totm january 2020 Community Caveman Jool 5 mission

[IncongruousGoat]

@ManEatingApe

KILO launch: https://imgur.com/a/hU6aFug

Tylo landing: https://imgur.com/a/N0b0kz9

Jool->Kerbin transfer: https://imgur.com/a/Gv76004

[dvader]

Here are my contributions:

2xFOXTROT Launches
https://imgur.com/a/NfgvDOH

VICTOR KILO LIMA Setup
https://imgur.com/a/btKYyBe

LIMA - Laythe Landing and Rescue
Screenshots https://imgur.com/a/uZsrkKi
Video https://youtu.be/DV9TxOlBg2o

 

[LittleBitMore]

This sounded amazing! If only I was available at the time to help. I joined after this finished. I hoped it was as entertaining for you guys to proceed with as it was for me to observe. This sounded like quite the challenge. Great job!

[ManEatingApe]

1 hour ago, LittleBitMore said:

This sounded amazing! If only I was available at the time to help. I joined after this finished. I hoped it was as entertaining for you guys to proceed with as it was for me to observe. This sounded like quite the challenge. Great job!

Welcome aboard!
Glad you enjoyed it, we would have loved to have you contribute!

The entire mission history is preserved in the GitHub repo, so if you want you can try any of the missions yourself e.g chance your arm on Tylo or give Laythe a shot.

I'd heartily recommend trying the KSP Caveman Challenge which is still live.

[LittleBitMore]

8 hours ago, ManEatingApe said:

Welcome aboard!
Glad you enjoyed it, we would have loved to have you contribute!

Thanks! 

8 hours ago, ManEatingApe said:

The entire mission history is preserved in the GitHub repo, so if you want you can try any of the missions yourself e.g chance your arm on Tylo or give Laythe a shot.

...Tylo? Or even Laythe? Haha, maybe I should just try one of the easier things like "hitting the launch button and then saying 'I did something productive!'".

In all seriousness, how did you do these things without patched conics? How did you get the SOI changes on point?

[Jacke]

1 hour ago, LittleBitMore said:

Thanks! 

...Tylo? Or even Laythe? Haha, maybe I should just try one of the easier things like "hitting the launch button and then saying 'I did something productive!'".

In all seriousness, how did you do these things without patched conics? How did you get the SOI changes on point?

It's a little bit harder than going to Mun (have enough delta-V, burn prograde just as or before Mun rises, work it to get the pass, refine the pass in Mun SoI).  Basically, it's going back to astronautics first principles.  Helps that we could take as much time as we need (no life support limits, solar power).  Others did the complex interplanetary stuff, so they could explain it better.  I just tooled around doing a couple launches, rendezvous and docking, and the initial return Kerbin Orbit Burn.

[LittleBitMore]

52 minutes ago, Jacke said:

It's a little bit harder than going to Mun (have enough delta-V, burn prograde just as or before Mun rises, work it to get the pass, refine the pass in Mun SoI).  Basically, it's going back to astronautics first principles.  Helps that we could take as much time as we need (no life support limits, solar power).  Others did the complex interplanetary stuff, so they could explain it better.  I just tooled around doing a couple launches, rendezvous and docking, and the initial return Kerbin Orbit Burn.

You make it sound so easy. For me, I'd probably have to quicksave and quickload about a hundred times before a reasonable intercept. Oh, jeez, I can't imagine what it's going to be like if I ever plot an interplanetary Caveman mission. I rely too heavily on patched conics and maneuver nodes.

(Also, odd how there's another sleepless person who's awake on Christmas Eve night. The reason I'm awake is I made the mistake of attempting to jump my sleep schedule from 1 AM sleep time to 7 PM sleep time, in one night... I got 1.5 hours of sleep before I woke up again. Why are you awake this late?)

Edited December 25, 2019 by LittleBitMore
I accidentally typoed 1.5 hours as 15... Whoops...

[Jacke]

52 minutes ago, LittleBitMore said:

You make it sound so easy. For me, I'd probably have to quicksave and quickload about a hundred times before a reasonable intercept. Oh, jeez, I can't imagine what it's going to be like if I ever plot an interplanetary Caveman mission. I rely too heavily on patched conics and maneuver nodes.

Only the practised things are relatively easy.  You can look back in the topic for me working through my last part, where I had to rendezvous the spacecraft with Jeb, change the periapsis pass, then perform the Kerbin Orbit Injection.  Except for the rendezvous (which I know very well), I needed a lot of reloads to get this done.

Patched conics and maneuver nodes are the way to package well doing astronautics.  It is possible to go back to first principles and do the maneuvers by hand.  Like all manual use of automatic procedures, there's a learning curve but you then understand the tool better.  I like Caveman as a challenge, but I would prefer to have the tools in my KSP play.

 

52 minutes ago, LittleBitMore said:

(Also, odd how there's another sleepless person who's awake on Christmas Eve night. The reason I'm awake is I made the mistake of attempting to jump my sleep schedule from 1 AM sleep time to 7 PM sleep time, in one night... I got 1.5 hours of sleep before I woke up again. Why are you awake this late?)

My sleep schedule has been rubbish for a while due to being hard to fall asleep, which I'm still working on.  And it's still just past 23h30 by my local clock.

[IncongruousGoat]

21 hours ago, LittleBitMore said:

You make it sound so easy. For me, I'd probably have to quicksave and quickload about a hundred times before a reasonable intercept. Oh, jeez, I can't imagine what it's going to be like if I ever plot an interplanetary Caveman mission. I rely too heavily on patched conics and maneuver nodes.

It was a learning process for all of us. But it's really not as bad as it looks, especially if you're willing to take a bit longer with the transfer than you might otherwise, and to throw a little bit of math at the problem. For the return to Kerbin (for example), I just warped to Pe without trying to set up a transfer, used Kepler's 3rd Law to compute my orbital period from Kerbin's period, squinted at the screen and made an educated guess at Kerbin's position, and used some math to work out a rough intercept on the next pass. From there it only took me 3 or 4 quickloads to narrow down the intercept to something reasonable. Plus or minus a couple inclination correction burns, but that's the general idea.

Of course, there's always what @Muetdhiver did for the transfer to Jool, which was to throw a lot of math at the problem and work out the precise orientation and duration of the transfer burn, as well as the needed correction burns. But explaining that in detail goes a fair bit beyond my knowledge of physics.

[Jacke]

7 hours ago, IncongruousGoat said:

Of course, there's always what @Muetdhiver did for the transfer to Jool, which was to throw a lot of math at the problem and work out the precise orientation and duration of the transfer burn, as well as the needed correction burns. But explaining that in detail goes a fair bit beyond my knowledge of physics.

The hardest part of that is if the Trans Joolean Injection burn out of Kerbin orbit is long due to low acceleration, quite common with fully loaded spacecraft.  Then to figure out the angle of the power track (the trajectory under thrust) around Kerbin, which angle being large and significant to get the point in orbit to start the burn, you really need to solve the Lagrangian of the maneuver (using Lagrangian mechanics.  It should be the Lagrangian equations of the second kind, in circular coordinates, with each coordinate equation (2 of them) equal to the acceleration in the particular coordinate due to engine thrust).

Where I learned this was from reading Werner von Braun's book The Mars Project (1953), where he shows and does an approximation solution for the Lagrangian of the Trans Mars Injection for his spacecraft.  Back around 1980, I coded a graphing program in Pascal to solve and draw out an example.  Still got that program someplace, on a deck of cards (it was a different time).

I don't know if @Muetdhiver did that level of solution, as a Caveman design usually has a significant reserve and he could have just calculated a more approximate angle with a more simpler method.

Edited December 26, 2019 by Jacke

[Muetdhiver]

I did something both simpler and more complicated : with the data available in caveman (alt, speed, clock time, planet ephemeris) I computed an approximate heliocentric elipse solution with Kepler equation shenanigans. Then I solved the Lambert problem for the jool encounter, then translated the solution in the ship frame, the turned DV into fuel units.

I got the initial launch angles and time from kerbin also with a Lambert solver from an ESA package. But I planed on the burn to be sloppy, which is why I went for the course correction  approach. 

This was made quite a bit harder by the Jeb pickup routine affecting the orbit and me doing the correction burn in the wrong direction for the radial component.

I must say it was loads of fun to figure it out.

Edited December 26, 2019 by Muetdhiver

[James Kerman]

Congratulations, in addition to winning thread of the year for your catapult challenge @ManEatingApe, this excellent challenge has been awarded thread of the month.

[ManEatingApe]

1 hour ago, James Kerman said:

Congratulations, in addition to winning thread of the year for your catapult challenge @ManEatingApe, this excellent challenge has been awarded thread of the month.

Thank you on behalf of @Muetdhiver @Jacke @Pds314 @IncongruousGoat @dvader @Rakaydos who made the thread and mission a success!

Shout out to contributors  - if you haven't already put all your photos into an Imgur album, please do so and I'll add it to the OP, so folks can see the entire mission in a single location.

Edited January 3, 2020 by ManEatingApe

[Muetdhiver]

Here are the Caveman launch, dock and transfer to Jool video as well as the Pol - Bop - Vall video.

https://youtu.be/vonjAX5r3_0

https://youtu.be/9UKmKqCkktw

It's a first for me, so no fancy stuff here. Funny thing is that I lose an engine at lauch XD.

 

[sevenperforce]

On 8/11/2019 at 9:23 AM, Muetdhiver said:

I have made bit of work on a return vehicle based on the Laythe lander from IncongruousGoat.

1) I fixed the ascent stability and found a nice way to boost overall DV from the launch. Simply put, my craft was too light and made the rockets to top over. Which is a classical problem. What I did is put an FLT400 on top with the nose cone, which is then ditched in space.

2) The lower stage plus top tank is nearly enough (or likely enough with a really good turn) to get in orbit. In short : the craft gets in LKO with a fully filled FLT400, which is neat.

3) The capsule does just fine on rentry, tested with AP = 1000km, kerbal was fine. I assume we'll do a personal chute landing soviet style to land => we'll have to level up the dude.

  Hide contents

 

 

Do you still have this craft file? My attempts to replicate it for a different challenge keep torching the Kerbal.

[IncongruousGoat]

2 hours ago, sevenperforce said:

Do you still have this craft file? My attempts to replicate it for a different challenge keep torching the Kerbal.

This isn't that exact one, but I've uploaded the original re-entry tube design from back during initial testing here. Again, not exactly the same, but if you're just looking for any old fairing-based re-entry vehicle it should do fine.

[sevenperforce]

16 minutes ago, IncongruousGoat said:

This isn't that exact one, but I've uploaded the original re-entry tube design from back during initial testing here. Again, not exactly the same, but if you're just looking for any old fairing-based re-entry vehicle it should do fine.

Tested. Definitely does not work. 

Spoiler

 

I wonder if they changed the code since that version.

[ManEatingApe]

2 hours ago, sevenperforce said:

Tested. Definitely does not work.

Dang. This challenge used version 1.7.3 so it looks like the loophole exploit cunning workaround stroke of genius (take your pick ) has been closed in later versions.

[IncongruousGoat]

Well, that's a bummer. Oh well. I suppose this is what we get for building things that exploit strange behavior.

[ManEatingApe]

15 hours ago, IncongruousGoat said:

Well, that's a bummer. Oh well. I suppose this is what we get for building things that exploit strange behavior.

The good news is that this mission is possible using propulsive braking only. In fact it's nearly possible using the existing save game and craft.

There are 2 situations in which aero-braking comes in useful:

1. Capturing at Laythe
2. Capturing at Kerbin when returning from Jool

For situation 1 @dvader used propulsive braking only when landing on Laythe, so no changes needed there.

For situation 2 we did originally use aero-braking to capture at Kerbin. However that was mostly to preserve dV in order to make a fancy-pants landing right on top of the VAB. The craft had plenty of dV left when entering the Kerbin system.

Out of curiosity I reverted to version 27 of the save game, just as KILO is entering the edge of Kerbin's SOI. Using only propulsive braking I was able to get Jeb into a 67x83km orbit at a velocity of ~2,300 m/s.

This is still too fast for Jeb to survive when re-entering in an EVA suit, but close! A few hundred dV more and it would be perfectly feasible. As a bonus it would simplify the design as the combined Tylo, Bop, Pol and Vall lander design could be re-used for the return craft.

Edited June 6, 2020 by ManEatingApe

[Muetdhiver]

It seems something changed w.r.t the way heat is processed in this edge case.

It looks like Kerbals on ladders now don't benefit from the fairing potection anymore. Very disapointing  (Though it was already really tricky to make it work before)

It would be interesting if anything could be done with a radiator or something...

This Jool5 was quite epic TBH.

I would be really impressed by a caveman Moho landing using a single or double grav assist from Eve.  

 

[ManEatingApe]

29 minutes ago, Muetdhiver said:

This Jool5 was quite epic TBH.

It most certainly was!

I had a crazy idea to make things even more epic...what do people think about the feasibility of doing this again but with CommNet ON?

Edited June 8, 2020 by ManEatingApe

[IncongruousGoat]

Just now, ManEatingApe said:

I had a crazy idea to make things even more epic...what do people  think about the feasibility of doing this again but with CommNet ON?

I don't think it's very feasible, TBH. All of our landers would need to have crew capsules for control (CommNet on -> no probe control out at Jool), which just by the rocket equation would give us a minimum of ~7x overall vehicle mass (.60t for a lander can vs. .09t for a Kerbal on a ladder). It's certainly possible to design a ship that would work, but it would be an absolute mess of pad and orbital assembly to put together, and it would probably end up being painful to fly due to low control authority, wobbly joints, and low-tech solar panels.

I'd love to be proven wrong on that though.

[JAFO]

11 hours ago, ManEatingApe said:

I had a crazy idea to make things even more epic...what do people think about the feasibility of doing this again but with CommNet ON?

The word "impossible" comes to mind. Unless CommNet has been massively altered since I was last playing, a couple years back.

About 3 years back, I did some experiments to see how feasible it was to "stack" antennas for extended control range in Caveman missions. To quote myself from back then...

Quote

In theory, it's possible to combine enough HG-5 antennas on a probe to contact Kerbin from other planets, so I set myself the task of determining if the quantity necessary was practical to achieve.. obviously it would need a number of comms modules docked together in orbit, the question was, how many modules would be needed?

In a sandbox game, I placed 8 HG-5s around an OKTO core, then duplicated and stacked the setup as many times as I thought would be necessary. It turned out that with a Level 3 Tracking Station, and Eve at close approach to Kerbin, it took 88 HG-5s to get an 11% signal. I estimate that at very closest approach, that might go up as high as 15%. Some rough calcu-guess-timations suggest that with a Level 1 Tracking Station, something on the order of 500-1000 HG-5s would be needed to establish the same connection. And even then, Eve would be out of range for around 90% of its orbit!

As for Duna.. well, let's just forget about that.

So, for Cavemen considering Eve/Duna/beyond missions, there really is no option but crewed missions.

 

Edited June 9, 2020 by JAFO

[Pds314]

2 hours ago, JAFO said:

The word "impossible" comes to mind. Unless CommNet has been massively altered since I was last playing, a couple years back.

About 3 years back, I did some experiments to see how feasible it was to "stack" antennas for extended control range in Caveman missions. To quote myself from back then...

 

Is there any viability to the idea of relaying several HG-5s or groups thereof back to Kerbin so that they all align perfectly when the mission reaches Jool? It'd probably take a pretty huge number of them but it might be cheaper, if considerably harder to set up, than a direct commlink to Jool?

Hmm... Doing the math it seems to me like a craft with 64 HG-5s would have a signal power of 113 M. This is not particularly efficient, with only 22 times the power of a single unit, but I'd rather have fewer satellites that are less efficient if possible so that they don't all need to load at once.

So this to me would indicate that it *should* have a laughably microscopic communication range to another commsat. Just 113 Megameters. That means you need hundreds of relaysats precisely aligned on a straight path to Jool at a precise time.

So...
1 relay, at Jool, would need 50 million HG-5s in one craft.
2 relays, one at Jool and the other at Kerbin, would need two crafts with 400 thousand relays each. (Basically the extra 400000 relays replace the tracking station with an 80 G transmitter)
100 relays in a chained system going straight to Jool with perfect orbital alignment would need about 1000 HG-5s per satellite.
Make it 1000 relaysats and it's about 40 HG-5s per sat.
If you want singular HG-5s for each relaysat, you're gonna need 5 Megameters spacing between them, so about 16,000 links in the chain.

Edited June 9, 2020 by Pds314