Rocket design suggestion when payload is dry fuel tank?

[FancyMouse]

e.g. a vessel to transfer fuel from base to orbit, or a vessel of just fuel storage. It doesn't need to carry fuel at all when arriving at destination, but it doesn't have to be empty to launch as well. In that case, what is a good launch rocket?

To clarify what I mean, this is an example ship. Everything above the middle docking port is payload except for the fuel:

Ignore what KER shows. This is capable of landing at Mun/Minmus (a lot dV left). The launch profile goes like this: All fuel tanks are full, twin-boar brings the thing high enough, then cut engnine, move fuel up, decouple docking port, then use the Poodle for the vaccum thrust. With the Twin-boar consuming the S3-14400 amount of LFO, I can get ~3km/s on LKO.

Now, this above thing is mainly just randomly sticking pieces together (one evidence that I'm having fuels a lot more than I needed). So I'm seeking advices how other people do this kind of stuff. Specifically, you can answer these questions if you prefer specific questions (but general advice still welcome!)

1. How to design the stages better? I understand it's more like making the lower-stage tank dry mass into payload, but how it affects rocket design?

2. How to design so that it launches more easily? My current ship involves a bunch of non-stagable/actionable operations (move fuel, decouple docking port), and I'm afraid a couple of days later when I want to launch again I'll just totally forget how to launch it.

[Jimbimbibble]

I would design it so that the first stage (plus any boosters) will almost get you to LKO. That will allow you to send up almost full tanks with minimal effort during launch.

[Snark]

If you have anything special about "how to fly this rocket", why not add them as notes in the little window under the ship name? That way, whenever you're launching another one you can just refer to the notes.

Also, you've got a really blunt, flat front end, which is horrible for aerodynamics. For cases like that, I like to put on a disposable nosecone. Nosecone up top, with a decoupler underneath it (make the decoupler upside-down, so that it's discarded with the nosecone rather than sticking to the rocket). Move the decoupler fairly high up in the staging sequence. I then go to "Action Groups" and tie that decoupler to the "Abort" action, which I never use for anything else.

That way, as I'm launching, whenever I get to the place where I judge that the nosecone is no longer paying for itself (i.e. the cost of the extra nosecone mass outweighs the diminishing benefit from aerodynamics), usually around 25-30km, then I just pop the nosecone off.

The launch profile goes like this: All fuel tanks are full, twin-boar brings the thing high enough, then cut engnine, move fuel up, decouple docking port, then use the Poodle for the vaccum thrust. With the Twin-boar consuming the S3-14400 amount of LFO, I can get ~3km/s on LKO.

This part confuses me. It sounds like you have a straightforward two-stage-to-orbit design. Why do you need to move fuel? Why not design it so that the twin-boar uses its own fuel, then when it's empty, you discard it you just use the poodle with its own (still-full) fuel to complete the orbital insertion?

Edited July 27, 2015 by Snark

[Sharpy]

This part confuses me. It sounds like you have a straightforward two-stage-to-orbit design. Why do you need to move fuel? Why not design it so that the twin-boar uses its own fuel, then when it's empty, you discard it you just use the poodle with its own (still-full) fuel to complete the orbital insertion?

Because Boar will use more than it can store but less than the payload can store; instead of adding another tank you just drain fuel from the "payload" one, except you can't force engines to drain in bottom-first order - make the Boar first drain its own supply and only then start depleting the payload. It will drain payload all the way and when it's to be decoupled, fuel left in it needs to be moved up.

[GoSlash27]

If my objective is to simply have an empty fuel tank in orbit, I'd design it to hold the fuel required to get it there. Put a disposable nose cone on top and a detachable engine unit underneath with a small reserve tank for deorbit.

Best,

-Slashy

[Claw]

I agree with slashy. If the goal is to get an empty tank to orbit, then use that fuel if at all possible, and reduce tankage elsewhere. An SLS tank with a Rhino on it can SSTO even with other parts added on. Then you can jettison the engine section on orbit. I would build the engine section to contain all the extra stuff not needed by the tank so you can minimize debris. (Maybe even include a small amount of fuel or monopro to deorbit.)

If you are asking in the general "did I over design this rocket" sort of way, it really depends on the follow on goals. I can't tell if that's what you are asking though, but there are certainly a lot of tips in that area too.

Cheers,

-Claw

[FancyMouse]

Thanks all. The goal is mainly to get empty tank to orbit/surface of Mun. Poodle is for the purpose of circularizing LKO and arriving at destination. That part isn't that interesting, but more interesting is the lifter part.

I guess given that the lower stage dry tank becomes payload, you don't get too much benefit from having too many stages (because in perfect world you can only throw away engine mass, not tank mass). I guess my question is more specifically in this case is it better to do more stages than my current two-stage design?

Also, you've got a really blunt, flat front end, which is horrible for aerodynamics. For cases like that, I like to put on a disposable nosecone. Nosecone up top, with a decoupler underneath it (make the decoupler upside-down, so that it's discarded with the nosecone rather than sticking to the rocket). Move the decoupler fairly high up in the staging sequence. I then go to "Action Groups" and tie that decoupler to the "Abort" action, which I never use for anything else.

I do use decoupler-nosecone trick when dealing with 2.5m rocket. But for this one - I just can't even get close to terminal velocity no matter how I launch, so I don't even bother. I guess it's just so heavy and Twin-boar is just so powerful, and aero is no longer proportional to mass.

- - - Updated - - -

This part confuses me. It sounds like you have a straightforward two-stage-to-orbit design. Why do you need to move fuel? Why not design it so that the twin-boar uses its own fuel, then when it's empty, you discard it you just use the poodle with its own (still-full) fuel to complete the orbital insertion?

Sharpy was right. The usage of docking port for decoupler is to allow fuel crossfeed down, and I won't need to move fuel up if Twin-boar can drain from bottom tank first. If Twin-boar uses its fuel all the way, then there will be unused fuel tank and will be less efficient - I'm literally saving the fuel tank mass of Twin-boar's own fuel with this design. If I really carry full fuel for Poodle for its own use - that's way too overkill and will result in a even worse lifter.

Just to compare. I did have an earlier version which payload only carries Poodle+its own fuel (not full, but giving similar albeit fewer dV than my above design on LKO), and it was ~40t heavier.

Edited July 27, 2015 by FancyMouse

[GoSlash27]

Well darn it...

I knocked together a design to illustrate the concept. Then I found out I had the wrong concept...

http://s52.photobucket.com/user/GoSlash27/slideshow/KSP/OrbitalTank

I suppose this could make lunar orbit or even the lunar surface. It has plenty of DV remaining. But of course you wouldn't make a recoverable lift stage to do that...

I'll check and see.

*edit* yeah, it's got plenty of DV to reach Munar orbit, or the surface. Over 2km/sec. No need to 2-stage it since the Rhino has excellent Isp in vacuum. If it's going to the munar surface, does it really *need* a detachable booster?

Best,

-Slashy

Edited July 27, 2015 by GoSlash27

[mrclucks]

The fins you have seem a bit large for your rocket, have you tried using smaller ones for some minor savings? Based on what you said about aerodynamics, the draggynes of your flat top isn't enough to flip you.

Another thought, do you need the twin boar engine? You might be able to just use a mainsail with no fuel tank. Add a decoupler to the bottom of your poodle and the mainsail underneath that and use fuel lines to direct the fuel into the mainsail.

You could probably also save dv by not even stopping to circulize in lko and just go straight up to the mun, but that would require some good positioning of the mun before you did that. You would only need like 2k dv to get out of the atmosphere and then be able to use your high ISP poodle for space

[FancyMouse]

*edit* yeah, it's got plenty of DV to reach Munar orbit, or the surface. Over 2km/sec. No need to 2-stage it since the Rhino has excellent Isp in vacuum. If it's going to the munar surface, does it really *need* a detachable booster?

Best,

-Slashy

Um interesting question. I guess you're able to do that because you have two S3-14400's, where you could fill in enough fuel. (Not at home computer now) I remember yesterday when I was trying single stage it was giving something like ~4.5km/s dV, enough for LKO but a little short to land on Mun. But I wasn't trying optimal engine and just randomly sticked one, so maybe Rhino would work - I'll check later today.

The main reason I do two stages is that I want the payload to be able to take off from Mun by itself when refueled, but engine is not big enough to have design problems on landing. Poodle is the excellent option for this purpose. Other 2.5m engines or Rhino - I would imagine I need to spend some time placing the landing legs.

The fins you have seem a bit large for your rocket, have you tried using smaller ones for some minor savings? Based on what you said about aerodynamics, the draggynes of your flat top isn't enough to flip you.

Another thought, do you need the twin boar engine? You might be able to just use a mainsail with no fuel tank. Add a decoupler to the bottom of your poodle and the mainsail underneath that and use fuel lines to direct the fuel into the mainsail.

You could probably also save dv by not even stopping to circulize in lko and just go straight up to the mun, but that would require some good positioning of the mun before you did that. You would only need like 2k dv to get out of the atmosphere and then be able to use your high ISP poodle for space

Well, my earlier rocket is kind of done on purpose for every piece.

Fins - their mass is ignorable for this heavy ship, so I just put them for safety.

Engine - I need its TWR. I can't stick Mainsail in such a way that it gives ~1.5 TWR. I also don't want to stick fuel lines if possible, as my base is getting slow now and fuel line sticking that way will remain on payload.

Trajectory - this topic is by itself debatable, and I prefer not to do that even if it's slightly more efficient because I don't want to (read: I'm bad at) launch at the right spot.

[GoSlash27]

Um interesting question. I guess you're able to do that because you have two S3-14400's, where you could fill in enough fuel. (Not at home computer now) I remember yesterday when I was trying single stage it was giving something like ~4.5km/s dV, enough for LKO but a little short to land on Mun. But I wasn't trying optimal engine and just randomly sticked one, so maybe Rhino would work - I'll check later today.

The main reason I do two stages is that I want the payload to be able to take off from Mun by itself when refueled, but engine is not big enough to have design problems on landing. Poodle is the excellent option for this purpose. Other 2.5m engines or Rhino - I would imagine I need to spend some time placing the landing legs.

FancyMouse,

Really, it's just a case of KER not giving an accurate DV estimate in the VAB. Not that I blame them; it's a difficult estimate to pull off.

I didn't fill the tanks completely on this example. I put in enough for it to hold it's speed immediately after dumping the SRBs. I just designed it to put a big ol' tank in orbit with no debris and any extra fuel I had leftover was a bonus.

The landing legs for munar surface use adds another wrinkle, as does keeping the engine to return to Munar orbit. Neither obstacle is insurmountable with this approach, but now you're not talking about a "tank" as payload, but rather a "tanker". That makes this a different problem; a vehicle as payload rather than a tank.

The Rhino nearly matches the Poodle for efficiency and is capable of lifting a whole lot more mass, so I'd probably stick with it as a single stage. Why incur the waste and complexity of staging just to save 10s of Isp when you already have something that works in a single stage from the pad? How you plan to get the fuel *into* the tanker.... I dunno.

I'd think about setting it down sideways with tiny lift rockets. More capacity, less headaches.

*edit* I tried my example design as a sideways munar lander. Close, but no cigar. The addition of lift rockets suitable to lift it when fully-loaded cost just enough DV to where a zero descent rate munar approach was required. I'm assuming that if you want a tanker to bring fuel from the munar surface to orbit, you want to set it down in a precise spot. The ZDR approach is terrible for that.

It could probably be tweaked to do that job (it's only the first landing where this would be a problem), but if it were me I'd leave it as a munar orbital depot and have smaller dedicated tankers to fuel it from the surface.

Best,

-Slashy

Edited July 28, 2015 by GoSlash27

[FancyMouse]

The initial vessel is indeed my "tanker" to transport fuels from surface to orbit. The fuel storage (which lands and sits there forever) is another one which is supposed to be even bigger (but not yet designed - I want to figure this out first since it's smaller). Both need to land, though.

I just checked. I actually remembered the wrong number - single stage on the initial vessel is giving ~5km/s (whether using Mainsail or Rhino, just <100m/s difference), which is probably still going to be a few hundred less than needed, but close. And just to clarify - KER has no problem dealing with this number for single-stage dV. The 4.5km/s I claim in that reply wasn't from the same picture. I do know some bug in KER, but I don't want to over-blame KER that is in fact not true. Even for the picture I posted, I understand KER can't show the right number in that case because it can't possibly know I launch in that awkward way.

I saw your SRBs - that fills the gap (both dV and TWR) for the single-stage approach, but - it's not single stage any more, I suppose?

Anyway, still appreciate your ideas, especially let me think more about the possibility of single stage designs.

[Wolf Baginski]

Some of the half-stage options might be useful, such as SRBs. They won't get into orbit, so there'll be some debris down-range, and maybe you can do things with a small probe core and parachutes to get actual recovery. With the acceleration-limiter in MechJeb the core stage can be running at low throttle for most of the SRB burn. Attach some fins to the SRBs as the engine gimbaling of the core stage will be less useful at low throttle.

You don't want to greatly increase the flight speed with the SRBs as things are in the high-drag phase at low altitude. Effectively the launch pad is being jacked up a few km. You're also jettisoning some hot objects before they can transfer a lot of heat to the fuel tankage.