Some Disassembly Required

[Snark]

So, a completely silly and not super practical idea that occurred to me was to try designing a crewed vehicle that does a MIRV-style reentry:  it disassembles into multiple small parts, which then independently (and simultaneously) reenter and land.

There was a nugget of practical thinking there:  aside from the 16-kerbal Mk3 monster (well, okay, and the command chair), the Mk1 crew cabin has the best mass-per-kerbal of any crewed part.  In principle, a stack of them ought to make a reasonably mass-efficient crew carrier.  The problem is that such a vehicle has a very high ballistic coefficient, and reentry can be awkward.  Too much mass, piled up behind too little surface area.  So hey, what if it disassembled so that it has a whole lot more area to use?  How would it do?

But who am I kidding?  This wasn't for practical reasons at all.     I just wanted to see, "would it work", how the game would handle it, and (most importantly) would the components stay close enough to each other that the game wouldn't delete any of them out from under me.

Anyway, here's the tale of my adventure.  Invididual screenshots below, or you can see the imgur album.

 

Preparing to separate

Here's the ship, suborbital, shortly before separation.  (Before anyone brings it up-- and I know you will, don't bother denying it-- no, I'm not a noob who put an LFO tank on a ship powered by LV-Ns.  That's a modded LF-only tank.)

It holds 10 kerbals.  There's space for 11, but I ditch the Mk1 command pod on the way down, so one of the crew cabins has an empty seat so the pilot has a place to transfer to.  The idea is that the whole ship plows straight into atmosphere, stays whole as far down as it can stand, then activate the final staging action which blows all the separators and activates all the parachutes, all in one fell swoop.

Clever, huh?

 

Oops #1

Well, not quite as clever as I thought.  I blithely designed the ship without a probe core ("no problem, I actually have a pilot on board, for once!")... and like a total idiot, I didn't think of the fact that that made my original mission plan impossible.

If I transfer Jeb out of the command pod into a crew cabin... that leaves the ship with no control at all (since cabins don't have control authority).  Which means I can't activate the final stage.  So the ship can't decouple and the whole experiment is ruined.

Okay, Plan B.  (I gotta say that "figuring out Plan B" is one of my favorite things about KSP.  Yah, I'm that guy.)  I'll just decouple the ship while it's still in vacuum on a suborbital trajectory, then Jeb can quickly EVA across to his waiting cabin seat in Super Duper Economy Class, just seconds before hitting atmosphere.

It worked.  Yay!  The only concern, though, is that now I've been and gone and decoupled while still out in space, and there's a long way to go before touchdown.  So I'm a little worried about how much dispersion I'll get, and whether they'll end up separated enough that the game can't keep track of them and they get deleted out from under me.  We'll see what happens.

 

Oops #2

I went and flew the entire mission after that, all the way to landing, and discovered a problem:  I decoupled while the ship was aligned .

Why did I do it, and why was it a problem?

Well, it seemed like a sensible move, since it had all the cabins nicely aligned with the heat shield facing forward.  As you can see, each separate reentry vehicle is just a Mk1 crew cabin, a 1.25m heat shield, and a couple of radial parachutes.  I had already done a test in atmosphere that satisfied me that they're aerodynamically stable with the heat shield facing .  (This is important, because after separation, they're completely uncontrolled and have no reaction wheels or SAS or anything.  They're completely at the mercy of aerodynamic stability.)

I wanted to do it that way because I wasn't sure how well the fragile, melty cabins could handle the early part of reentry if they're not properly aligned:  what if they overheat and explode before aligning themselves with the heat shield facing forward?  So separating while already aligned just seemed to make sense.

The problem was the separators.  Those things are really lightweight, and also high-drag.  So as soon as the fragments enter atmosphere, those separators want to drift relative to the cabins.  What I found was that a couple of them got "pinned" by aero forces against the front end of a couple of the cabins... but the others tumbled off and were lost.  This now caused the cabins that had the separators pinned against them to decelerate harder than the ones who didn't.  It was very gentle at first, but became more pronounced as we descended, so that by the time it got down to landing, there was a 10-kilometer gap. 

"But wait!" I hear you cry.  "That's still well inside the atmospheric 'bubble' for going on rails, that shouldn't be a problem!"

Well, it seemed not, just about all the way down.  But as soon as the ships I was focused on touched down, the game suddenly decided that the other ships crashed and were lost. 

Okay, so that didn't work.  Fortunately, I had saved before reentry, so I tried the whole thing again.  This time, I aligned the ship instead of , so that the line of decoupled cabins was traveling sideways relative to the direction of travel.  My hope was that this would leave the separators with a clear "escape path".

(And lest someone ask "well, why didn't you just design it with decouplers rather than separators, so this entire problem would be avoided?",, there are two answers:  1. I didn't anticipate this problem, and 2. I just didn't wanna.  I thought the spent decoupler was ugly and I didn't want it stuck to my nice clean reentry pod.)

 

Separating the wheat from the chaff

...and hey, looky there, it worked!    Here you can see the separators quickly falling behind.  You can also see that the cabins quickly aligned themselves with the heat shields facing . 

All my ducks in a row. 

 

Reentry

Worked like a charm!  No overheating danger at all; never saw so much as an overheat bar on any of the cabins.

It was fun, watching them all come down in formation.  They stayed nicely clustered.

Flames gone, decelerating past Mach 1.

 

Landing

Uneventful landing, just your standard parachutes-deploy-and-then-touch-down sequence.  Really fun to watch, though. 

Almost there...

Aaaaaand.... ta dah!

Landed and safe.  The maximum separation between two cabins was 226 meters.  Four of them were within 100 m!  I thought that was remarkably precise, given that they separated while still outside of atmosphere going nearly orbital speed.  The game physics turned out to be remarkably consistent; they stayed clustered all the way down, and touched down within a couple of seconds of each other.

Anyway, silly and impractical-- I doubt I'll make a habit of this techique.  But it sure was a fun experiment. 

[KSK]

That's a whole lot of orbital tourists coming home.  Neat - and those screenshots of formation drop pods are sweet.

[Hotaru]

14 minutes ago, Snark said:

Anyway, silly and impractical-- I doubt I'll make a habit of this techique.  But it sure was a fun experiment. 

I don't know--they look like they'd make great little escape pods. Extremely cool. I wonder how they'd handle reentry at higher speeds.

[Snark]

18 minutes ago, KSK said:

That's a whole lot of orbital tourists coming home.

Actually, it wasn't tourists.  It was my entire space program, all 10 kerbonauts, whom I'd sent on a quick jaunt outside of the local SoI so they could pick up "orbit the Sun" experience points.

18 minutes ago, KSK said:

Neat - and those screenshots of formation drop pods are sweet.

It was just really fun to watch.  For example, before the parachutes fully opened, they were slightly vertically dispersed (the lowest one a few dozen meters lower than the highest one).  But when the chutes opened, they suddenly slammed to nearly the exact same height (since the bottommost one opens first), and suddenly they were all within a couple of meters of the same height.  That was cool. 

17 minutes ago, Hotaru said:

I wonder how they'd handle reentry at higher speeds.

Not super well.  An earlier attempt had them hitting atmosphere at ~2800 m/s, and they did get heat bars, and one of them ended up exploding (while still wobbling side-to-side) before they'd either stabilized enough to avoid exposing the cabins to the airstream, or slowed down enough not to fry.  Though to be fair, that was with a lowish 30 km Pe.  Something as lightweight and draggy as these pods can probably make do with a higher Pe.  (And of course, I could have helped them along by adding airbrakes, if I had wanted to, but I wanted to keep it simple.)

Fortunately, my ship still had a fair amount of dV left, so I tried again and this time used the engine to slow down to LKO before reentering.  Worked a lot better.

Or... if I'd not been such a total doofus, and put a probe core on the ship so I could have stuck to my original plan, then I may have fared better-- i.e. don't decouple until I get down inside the atmosphere reasonably deep.

[goldenpeach]

Snark, if you ever start a private space tourism company, I am not going to be your first client. I don't want to discover just before reentry that you used a similar design.

[Snark]

3 minutes ago, goldenpeach said:

Snark, if you ever start a private space tourism company, I am not going to be your first client. I don't want to discover just before reentry that you used a similar design.

Awww, you wouldn't have to worry.  It would be guaranteed:  safe landing, or your money back!  (Must apply in person.  Additional restrictions may apply.  Not responsible for personal items or body parts left in cabin.)

[Geonovast]

1 hour ago, Snark said:

2. I just didn't wanna.  I thought the spent decoupler was ugly and I didn't want it stuck to my nice clean reentry pod.)

A decoupler on the top would help with alignment though, right?

Also it bothers me that you covered up some of the windows on your nice clean reentry pod

[KSK]

29 minutes ago, Geonovast said:

A decoupler on the top would help with alignment though, right?

Also it bothers me that you covered up some of the windows on your nice clean reentry pod

That's for crew safety. What they can't see won't hurt them.

[Snark]

27 minutes ago, Geonovast said:

A decoupler on the top would help with alignment though, right?

Hard to say, without testing.  I expect it would add hardly any drag at all.  It would help to move CoM very slightly forward, I think (it has 1/20th the mass of the cabin, but more than 1/20th the length), but I'd expect that to be a fairly small effect compared to, for example, adjusting the position of the parachutes.

29 minutes ago, Geonovast said:

Also it bothers me that you covered up some of the windows on your nice clean reentry pod

Yeah, me too.  But the only other option would be to put them on the top/bottom of the pod, and that would have obstructed the crew indicator, and obviously we can't have that...

[GarrisonChisholm]

This reminds me of my very first multi-kerbal re-entry, which as a rookie I didn't realize would be a problem if they drifted apart.  Consequently, it was both my first kerbal fatality and my first muti-kerbal fatality.  >,<  ...I think that was also my very first post here?  ...heavens, bet that's impossible to find now...

Good Plan-B-ing though @Snark!

[Snark]

41 minutes ago, GarrisonChisholm said:

...I think that was also my very first post here?  ...heavens, bet that's impossible to find now...

No, not really. 

(that's "no" to both of your comments quoted here)

 

[Just Jim]

hehehe... that was great! 

[Geschosskopf]

Hehehe, funny concept, Snark!  I"m a bit disappointed you only killed 8 Kerals, though.  This had such potential .....  

[GarrisonChisholm]

21 hours ago, Snark said:

No, not really. 

(that's "no" to both of your comments quoted here)

 

LOL  Snark, you just made (abetted at least) me 'Like' a comment in a 2 1/2 year old thread lol.

But I was right at least in remembering a multi-capsule re-entry attempt!  Small victories...

[linuxgurugamer]

 

So, a completely silly and not super practical idea that occurred to me was to try designing a crewed vehicle that does a MIRV-style reentry:  it disassembles into multiple small parts, which then independently (and simultaneously) reenter and land.

There was a nugget of practical thinking there:  aside from the 16-kerbal Mk3 monster (well, okay, and the command chair), the Mk1 crew cabin has the best mass-per-kerbal of any crewed part.  In principle, a stack of them ought to make a reasonably mass-efficient crew carrier.  The problem is that such a vehicle has a very high ballistic coefficient, and reentry can be awkward.  Too much mass, piled up behind too little surface area.  So hey, what if it disassembled so that it has a whole lot more area to use?  How would it do?

But who am I kidding?  This wasn't for practical reasons at all.     I just wanted to see, "would it work", how the game would handle it, and (most importantly) would the components stay close enough to each other that the game wouldn't delete any of them out from under me.

Anyway, here's the tale of my adventure.  Invididual screenshots below, or you can see the imgur album.

 

Preparing to separate

Here's the ship, suborbital, shortly before separation.  (Before anyone brings it up-- and I know you will, don't bother denying it-- no, I'm not a noob who put an LFO tank on a ship powered by LV-Ns.  That's a modded LF-only tank.)

It holds 10 kerbals.  There's space for 11, but I ditch the Mk1 command pod on the way down, so one of the crew cabins has an empty seat so the pilot has a place to transfer to.  The idea is that the whole ship plows straight into atmosphere, stays whole as far down as it can stand, then activate the final staging action which blows all the separators and activates all the parachutes, all in one fell swoop.

Clever, huh?

 

Oops #1

Well, not quite as clever as I thought.  I blithely designed the ship without a probe core ("no problem, I actually have a pilot on board, for once!")... and like a total idiot, I didn't think of the fact that that made my original mission plan impossible.

If I transfer Jeb out of the command pod into a crew cabin... that leaves the ship with no control at all (since cabins don't have control authority).  Which means I can't activate the final stage.  So the ship can't decouple and the whole experiment is ruined.

Okay, Plan B.  (I gotta say that "figuring out Plan B" is one of my favorite things about KSP.  Yah, I'm that guy.)  I'll just decouple the ship while it's still in vacuum on a suborbital trajectory, then Jeb can quickly EVA across to his waiting cabin seat in Super Duper Economy Class, just seconds before hitting atmosphere.

It worked.  Yay!  The only concern, though, is that now I've been and gone and decoupled while still out in space, and there's a long way to go before touchdown.  So I'm a little worried about how much dispersion I'll get, and whether they'll end up separated enough that the game can't keep track of them and they get deleted out from under me.  We'll see what happens.

 

Oops #2

I went and flew the entire mission after that, all the way to landing, and discovered a problem:  I decoupled while the ship was aligned .

Why did I do it, and why was it a problem?

Well, it seemed like a sensible move, since it had all the cabins nicely aligned with the heat shield facing forward.  As you can see, each separate reentry vehicle is just a Mk1 crew cabin, a 1.25m heat shield, and a couple of radial parachutes.  I had already done a test in atmosphere that satisfied me that they're aerodynamically stable with the heat shield facing .  (This is important, because after separation, they're completely uncontrolled and have no reaction wheels or SAS or anything.  They're completely at the mercy of aerodynamic stability.)

I wanted to do it that way because I wasn't sure how well the fragile, melty cabins could handle the early part of reentry if they're not properly aligned:  what if they overheat and explode before aligning themselves with the heat shield facing forward?  So separating while already aligned just seemed to make sense.

The problem was the separators.  Those things are really lightweight, and also high-drag.  So as soon as the fragments enter atmosphere, those separators want to drift relative to the cabins.  What I found was that a couple of them got "pinned" by aero forces against the front end of a couple of the cabins... but the others tumbled off and were lost.  This now caused the cabins that had the separators pinned against them to decelerate harder than the ones who didn't.  It was very gentle at first, but became more pronounced as we descended, so that by the time it got down to landing, there was a 10-kilometer gap. 

"But wait!" I hear you cry.  "That's still well inside the atmospheric 'bubble' for going on rails, that shouldn't be a problem!"

Well, it seemed not, just about all the way down.  But as soon as the ships I was focused on touched down, the game suddenly decided that the other ships crashed and were lost. 

Okay, so that didn't work.  Fortunately, I had saved before reentry, so I tried the whole thing again.  This time, I aligned the ship instead of , so that the line of decoupled cabins was traveling sideways relative to the direction of travel.  My hope was that this would leave the separators with a clear "escape path".

(And lest someone ask "well, why didn't you just design it with decouplers rather than separators, so this entire problem would be avoided?",, there are two answers:  1. I didn't anticipate this problem, and 2. I just didn't wanna.  I thought the spent decoupler was ugly and I didn't want it stuck to my nice clean reentry pod.)

 

Separating the wheat from the chaff

...and hey, looky there, it worked!    Here you can see the separators quickly falling behind.  You can also see that the cabins quickly aligned themselves with the heat shields facing . 

All my ducks in a row. 

 

Reentry

Worked like a charm!  No overheating danger at all; never saw so much as an overheat bar on any of the cabins.

It was fun, watching them all come down in formation.  They stayed nicely clustered.

Flames gone, decelerating past Mach 1.

 

Landing

Uneventful landing, just your standard parachutes-deploy-and-then-touch-down sequence.  Really fun to watch, though. 

Almost there...

Aaaaaand.... ta dah!

Landed and safe.  The maximum separation between two cabins was 226 meters.  Four of them were within 100 m!  I thought that was remarkably precise, given that they separated while still outside of atmosphere going nearly orbital speed.  The game physics turned out to be remarkably consistent; they stayed clustered all the way down, and touched down within a couple of seconds of each other.

Anyway, silly and impractical-- I doubt I'll make a habit of this techique.  But it sure was a fun experiment. 

So I've been trying this, and ran into a problem.

When I recovered the first tourist, and then went to get the next, they all seem to have disappeared, even though the cockpits (I was using the Mk1 inline cockpit) were still there.

Any ideas?

[RedPandaz]

What this sounds like is my solution to Career Tourist contracts- I make a stack of mk1 command pods on a rocket, each with a parachute attatched radially, and stick it on a rocket XD

[Snark]

 

When I recovered the first tourist, and then went to get the next, they all seem to have disappeared, even though the cockpits (I was using the Mk1 inline cockpit) were still there.

Any ideas?

Wait, so they were inside your cockpits and they just vanished without a trace?

That's... bizarre.  Have they hired Stephen King as a developer?  I mean, seriously, you've just precisely re-created The Langoliers.

My guess would be that it's some kind of a bug, maybe the game treats tourists specially somehow?  My own experiment didn't involve tourists-- all of the kerbals involved were astronauts.

Is it reproducible?  What happens if you use astronauts rather than kerbals?  If it's reproducible, sounds like a good candidate for a bug report.

[linuxgurugamer]

13 hours ago, Snark said:

Wait, so they were inside your cockpits and they just vanished without a trace?

That's... bizarre.  Have they hired Stephen King as a developer?  I mean, seriously, you've just precisely re-created The Langoliers.

My guess would be that it's some kind of a bug, maybe the game treats tourists specially somehow?  My own experiment didn't involve tourists-- all of the kerbals involved were astronauts.

Is it reproducible?  What happens if you use astronauts rather than kerbals?  If it's reproducible, sounds like a good candidate for a bug report.

I'll retry it with astronauts in a day or so when I have time.