in the first stage boost should help with the problem of the first stage flying into the second stage, we don’t know if it will completely eliminate it. Right now we are working with USI to see how much we can increase the force of the decoupler, which will add a bit of weight for the larger charge of explosives. The force of the larger charge could still have an adverse affect on the second stage however, so we are also looking into very small retro-rockets that would fire to help push the stage further away upon decoupling.

Early Boost Ignition

This has been a problem in previous Progeny flights, but we’re still not sure of a good solution. As pointed out earlier the large change in Angle of Attack is a clear indication that the rocket is not in a stable flight state, which is something the AFCS could easily detect and therefore ignore pitch change as a trigger for the next boost phase. The problem is that we have no idea how the rocket would behave coasting while in a wobble and whether the delayed activation of the engine would help it stabilize or make things worse when its traveling slower but still in precession as the next boost begins. Right now we are simply working to reduce the chances of this happening by preventing the stages from contacting each other after separation but programmers are talking with Flight Director Lanalye about the consideration of putting some AoA detection into the AFCS that makes it wait for controllers to make the decision to ignite the booster under unstable flight conditions.

Fin Shred Under Thrust

Blowing up the fins while the rocket was under acceleration probably didn’t help matters. It’s not like the fins were really having much of an effect anymore that high up in the atmosphere but the force of the explosions likely increased the instability of the rocket. The solution here is simple: remove the fins entirely. With how high the rocket is flying now by the time the second stage is expended there’s really no more need for fins at all by that point. The rocket will either be in a stable roll and continue to fly straight or be unstable and there’s not enough air density by then for fins to help stabilize it anyways. This will also help to offset the mass we will gain by adding a more powerful decoupler or several small retro-rockets to the first stage.

AFCS Crash

The team behind the AFCS is still uncertain as to exactly what made the control system crash out shortly after reaching space. The error log was corrupted by radiation and what was recovered doesn’t really make any sense, pointing to a closing bracket of one of the triggers that activates once the spacecraft passes 70km and calling it an unexpected token. The call stack unwound from there doesn’t make things much clearer either. It’s possible a stray cosmic ray could have affected the probe core electronics. Despite the crash, control of the rocket was successfully returned to ground controllers as designed.

Radiation Damage

The biggest surprise of the entire mission was the radiation data. The launch was contracted for gravimetric measurements and the radiation detector was added to the second empty payload bay by our resident science team. What data was recovered shows a steady increase in radiation dosage once the rocket passed around 340km ASL. The dosage plateaued at 10.010rad/h at about 380km before falling again after the rocket passed apokee and reached 380km on its way back down, eventually passing through 340km and reaching the “normal” range of 0.010 rad/h that has been recorded on all previous space flights carrying a radiation detector.

Scientists are still working through several theories on what we flew through, from a bubble of captured charged particles that could have come from a (hypothesized) recent kerbolar storm to a band of charged particles surrounding our planet to leaving the protective shell of our magnetosphere entirely and experiencing the harsh radiation of interplanetary space for the first time. Based on this single flight, it’s not possible to know for sure if the data we collected was simply because we flew higher than previously or what is the exact nature of the radiation field we flew through.

Still, the damage it caused the sensitive electronics on board was not something we anticipated and to guard against future high-levels of radiation the fairings will be bulked up slightly to provide better shielding, as will the probe core casing. We regret every gram of mass that needs to be added to the rocket but in this case we don’t see what choice we have otherwise.

It’s also worth noting that experiencing such high levels of radiation so close to the planet is troubling news in regards to future kerbed spaceflight. It would appear a lot more exploring needs to be done by automated probes to better determine what the local space environment is like, which thankfully fits well with our plan to not launch any kerbed missions until 2019.

Increased Radio Range

We actually almost flew entirely out of range of the rocket’s onboard antenna on this last flight! The rocket’s nose cone contains a wire antenna that can send a signal up to 500km, and our apokee reached 493km. The worst that would have happened is a bit of lost telemetry data if the rocket hadn’t been recovered, but we would still rather be able to stay in touch throughout the entire flight so work is also being done to boost the power output up to 1Mm. Based on average power use with the AFCS running, the increased draw of the more-powerful antenna would still give us an estimated 25-28 minutes of flight time, which could be a problem for flights near max range but we will further address that issue if needed.

The Future

As you can see, there is a lot of work to be done and not a lot of time left in this year to do it. KSA operations are planned to end on December 15th for the first full survey of all campus buildings after more than a year of surface wear. In that time to get another rocket launched we need USI to get us a new booster core and also test it with a static fire on our launch pad as USI still has no facilities to handle an engine that big underground. The various other more minor adjustments should be completed in this time frame as well, which right now should be about 3-4 weeks. We can also assemble most of the rocket during this time in preparation for integration of the solid fuel boosters. If the test booster is delivered on schedule and the test is successful, we should just be able to get a launch in on 12/15, maybe earlier. Right now we are saying no earlier than 12/12. Unfortunately this means the Block II’s first launch has to be pushed further back to no earlier than January 2018.