Face to Face with The Tyranny

[GarrisonChisholm]

So, in my game I have as my far distant-dream target Valentine, the red-dwarf from the Extrasolar mod (which seems to work fine on my install btw).  However, I have *not* installed any interstellar mods, and wanted another star around more just to represent what the closest other neighbor in space would be.  Well after playing a while I decided to attempt to send something out that way, just to have a try.

The famous top speed video (without breakage or cheats) I think illustrates a ship hitting 80,000 m/s.  I figured if I could get a ship to 300,000 m/s it would take only 1,000 years (3400 Kerbin years) to get there at 9.5 trillion km away.  So, I could launch something, play my Kerbol system exploration out, and then before restarting I would just leave it run at max warp for a few days to get there.  300,000 is only 4.5 times faster than 80k, so that had to be doable, right?

Wrong!

The first stage sported 17k dV, and I thought "no problem!"  The second stage only lifted it to 25k dV, and I thought "hmm..."  The third stage- well, you know where this is going.  Its not just that successive stages have to be Bigger, its that each stage does less and less *work* per ton.  I know this is what the math says, and I know it has been discussed frequently & intelligently in these forums, but until you come face to face with the diminishing returns of the Equation... its just hard to realize how tyrannical it can be.

That being said, if anyone has a craft-file for something approaching 100,000 dv, I'd love to see it-  if only to pick it apart and have its engineering inform my own decisions.

I will upload pictures if I am satisfied with my final result, but- if we don't keep trying to design impossible things, how will the magic happen?  

 

[Gaarst]

A stage powered by a massless nuke, with no payload and using the highest full/wet masses ratio (9) tanks will only have a maximum delta-V of about 17.2km/s, even with infinite amounts of fuel. (Applying the same reasoning to a ion powered ship yields about 36km/s).

If you do it with the smallest amount of stages (6, so 16.667km/s per stage), with only one nuke per stage and a 10kg probe at the top (no decoupler, no structural stuff, just tanks, engines and a probe), here is what you 100k dV rocket would look like stage by stage:

Stage number	Contents	Stage mass (t)	Stage fuel mass (t)	Rocket mass (t)	Stage dV (km/s)	Stage TWR	Stage burn time
0	Probe core	0.01	0	0.01	0	0	0
1	1 Nuke + fuel tanks	316	278	316	16.667	1.94e-2	10.1h
2	1 Nuke + fuel tanks	33,434	27,717	33,750	16.667	1.81e-4	180 Kerbin days
3	1 Nuke + fuel tanks	3,540,494	3,147,103	3,574,244	16.667	1.71e-6	44.7 Kerbin years
4	1 Nuke + fuel tanks	374,918,264	333,260,676	378,492,508	16.667	1.62e-8	4736.3 Kerbin years
5	1 Nuke + fuel tanks	3.97e+10	3.53e+10	4.01e+10	16.667	1.53e-10	501,544 Kerbin years
6	1 Nuke + fuel tanks	4.20e+12	3.74e+12	4.24e+12	16.667	1.44e-12	53,110,706 Kerbin years

You're ending up with a rocket weighing 4.24e+12 tons, that is 4.24 trillion tons (or about 700,000 Great Giza Pyramids or 1.3 billion Saturn V). And your burn will last about 15million Earth years, assuming no physics lag.

(Note that this might not be the lightest way to do it, reducing dV requirements per stage will exponentially reduce their mass, and the increased mass because of more stages will probably be compensated by the reduced mass of each stage, I'm too lazy to do the maths though)

 

Now, let's see the what an ion powered rocket would look like:

Let's take 33.333km/s of dV per stage, giving us the 100km/s dV in 3 stages only, keeping a single engine per stage and a 10kg probe at the top.

Stage number	Contents	Stage mass (t)	Stage fuel mass (t)	Rocket mass (t)	Stage dV (km/s)	Stage TWR	Stage burn time
0	Probe core	0.01	0	0.01	0	0	0
1	1 ion engine + fuel tanks	15.96	8.86	15.97	33.333	1.28e-2	8.45 Kerbin days
2	1 ion engine + fuel tanks	995.5	561.2	1,011.5	33.333	2.02e-4	1.26 Kerbin years
3	1 ion engine + fuel tanks	62,098	35,017	63,110	33.333	3.23e-6	78.4 Kerbin years

This definitely looks more reasonable, our rocket is not "only" 63,110 tons heavy (or only one hundreth of a Great Giza Pyramid or 21 Saturn V), furthermore, our burn is reduced to a more reasonable 20 Earth years.

 

Conclusion: use warp drive, and start thinking about overthrowing the tyrant.

[Norcalplanner]

The best I could do is just under 32,000 m/s with Buster. It was a challenge to see how fast you could go before leaving Kerbin's SOI, so the thrusts are all above 0.5. This was back in 0.90, so not much concern was given to aerodynamic considerations.

 

 

Edited July 8, 2016 by Norcalplanner
Added version

[Archgeek]

I've been working on a small monster of an ion craft that'll get a ~5 tonne payload 47.6km/s in 11 stages for the low cost of only 107 tonnes, or 75km/s in 18 for 389 if I push my spreadsheet to its limits.  (The drive system sans-fuel-pod is around 7 tonnes on its own, and has another 2km/s onboard it.)  I figure if you whip up a beast like that, and precede it with a nuke stage with a decent number of say, 4km/s drop tanks, preceded by successively mightier if lower-ranged stages to eke out what you can, you might be surprised the degree to which you can tell the equation to go suck on a salmon.

[Streetwind]

An ion driven spacecraft would score even better than what @Gaarst calculated, if the mass ratios for xenon tanks weren't so awful. It's been scheduled for buffing by Squad, but now that it's questionable if the big rocket part revamp makes it into 1.2, I'm not sure when it will actually happen...

Still, this scenario here is a great example of the relative usefulness of ion engines. The stock KSP solar system is so small that the dV costs to go anywhere are negligible. Even the most expensive destination, Moho, can be done in less than 5 km/s. Which often leads people to ask: what even is the purpose of the Dawn?

Look at Gaarst's post. That is the purpose of the Dawn, and high Isp engines in general.

@GarrisonChisholm - you definitely need some higher Isp engines if you want to make things work. May I recommend Near Future Technogies?

Edited July 8, 2016 by Streetwind

[GarrisonChisholm]

13 hours ago, Archgeek said:

I've been working on a small monster of an ion craft that'll get a ~5 tonne payload 47.6km/s in 11 stages for the low cost of only 107 tonnes, or 75km/s in 18 for 389 if I push my spreadsheet to its limits.  (The drive system sans-fuel-pod is around 7 tonnes on its own, and has another 2km/s onboard it.)  I figure if you whip up a beast like that, and precede it with a nuke stage with a decent number of say, 4km/s drop tanks, preceded by successively mightier if lower-ranged stages to eke out what you can, you might be surprised the degree to which you can tell the equation to go suck on a salmon.

Wow!  47km/s in 107 tonnes?  I had no idea such a thing was possible!  When you finalize it, I'd be Very interested in that craft file.    The best I could manage last night was ~55 km/s in ...oh, gosh, probably 900 tons?  It costs 1.4 mil funds, though not nearly as big as Buster (above) .

If it were worded properly there might be a Challenge buried in this thread; for stock-only parts, greatest dV in 100 tons?... - or, something like that.

And, yes, I think if I bother with extra solar Anything in my next game, I will at the least use Near Future Tech for engines.

[Wcmille]

Are you putting an engine on every stage, or just using drop tanks?

[Archgeek]

2 hours ago, GarrisonChisholm said:

Wow!  47km/s in 107 tonnes?  I had no idea such a thing was possible!  When you finalize it, I'd be Very interested in that craft file. 

I was actually just thinking on whipping up something suitable for your goal.  Of note, though, the XenonTempest Mk3 is a bit beastly, with a dozen ion engines to keep the min TWR not too far below .02.  If what you're after is just a probe to go fly to another system and not worry about any science or return capabilities, I can probably come up with something more interesting.  For instance, I think 6 or 8 1x6 panels will power two Dawns, if the low min TWR's not a problem, which means a lot less mass than my 6 gigantors, and we can get away with using a decoupler between probe and fuel pod instead of two docking ports, saving more mass.

7 minutes ago, Wcmille said:

Are you putting an engine on every stage, or just using drop tanks?

Blatantly drop tanks.  The XT series is meant to be reusable, with mission packages of payload and fuel pod flown up seperately and docked on either end in orbit.

Edited July 8, 2016 by Archgeek
Minor tyop correction

[LaytheDragon]

1 minute ago, Archgeek said:

Blatantly drop tanks.  The XT series is meant to be reusable, with mission packages of payload and fuel pod flown up seperately and docked on either end in robit.

How many are dropped at a time?

[Archgeek]

11 minutes ago, LaytheDragon said:

How many are dropped at a time?

Various.  Due to the way the math works out, I try to make each stage wind up with as close to the same mass ratio as possible, by taking the mass ratio of the first pod, and calculating the deviation from it of each subsequent stage and varying the number and types of tanks (within reason) to minimize it.  The precise optimal(ish) arangment has been found to vary with payload mass, though.

[Gaarst]

So I went a bit further with theoretical experimentations, and found that the lightest nuclear powered ship with a single engine and a 10kg probe core to achieve 100k dV would have over 200 stages (Excel doesn't want to let me go further and I don't know how to write a more optimised spreadsheet, also I don't want to go into coding something right now) and most likely weigh between 5 and 6 million tons.

Using an ion engine, you'd only need 22 stages and you ship will only weigh 37.458t with a burn time of 19.3 Kerbin days !

(Note that I did not account for discrete fuel quantities caused by tanks of given sizes, let's just assume you're using procedural parts)

Spoiler

Extremely accurate scheme of the design used for the calculations

 

[GarrisonChisholm]

13 minutes ago, Gaarst said:

So I went a bit further with theoretical experimentations, and found that the lightest nuclear powered ship with a single engine and a 10kg probe core to achieve 100k dV would have over 200 stages (Excel doesn't want to let me go further and I don't know how to write a more optimised spreadsheet, also I don't want to go into coding something right now) and most likely weigh between 5 and 6 million tons.

Using an ion engine, you'd only need 22 stages and you ship will only weigh 37.458t with a burn time of 19.3 Kerbin days !

(Note that I did not account for discrete fuel quantities caused by tanks of given sizes, let's just assume you're using procedural parts)

  Reveal hidden contents

Extremely accurate scheme of the design used for the calculations

 

...wow.  I *suck* at engineering, because I would have bet $100,000,000 that what you just said was impossible. :\   I guess the only way to improve is to try, as in all things.  One seldom learns without a "do".

It sounds like a 300,000 m/s craft is eminently possible, especially if using Tweakscale.

 

But, wait- doesn't Xenon drain from All tanks equally, like Jet engines do?  I suppose you'd have to turn on each stage's tank as its turn to be used came up?

Edited July 8, 2016 by GarrisonChisholm

[Gaarst]

2 hours ago, GarrisonChisholm said:

Wow!  47km/s in 107 tonnes?  I had no idea such a thing was possible!  When you finalize it, I'd be Very interested in that craft file.    The best I could manage last night was ~55 km/s in ...oh, gosh, probably 900 tons?  It costs 1.4 mil funds, though not nearly as big as Buster (above) .

If it were worded properly there might be a Challenge buried in this thread; for stock-only parts, greatest dV in 100 tons?... - or, something like that.

I have a (theoretical, of course ) craft powered by a single ion, with 36 stages, with a burn time of 53.8 Kerbin days, weighing 100t and giving out a respectable 122,128m/s of delta-V.

[KerikBalm]

keep in mind that to get to 80 km/sec, you don't need 80 km/s of dV.

Just dropping your PE down to as close to the sun as you ship can tolerate will get you a pretty high velocity at PE. Then of course you have a conversation with Dr. Oberth... another way of looking at it is that you burn enough so that you're getting out of the sun's gravity before it has time to take the velocity away from you (yes this is very simplified).

Oberth may not give you more dV... but it does let you get more excess hyperbolic escape velocity for the same amount of dV

[Archgeek]

15 minutes ago, GarrisonChisholm said:

But, wait- doesn't Xenon drain from All tanks equally, like Jet engines do?  I suppose you'd have to turn on each stage's tank as its turn to be used came up?

Neither xenon nor monoprop do exactly that anymore -- Now they try to respect staging, drawing equally from all tanks in a stage (regardless of where they might be).  Before that, I was stuck with things like the XT Mk1 in my avatar, there -- a lot of radial tanks attached to the sides of a structural panel cube with a docking port on two ends.

[GarrisonChisholm]

Just now, KerikBalm said:

keep in mind that to get to 80 km/sec, you don't need 80 km/s of dV.

Just dropping your PE down to as close to the sun as you ship can tolerate will get you a pretty high velocity at PE. Then of course you have a conversation with Dr. Oberth... another way of looking at it is that you burn enough so that you're getting out of the sun's gravity before it has time to take the velocity away from you (yes this is very simplified).

Oberth may not give you more dV... but it does let you get more excess hyperbolic escape velocity for the same amount of dV

Yes, I thought about using a solar escape, but knowing my luck it would get too close and burn, and I thought extra mass for radiators would be a liability.  I will not be so fearful of this plan, and will give it a fair consideration.

Also, I have the as-yet unsolved problem of how to Intercept Valentine.  When I draw out a craft's maneuver node, it does not extend out to Valentine's 9.5 trillion km orbit.  I don't know if this is due to the mod or the install, but it looks like the intercept would involve actual Math, and I haven't even thought about how to crack that nut yet...

7 minutes ago, Gaarst said:

I have a (theoretical, of course ) craft powered by a single ion, with 36 stages, with a burn time of 53.8 Kerbin days, weighing 100t and giving out a respectable 122,128m/s of delta-V.

And, I think you're the first winner of whatever-it-is-called that you just won.  :]

[GarrisonChisholm]

22 minutes ago, Archgeek said:

Neither xenon nor monoprop do exactly that anymore -- Now they try to respect staging, drawing equally from all tanks in a stage (regardless of where they might be).  Before that, I was stuck with things like the XT Mk1 in my avatar, there -- a lot of radial tanks attached to the sides of a structural panel cube with a docking port on two ends.

Oh cool-  so, if you can use Tweakscale to make a xenon tank whatever size you wish, is there a size that is most ideal, considering that the tank mass is such a detriment?

[Gaarst]

41 minutes ago, GarrisonChisholm said:

...wow.  I *suck* at engineering, because I would have bet $100,000,000 that what you just said was impossible. :\   I guess the only way to improve is to try, as in all things.  One seldom learns without a "do".

It sounds like a 300,000 m/s craft is eminently possible, especially if using Tweakscale.

We can get away from the tyranny of the rocket equation but not that much, the mass growth is still exponentional.

While a 100km/s dV ship would only weigh 37t, a 300km/s ship has over 200 stages and weighs about 230,000t.

[GarrisonChisholm]

Just now, Gaarst said:

We can get away from the tyranny of the rocket equation but not that much, the mass growth is still exponentional.

While a 100km/s dV ship would only weigh 37t, a 300km/s ship has over 200 stages and weighs about 230,000t.

...Holy...krap.  Yeah, that's the Equation iceberg that I didn't know was there.  So, it sounds like the real limit of any rocket isn't what speed you want it to be, but how much you're willing to load on your pad.  - mind you, I am excluding orbital construction from my plans, as - well, I'm just not interested in all of the tremendous focus all that orbital rendezvous would involve.  It would make my hobby too much like work.  :\

So, it looks like my best-case transit time with non-futuristic propulsion would be 2k-3k- years.

[Archgeek]

18 minutes ago, GarrisonChisholm said:

Oh cool-  so, if you can use Tweakscale to make a xenon tank whatever size you wish, is there a size that is most ideal, considering that the tank mass is such a detriment?

Welp, if you're using tweakscale, then the ideal size would be whatever's right for that stage, substituting tank size in for number of tanks.  You could start just small enough to equal the mass ratio of the ship's on-board xenon, which you could tweak-scale into a very narrow spine indeed, then just make the single tank for each stage big enough to give the same mass ratio as the stage before it.  For bonus points, use a set length and just increase diameter, making your rocket a great and hilarious exponential cone.

[wumpus]

2 hours ago, GarrisonChisholm said:

...Holy...krap.  Yeah, that's the Equation iceberg that I didn't know was there.  So, it sounds like the real limit of any rocket isn't what speed you want it to be, but how much you're willing to load on your pad.  - mind you, I am excluding orbital construction from my plans, as - well, I'm just not interested in all of the tremendous focus all that orbital rendezvous would involve.  It would make my hobby too much like work.  :\

So, it looks like my best-case transit time with non-futuristic propulsion would be 2k-3k- years.

Don't forget you can add a ton of asparagused ion engines in there, and it will barely add mass (just be careful in the last few stages), this should cut the burn time to "leave the thing on for awhile".  Absolutely nothing can be done about the transit time (c is 300,000,000m/s), except that I'd have to assume that the nearest star would be 10x closer in the Kerbalverse the way everything else is scaled down.

[Wcmille]

If time is not a thing, can't you just do lots of gravity assists of Kerbol?

[Bill Phil]

If I know what I'm talking about (I don't think I do...), then isn't there an issue with timewarp ing beyond a few hundred years?

[Deddly]

2 hours ago, Bill Phil said:

If I know what I'm talking about (I don't think I do...), then isn't there an issue with timewarp ing beyond a few hundred years?

Only in the 32 bit version, if I understand correctly.

[Xyphos]

 

it's an ion-powered lander, fitted with a decoupling 2.5m heatshield and a sinlge radial parachute under the decoupler.
if piloted correctly, it can take 1 (un)lucky kerbal to a foreign star. (getting back is not guaranteed)
the Action Group 1 toggles the ion engines and fuel cells used to power them, so be sure to toggle them off when not in use or you'll run out of fuel for the fuel cells.
I also packed a single RTG at the bottom of the lander so you don't have to worry about running out of electricity. (it's also used as the attachment point for the decoupler)

download the craft file here: https://drive.google.com/file/d/0BxtArMQEkcH5VzYzOWJtZ1J3Qkk/view?usp=sharing

instructions: once you get it into orbit, make two or three gravity assists around Eve then start gravity assisting around Kerbol until your AP/PE reaches the foreign star, then alter your trajectory to gravity assist around that star and whatever celestial bodies it holds until you're captured. try to find a moon or planet to land on that has an atmosphere, as the ion engines aren't guaranteed to withstand the gravity.

Getting enough delta-v to go anywhere soley relies on your ability to navigate. there is no such thing as a fast trip, and the best route to your destination isn't a direct flight.

Edited July 9, 2016 by Xyphos