Welcome to Jester's Trek.
I'm your host, Jester. I've been an EVE Online player for about six years. One of my four mains is Ripard Teg, pictured at left. Sadly, I've succumbed to "bittervet" disease, but I'm wandering the New Eden landscape (and from time to time, the MMO landscape) in search of a cure.
You can follow along, if you want...

Friday, February 28, 2014

We feel her day is over

EVE players can move on to the next EVE-related post, but I wanted to do a pair of recap and wrap-up posts on my Duna mission in KSP over the weekend. They both illustrate some interesting KSP points that I wanted to talk about.

First up, a final note on Enterprise's disposition. Enterprise went to Duna with seven crew on board: three in the command capsule, two in the science lab, two in the Orion lander. That was clearly excessive but I didn't see any particular reason not to do it. I find the necessity of using an EVA to switch between vehicles an annoyance within KSP and I generally try to keep them to a minimum. But it did mean that when the Orion lander set down on Kerbin, I still had five crew on Enterprise in orbit (including Jeb and two other kerbals I like to use, "Danger" and "Gus"). How to get them home?

I could send a space-only "shuttle" over from Liberty station to pick up half of them and a launch from the surface for the others. But I wanted to keep things simple if I could and get all five home in a single mission. I considered launching a "Hitchhiker" can attached to a command pod but then realized I could do the same job with a Mk2 Lander attached to a command pod, so that's what I did. I put the combination plus a couple of extra tanks of fuel and monoprop, on top of my standard Athena launcher and sent them up.

Part of the plan was to fully evacuate Enterprise but still be able to control her. There was a "senior" docking port on top. So at the top of the Athena launch, I attached a large size probe core with docking ports on either side. I could dock the combined space craft to Enterprise, then when I left, leave the probe core docked to Enterprise and control her with that. This is a trick that I use now and again and it always works beautifully. One probe core with docking ports on either side can control pretty much any space craft with an open docking port. You just have to make sure all of the docking ports face the correct way, plug it in, then leave it there when you undock.

Most ships I've built since Enterprise have a probe core somewhere on them as a matter of course so I can fly them unmanned if needed.

Aerobraking had left Enterprise in an elliptical orbit (p: 120km, a: 475km) with an unusual inclination (about 70 degrees). She'd come home with insufficient fuel to fully correct either of these situations. So this was the trickiest rendezvous in my KSP career to date. I matched inclination first, then found the proper moment to do a prograde burn to intercept. This left me in a position to swing very close to Enterprise (about 2km) but my rendezvous would not be near periapsis or apoapsis. That meant having to do a radial burn at rendezvous, the first time I've ever had to do that.

Think of an elliptical orbit as having its apoapsis at a "clock" position. In this case, Enterprise's apoapsis was at about 11 o'clock relative to the prograde direction of kerbin's orbit. My intercept location was at about 3 o'clock with an eventual apoapsis (if I left my orbit uncorrected) of 9 o'clock. A radial burn is used to swing the apoapsis "clock position" of an orbit from one clock position to another. They're very energy-intensive burns; this one required almost 500m/s of delta-v. But I'd used my heavy lifter to get up there so I had plenty of fuel.

The maneuver tool and its rarely used blue radial markers was instrumental for setting this up; once I was close enough, the friendly "Target" orbital indicators popped up and I did the final rendezvous burn with that, just like normal. I love the maneuver tool so much!

After that, docking was straight-forward. I transferred most of the fuel I had left so Enterprise would have enough to maneuver. I then made the appropriate joke from Star Trek III...
Scotty: All systems automated and ready. A chimpanzee and two trainees could run her.
Kirk: Thank you, Mr. Scott. I'll try not to take that personally.
The lander/command pod combo didn't need much fuel; I've gotten into the habit of sometimes using monoprop to adjust the orbit of light vehicles, just like the U.S. Gemini command pod did. I only carry just enough fuel for emergencies. Once the crew were safely off, I used the probe core to shift Enterprise's orbital inclination to 90 degrees and circularized the orbit at around 175km. Then I left her there in case I need her at some point in the future.

Enterprise with her new automation system (at picture bottom)

One more amusing note; I usually don't use KSP's quick save feature but since I was going to be landing an untested vehicle for the first time, this time I did. That turned out to be a good decision. I took 100 liquid fuel (the final stage of the space craft had a "Poodle" engine just in case), and did my de-orbit burn aiming for a water landing just east of the KSC. I came through the atmosphere with no problem, and deployed the six chutes I had attached to the lander. And at 500 meters above the water at around 110m/s, the chutes popped open... and the command pod was forcefully flung away from the lander.

This was kerbal engineering at its finest. Sigh. I hate this game sometimes. ;-)

I've run into this problem a couple of times before trying to land heavy vehicles: the shock of the chutes opening causes parts to fall right off! But this was the first time that I'd had this happen with two parts mounted in line. Rather than start from scratch, I used a technique that I've had luck with in this situation before. Instead of ditching the final fuel tank and the Poodle after the de-orbit burn was done, I reloaded from quick save and kept them all the way through to 10000m. Then I deployed the parachutes, waited until about 2000m above the water, locked the engine's gimbals, then fired the engine to slow my descent rate. As long as you don't reduce falling speed below 0m/s, the parachutes will continue to function.

At 600m, my falling speed was around 40m/s instead of 110m/s. Only at that point did I cut the engine and fuel tank loose. The command pod/lander accelerated back up to around 55m/s but when the chutes opened at 500m, the thing happily didn't break apart. I recovered Jeb, Danger, Gus, and their two flunkies successfully.

So kind of a long story for a simple outcome: just get five guys out of orbit all at the same time. But hopefully there's some lessons learned there for anyone else playing this game. Next up: my notes from the Duna mission itself.


  1. Your craft will tend to fall apart if the acceleration between parts gets too high - this is both true for rocket design and when aerobraking. Basically, if all of your parachutes are on the retrograde-most part, and you're going fast enough for the deployment of the parachutes to exceed the structural g-force limit, the "weld" will break and your ship will have a bad time.

    There are a couple of solutions to that problem:
    - Reduce acceleration: when the parachutes deploy (what you did); most craft will survive the transition from ~100 m/s to 10 m/s. If you're going too fast you can either fully deploy the parachutes earlier (where the air is thinner), and/or you can use propulsion to directly (and inefficiently, but quickly) reduce your velocity.
    - Spread the parachutes: distribute the force of acceleration evenly across the craft; the fewer non parachuted parts attached to a parachuted part, the better.
    - Add more parachutes: more parachutes means more drag means more mass reduced to acceptable acceleration parameters when the parachutes deploy.
    - Add struts: struts count as extra welds and will increase the amount of force required to break or bend two parts.
    - Take a more shallow descent path: the longer you spend in the atmosphere, the more time the partially deployed parachutes have to reduce your velocity.

    Some of those can be applied in situations where you can't change the design of your craft.

    For rocket design, this is the reason high Thrust to Weight Ratio rockets can disintegrate in the low atmosphere if you keep pushing the drag coefficient; the drag force on the top fights the thrust force on the bottom and any props in-between are very likely to spontaneously explode as they are pushed against one another.

    You can get a hint for how much punishment a part can receive in terms of acceleration in the info tab for said item (impact tolerance or something along those lines, in m/s); if you exceed that in terms of acceleration, you should probably think carefully about the design.

    1. Really good tips, thanks! I particularly like the idea of deploying the parachutes earlier for this kind of application. Struts wouldn't have helped in this situation because the parts were mounted in-line. I also like the info about looking more at the stats for parts for situations like this. I tend to go much more with gut feel and instinct.

    2. You're welcome!

      On the topic of struts, you can actually use them in-line to reinforce a regular part weld. Struts are far more rigid than regular welds, so that increases the tolerance between those parts. Strutting the bottom-most and top-most part together will "suspend" all the parts in-between, and the craft won't wobble nearly as much and will generally handle acceleration much better.

  2. I'll be glad when you finally get fed up with KSP and get back to what you do best-EVE.

    1. Sorry about that. ;-)

      But remember that this blog does venture into other games from time to time. Black Prophecy, Perpetuum Online, Global Agenda, Guild Wars 2, et cetera.

  3. Jester, you can ease the stresses of parachuting by not placing 1 large group of parachutes, but multiple smaller groups. I.e. instead of 1 group of six parachutes, use three pairs of them. You can then tweak them that the chutes don't deploy all at once, but open in pairs.

    1. Hm, that's very interesting! I'll give it a try.


Note: Only a member of this blog may post a comment.