How to achieve good TWR when you need lots of fuel

[Chaos_Klaus]

There is two things you could maximize with a lander. 

You could go for a large delta v budget. It is easy to build a lander with conventional rockets that has 3500m/s+. Here I built one that has 4000m/s without staging. It's also hugely overpowered and not a very sophisticated design:

It takes about 600m/s to land on the mun and another 600m/s to take off again. This thing can land and take off from the mun 3 times. 

 

The other thing you can maximize is weight efficiency. Build a very light lander. In that case you would design it in a way so it would have no fuel left at all when it got back into munar orbit. 1200m/s of delta v and nothing more.

 

This design has very little room for error but only weighs 1265kg. It will just barely make it back into orbit with almost no fuel left.

 

Edited January 5, 2016 by Chaos_Klaus

[Plusck]

On 5/1/2016 at 4:47 AM, Chaos_Klaus said:

There is two things you could maximize with a lander. 

You could go for a large delta v budget. It is easy to build a lander with conventional rockets that has 3500m/s+. Here I built one that has 4000m/s without staging. It's also hugely overpowered and not a very sophisticated design;

[...]

It takes about 600m/s to land on the mun and another 600m/s to take off again. This thing can land and take off from the mun 3 times.

 [...]

OK, I'll bite. : D

So I made your "not very sophisticated" design. However, since we're talking about landers capable of doing something, I added some monoprop (minimum possible, but enough to allow a fair amount of orbital manouvering), some solar panels (again absolute minimum possible), some batteries (idem), a hex probe core (because we want to send a scientist to do science, or be able to pick people up), and some experiments (everything except for the materials bay, which would have been excessive, so five in all, plus an antenna).

I was tempted to add a spotlight but decided against it, to keep to what is strictly necessary for a fully functioning lander (dark-side landings are optional, after all).

All of this takes KER's estimate of delta-v from 4000 odd to 3245.

I sent it up to my Mun orbiting station, refilled it, then took it down and back up.

Since I hadn't flown the vessel before, I think I did about 6 tries on the landing to get the most efficient burn possible. From 3245 m/s dv on leaving the station, this dropped to 3208 after the deorbit burn, and 2501 on landing. Fuel went from 405 to 275 (67%).

I then went back up, lost about 40 m/s on fluffing the orbital rendezvous with the station, and spent exactly 1 unit of monoprop manouvering (yes, I'm better at docking than landing, it seems). This took me down to 1766 m/s according to KER, and 172 fuel (42%).

Since there was clearly enough to go back down again, I did it all again, getting a best-effort landing with 1050 m/s dv remaining (91.5 fuel or 22.5%) and back up with about 300 m/s dv (22 fuel or 5.5%) left, though admittedly it was somewhat approximate return to orbit since I was just wanting it all to end by that time.

I did try to take screenshots at each stage but I'm not sure if it's worth putting them all online. This is a pic of the craft leaving the surface the first time around:

Spoiler

 

So firstly, yes my "rule of thumb" was a bit off. I should correct it:

"To get down to the surface of the Mun, you will generally use a bit less than half* of your fuel (assuming your vessel is mostly fuel). Getting back up will generally take a bit more than half of what's left.

* NB: If you add even more fuel to your lander, you might make that proportion more one of thirds (i.e. a third to get down, slightly more than a third of what's left to get back up) but you'll end up using a lot more fuel in total. This is a variation on the law of diminishing returns: you can keep adding fuel, but it'll end up only marginally increasing your Mun landing / reorbiting capabilities. At the other end of the scale, you can reduce your fuel mass to slightly more than half your total lander mass, but you'll have no leeway for the slightest inefficiency."

 

To be honest, I still prefer my original rule of thumb. : p

[Chaos_Klaus]

Well, by adding more stuff you basically negate your assumtion that your vessel is "mostly fuel".

 

The rocket equation has an interesting twist. Let's assume your vessel has no payload at all and the engine does not weigh anything. That basically leaves you with fuel tanks only. In KSP, fuel tanks have a wet/dry-mass ratio of 9:1. That's as close to "mostly fuel" as we can get. In this case the rocket equation simplifies to:

 

delta v = ISP * 9,81m/s² * ln 9

delta v = ISP * 9,81m/s² * 2,20 

delta v = ISP * 21,58m/s²

 

That implies that the maximum amount of delta v depends on ISP only. With a Terrier that would be 7445m/s. With a nuke it would be 17264m/s.

Now since the natural logarithm in the equations gives us diminishing returns, we have a problem. If we add some payload, we would need an infinite number of fuel tanks to get this maximum delta v. There is a point where adding more tanks just isn't adding enough delta v. To me that point is around 10 * ISP. That is why I tend to build my stages so that each stage has this amount of delta v. With a Terrier that would make 3450m/s. Anything above that and I would actually use staging. 

 

I think it boils down to how you build your landers in general. I tend to build with 10*ISP in mind. That implies a certain wet/dry ratio. I usually do not put RCS on my landers and have the orbiter do the active docking, by the way.

Maybe your desings tend to always come out with a certain wet/dry ratio because that's what your intuition tells you. In that case, your rule of thumb would work well for your designs.