Dec 21 2018

Operations Summary – Closing Out 2018

View post on imgur.com

Monolith Allows Operations to Resume

At the beginning of August we received another interruption by the Monolith that lasted just over 3 months, crippling KSC with a strong electromagnetic field and even affecting the nearby caverns of Umbarg and Kravass. We won’t go into details here – you can read the whole story with this chronological timeline of all the tweets and posts pertaining to the event. Suffice to say, we are happy that it was a transient event and did not permanently exclude us from performing operations at KSC – moving to a new location may not have been possible due to our currently strapped financial state.

During this time there were two major scientific discoveries we reported on. First, scientists used the data collected from our Progeny Mk5 and Mk6 rockets to prove that there is a radiation belt surrounding our planet, with a second, larger one possibly still waiting to be fully explored by our Progeny Mk6 Block II. The second discovery was of a new minor body in our solar system, which astronomers now believe is the dead remnant of a comet after consulting historical records. We expect to hear more about both discoveries, and hopefully many more new ones, in the coming new year.

Also while operations were down we celebrated our 2nd birthday. It was a bit melancholic to do so while being unable to carry out missions but did offer a good perspective for reflection. See what our founder and Operations Director Drew Kerman had to say about it in this post.

Civvies Rotate Service Craft

We have a fleet of three Civvie aircraft, from back in more profitable times, and with our current small cadre of pilots we only operate one at a time. Civvie #2 was used throughout this year and has now been replaced by Civvie #3, which will operate for the first half of 2019 before Civvie #1 comes back into service. A checkout flight was carried out to ensure the aircraft was still operationally capable after over a year of inactivity. Seeing no problems, it has already completed its first mission to sample pollen levels. Thanks to the new night-time air operation capabilities of KSC, data on pollen levels later in the day was able to be collected for the first time, allowing scientists to get a better idea of a full day cycle’s worth of pollen levels.

Genesis and Progenitor Airlaunch a Rocket

After about a year and a half of planning and preparation, we have finally successfully launched a rocket off the back of an aircraft while it was flying at 7km ASL! This is a very satisfying culmination of all the work that went into the collaboration between the two programs, which also included a checkout flight earlier this week to ensure the Deuce systems were okay after being replaced since they shorted out during the initial EM field incident. A final launch rehearsal was also performed two days ago to make double-sure everything was in place to succeed. Congrats to all involved and we can’t wait to dig in for some deep analysis of the mission when we return to operations next year. Until then though you can check out this video for a look at the release and ascent.

Ascension Continues Prep for Launches in 2019 and Capsule Testing

The Ascension team is confident enough in the design of the Mk1 that 5 more have been ordered to launch throughout the first quarter of 2019. They can all carry either a test weight or an actual payload – which one will depend on how many more tries it takes to reach orbit, after which we can place an actual probe atop the rocket. Some design notes about Kerbin I have been tweeted. We will have a launch schedule released at the start of next year.

We’ve finally received delivery of the final capsule design, which turned out to be a sort of middle ground between the two prototypes that were tested this year. The astronauts have already gotten to spend almost an entire month with it and so far have mostly good things to say. Minor problems have been noted for future design iterations but confidence is high that this could be the first capsule to take a kerbal into space in 2019. More extensive testing will be detailed next year.

ATN Continues Tracking New Moonlet Vieras

Discovered back in early September already in orbit, the new Class-D moonlet has yet to encounter Mun again, which has been projected to happen in February of 2019. Given the highly-eccentric nature of its current orbit astronomers suspect that it has only recently been captured and the next pass in 2019 will be its second Mun encounter. Like with all past and current asteroid companions, the moonlet will be closely studied, monitored and used to help increase the accuracy of our future trajectory projections. You can learn more about the moonlet on our Ops Tracker.

Ops Tracker Updates to v9

We were at least able to use the operational downtime to focus on some cool new usability features and bug fixes for our Ops Tracker. You can check out all the details in the change log. The one big feature which did not make it into this update however was real-time streaming launch telemetry. We are definitely focused on getting that implemented for v10 in early 2019 however. Even better news is it will be compatible even with older telemetry data, which means you will able to go check out almost all the rockets we’ve flown in the past to re-watch their launches with the telemetry display. It will look similar to this (minus the video playback, still at least one more major revision away).

ATN Database

The latest update for the Asteroid Tracking Network database is available here, containing 2,929 asteroids and 1 updated with new observation data.

Next week we will post our December financial statement and will update the post with a link. Update: here it is!

From the Desk of Drew Kerman

Out of Character Behind the Scenes stuff

Written on 11/14/18

My last update was before I went on my 4-week road trip. It was a fairly amazing success. I haven’t bothered to write about it or really document it outside of Facebook, which sucks because although all my posts are public, I can’t really find a good way to group them all together and share with people. Anyways, it was a fabulous experience and I came back without injuries or major vehicle problems. I did end up about $800 over budget which has me living paycheck to paycheck at the moment but I’ll tackle that situation next year after I get a bit more financially stable from a fireworks show over New Years.

KSP v1.4.5

Besides working on the Ops Tracker, another thing I used the operational downtime for was to upgrade from 1.3.1 to 1.4.5. Of course, 1.5.1 is out right now and 1.6 will drop early next year if Squad sticks to the ~3mo update cycle they want to do. Still, this jump from 1.3.1 is much more significant than it will be going up to the next two versions and I’m glad it came off fairly well – so far anyways. The more I play of course the more problems I will uncover. It took me two straight days to download and install and properly organize everything. The only major impact so far has been having to update the saved craft files to deal with some changes to Kerbalism. So for example the Civvie cockpit was throwing Nullrefs if I loaded it from a save file, but not when I placed it from the part listing. So I had to place the part, save it, then do a file diff comparison to figure out what modules in the 1.3.1 craft save file I needed to modify. Also I had to do a pass on all my probe cores and aircraft cockpit MM scripts to implement RemoteTech antennas since Kerbalism stripped out its own signaling system under the new maintainer. I did like the signal features of Kerbalism but I’m not unhappy about the move back to RemoteTech since it seems to be well-maintained again.

Aircraft flying okay

Both the Civvie and the Deuce, once I fixed up their save files, were able to be used without major issue. Wheels are still a problem though. I really have no idea how to deal with wheels but yea they suck so bad still. The Deuce can actually now takeoff without me running a kOS script to disable wheel friction, which was causing it to veer off the runway at around 25m/s. The Civvie however I still need to run the script. Landing and even taxiing around I can get a lot of bouncing. It’s very frustrating but doesn’t really affect the overall mission. If I break up on landing I don’t really care (unless I want it to affect the story), I just have to make sure I can get off the ground in one piece.

Airborne launch

The logistical challenge of launching the rocket off the plane was a nice puzzle to work out. I only raged and pulled my hair out a little bit during the process but managed to get it done to my satisfaction.

I did want to have a chase plane up there to get closer video than just from on the ground and onboard, but the release maneuver was just too complex to work out a means of using BDArmory‘s wingman AI or just some sort of kOS scripted control. In future launches more similar to how the Pegasus rocket drops from a carry plane that is just cruising level, that’s more feasible.

The “comms issue” was inspired by the fact that the during the rehearsal flight the rocket did indeed seem to pause on its release like it got the signal to detach but didn’t actually let go right away. It was weird and I didn’t have a good angle on the flight video recording to really see what happened. Thankfully it wasn’t replicated during the actual launch flight.

So yea, the launch flight had to be flown a few bazillion times – not just because I needed various views and data but of course due to issues with the game. In a perfect world I would have flown it just four times. The first time I would fly the aircraft from takeoff to landing, with a save game point after lining up for release. I’d actually release the rocket but just not switch to it or do anything with it. Then I would reload the save point and fly from there until the rocket released, following the rocket to collect its ascent data. Then I would reload again, this time with visual enhancements installed, and fly the release to capture footage from a ground camera. Finally I would reload once more to fly the release and capture footage from the onboard camera.

The first flight turned out okay, it was pretty much doing what I had done countless times before in practice missions, so no problems occurred.

The second flight also went well, since it was also a procedure I had done before. I edited the save file to replace the trim value of the aircraft controls with the current pitch value (AFBW joystick mod uses its own trim feature, not the game’s, so when you trim it affects the actual pitch value). This allowed me to resume flight and have the plane flying almost perfectly level from the start, with the throttle also set to the same point. Then I was able to use kOS to automatically switch over focus to the rocket as soon as it detached. The VOID data logger also switches its source values, which is how I compiled the rocket ascent data. I also made the Persistent Trails path made during the first flight visible – I came very close to an accurate recreation of the release conditions from the first flight.

Grabbing the video footage from the ground is where all the problems started showing up. Two were easy at least: I was able to use an external kOS Telnet terminal window to print out the current altitude, pitch and time so I could fly the release maneuver without the on-screen displays. I was also able to use the Tracking Station to find a good offset value for the clouds to ensure the skies were clear enough for good footage. EVE cloud layer changes update the fastest there.

So of the many various attempts some of the issues were:

  • Confused the flight computers. A script on the aircraft tells the rocket when to detach and a script on the rocket tells it when to ignite. The Telnet terminal windows give you a list of computers to link to – but I didn’t realize at first that list order could change. So I would have the rocket detach, but then that flight computer wasn’t running the script that told it to ignite – the aircraft one was. Grrr!
  • No rocket plume effects. I didn’t realize I never updated my plume configs to change over to the new Shuriken particle system, so the plume just didn’t show up after I successfully launched the rocket. That required a whole game restart. Argh!
  • Kerbalism messages don’t respect the F2 key. I had a great launch and then it pops up a message about signal linkage dropout or whatever. So I muted the messages on my next attempt using the Pause Break key. Oh wait, that’s also the key that disables my SweetFX post-processing. I will admit this was a point where I had a bad flip-out episode. After recovering I tried to mute the messages, then pause the game and hit the Pause Break key again to bring back SweetFX. Turns out Kerbalism doesn’t care if the game is paused either. AUUUGHHH!!! The SweetFX config file had some weird-ass syntax for keybinds I had never seen before and trying to search up how to change it was fruitless. So I tried to just remove Kerbalism, but being such a deep-seated mod the game save just refused to load. I went back to looking into remapping the SweetFX keys but no matter how I tried to format it the game would crash instantly on startup. Finally I just deleted the whole keymapping section altogether (it was by default all commented out anyways) and decided I would try muting the messages, then try reloading the save and toggling SweetFX back on while the game reloaded. But then I discovered none of the SweetFX keys were working anymore. But the post-processing was showing up fine. So I guess those commented out lines actually were doing something? I don’t know, I don’t care I managed to “fix” the problem at least. Ay!

Getting the shot from onboard the aircraft was relatively simple after all that thankfully.

The ground shot BTW was done with CameraTools using a manual offset which took a good deal of trial and error plugging in values and activating it at the right time until I realized I could just use AirPark to freeze the aircraft in place from where I wanted to reference the camera position. That did weird things to the height value but really all I was interested in was the Forward and Right offsets – the height was easy to determine just based on the aircraft’s altitude. The camera really was positioned right near KSC from that angle.

All told the airlaunch mission took a good 8-10 hours of work, at least,  to complete the mission itself and all the mission data and assets.

As a final note – BACK UP YOUR DATA!!! I can’t stress enough how valuable CrashPlan is. The fact that it can store revisions and deleted files is key. I opened up the rocket telemetry data CSV file after editing it before and it was just all the data columns condensed into rows of a single column with all the delimiters removed. It would have been an enormous pain in the ass to reconstruct. Thankfully tho my backups had stashed the original version of the file I was able to recover in like a minute. This is far from the first time backups have saved me time and effort like this.