[0.22-0.23.0] Payload Fraction Challenge

[Nao]

Some new strides in launch technology had been made!

I present to you the "SoLucian" (probably nobody here will get that reference)

New look with pleasant aesthetic!

New performance that has never been reached before!

And an Automatic flight certificate!

Starting mass: 116,48t

Payload mass in space:23,72t

Drag: above 0.2 (two docking ports)

Mass fraction: 20,36%

Mechjeb ascent settings are in the screenshot (everything off except autostage, 7/40/30/66 settings). Other procedures: before launch, disable crossfeed on docking port, after reaching orbit decouple the ascent stage from it.

At this point i start to believe that 21% would be possible

Edited November 20, 2013 by Nao

[mhoram]

Some new strides in launch technology had been made!

I present to you the "SoLucian" (probably nobody here will get that reference)

And an Automatic flight certificate!

Mass fraction: 20,36%

Thanks for the great entry. Nice combination of LV-N, 48-7S and aerospikes. You are taking the lead! Moved your previous entry in the medium category to the attic.

[quietsamurai98]

Is there a reason for the lack of SRBs? An RT-10 SRB alone has 4742 dV (vac), while a aerospike and FL-T400 has 2916 dV (vac).

The reason why I chose aerospike and FT-T400 is that it weighs only 0.1 tonnes more than an RT-10, and the fuel is simaler (liquidfuel+oxidizer=400 and solidfuel=433).

Also, the T/W ratio is about 4 for the liquid, and about 6 for solid.

So, any reason for not having srbs?

[mhoram]

So, any reason for not having srbs?

As far as I can recall, there was only one entry with SRBs: http://forum.kerbalspaceprogram.com/threads/55615-Payload-Fraction-Challenge?p=738199&viewfull=1#post738199

[Nao]

Is there a reason for the lack of SRBs? An RT-10 SRB alone has 4742 dV (vac), while a aerospike and FL-T400 has 2916 dV (vac).

The reason why I chose aerospike and FT-T400 is that it weighs only 0.1 tonnes more than an RT-10, and the fuel is simaler (liquidfuel+oxidizer=400 and solidfuel=433).

Also, the T/W ratio is about 4 for the liquid, and about 6 for solid.

So, any reason for not having srbs?

BACC was always inferior to LVT-30 (a stack of 3 LVT-30 with fuel tanks > 2 BACC). In current patch RT-10 got shafted by 48-7S which is tremendously overpowered. The difference is quite small but 9x 48-7S with a FL-T400 and LF-T100 is lighter, and give more thrust for longer burn time than one RT-10.

The RT-10 do have more dV if placed on a probe, but that is not an accurate measurement of engine performance, because it's heavily affected by empty mass. For example counting dV for BACC, its worse than RT-10 even thou it expends much more energy and can grant much more dV to an actual rocket.

Personally i use: thrust times burn time divided by launch mass as a performance indicator (this goes around and doesn't count the effect of empty mass at all, but the results are much closer to actual performance on a rocket).

[numerobis]

Nao: I'm having trouble understanding exactly what you spacecraft is. Can you write down a dump of the stages (#engines, type, and amount of fuel in each stage) ?

I used aerospikes on the first stages of my own rocket to shave off a full 100 kg, but get almost the same deltaV and TWR. It wasn't enough to write home about, though I would include it in a future submission.

I've worked out that in the first three seconds, a sepratron is more efficient than a 48-7S. So if you were tempted to give your spacecraft a short boost at launch to get up to speed quickly (thus saving gravity loss), that's the way to go. That said, I haven't calculated whether it matters enough to be worthwhile.

Edited November 21, 2013 by numerobis

[Nao]

If i remember correctly it goes like this. Stages in reverse (all fully asparagused):

engine	fuel [t]
1LVN	1
2LVN	4 <- edit: it was 4t not 3t
2spike	12
16roko	13
12roko	13
8roko	13
16roko	13

As for the separatrons i think they are too heavy to be useful, their TWR is crazy good but the actual energy per mass they give is pretty bad. I'll check it thou .

Edited November 22, 2013 by Nao
fuel for stage "1" fixed

[Seanner]

numerobis when you say your deltaV and TWR with aero such and such ... are you comparing at sea level pressures, vacuum, both? Because if they seem comparable with vacuum stats then in truth aeros are better overall.

[numerobis]

1 aerospike (1.5t) replaces 6 48-7S and 1 tonne of fuel (1.625t) for the first ~30s to 1 minute to get a few m/s less atmospheric deltaV; you lose a lot of vacuum deltaV. The thrust and mass are nearly unchanged, so TWR is basically the same. Overall, not a big change. It would be a bigger benefit, though it wouldn't show up in these stats, if the aerospike were still 0.1 drag like it used to be.

[Nao]

1 aerospike (1.5t) replaces 6 48-7S and 1 tonne of fuel (1.625t) for the first ~30s to 1 minute to get a few m/s less atmospheric deltaV; you lose a lot of vacuum deltaV. The thrust and mass are nearly unchanged, so TWR is basically the same. Overall, not a big change. It would be a bigger benefit, though it wouldn't show up in these stats, if the aerospike were still 0.1 drag like it used to be.

Wait is there a specific reason why you are using aerospike engine for first stages ? Isp means almost nothing compared to engine mass for short burns. (For example the most mass efficient stage1 propulsion for Eve ascent before the 48-7S madness was actually RT-10 with whooping 225 Isp)

Also at sea level for burns of around 32s 6x 48-7S compared to aerospikes give 10% more dv and 3% more TWR for the 20kg less.

[numerobis]

The specific reason is that the difference in Isp is quite large at low altitude, but declines as you climb.

Looking at my notes, the first stage was burning for 90s (I split it up into three substages, 30s each). I think that's up to 10km? Carrying the engine up higher was, in the configuration I tested, counterproductive.

It also seems like the T30 should be useful after burning for a few km/s, but I couldn't get that to work either.

[Nao]

I think most engines are better aerospike for the first ~30s of kerbin ascent because of it's weight, they simply don't burn enough (fuel) mass to save more than their price in it.

Also another thought: but for me the best place to *not* stage the engines is around 7-10km as the TWR requirement sharply goes up. Since we already have engines that we spent precious fuel to accelerate why not use them more to reduce the dV requirement. This shows a little in my recent craft as after staging the booster first stage I try to stage as few engines as possible and have the biggest TWR around that altitude.

[Seanner]

Hmm you're right...now that I look at it, the aero doesn't surpass the rocko until ~1750 dV, by which point you are at a height where the aero has no chance. However (!!) the aero generally uses less fuel, falling the tinniest bit behind in full vacuum at higher dVs, but overall, not having attempted the calculation, I suspect the aero uses less total fuel to reach orbit, using whatever stages are optimal to do so.

This is important because in reality (in game reality, or real life I guess), you attach a parachute to the engine and then you don't have to buy it again. So considering refueling, even if you want to strictly go by payload fraction, then the mass over repeated missions is minimized by aeros because the equation changes from payload/(total mass) = payload/(fuel+engines+etc) to payload/(fuel*#(missions)+engines+etc). I would argue that using less fuel *is* using less mass. Implicitly, it costs less over time which to me is the real goal, i.e. if person X builds rockets according to this challenge and person Y builds rockets with aeros, person Y will accomplish more with equal resources.

edit: (nvm)

Edited November 22, 2013 by Seanner

[Nao]

@Seanner I've made a replay to your post earlier, but deleted it because it wasn't any good, but now seeing your "nvm" I'm super curious what was that about

Oh and one small thing i wanted to point out, is that fuel itself (unless its antimater) is generally much cheaper than engines. So it would take a lot of launches to net a profit on efficient engine. Right now fuel goes for around 200 for most tanks (with a 162,5 minimum on the 16t one). Aerospike costs 2566c per ton and LV-T30 costs measly 680c per ton. Yes the numbers aren't balanced but i do believe that when we have "credits" implemented in career mode engines will be at least 5x more expensive than fuel tanks (that's when the second TWR revolution will come and things like 1,2-1,3 TWR will be the optimal for standard rockets )

And if we wanted to talk about real engines, i think even thou we could retrieve the engines they won't be able to do much launches before scrapping or total overhaul (maybe 5 or 10) so the benefit's of lower fuel usage would be minimal.

[Seanner]

The "(nvm)" was me making no sense due to a mistake on my spreadsheet that I caught.

I disagree about fuel costs being low because one carries more fuel than engine mass in a launch intended to reach orbit+.

edit: nvm again...I need to sleep more.

Edited November 22, 2013 by Seanner

[Nao]

Sleep is for the weak!, students need 3h anyways . As for the fuel, its a matter of relative numbers. Considering one-shot rocket, if we use two more tons of fuel (out of 50 or 100) in exchange for one LVT30 on board, that's a cost improvement. Also as a small KSP bonus, lower acceleration means lower "max Q" so designs that are just between controlled explosion (normal launch) and an uncontrolled one (fireworks) can benefit from lower TWR.

edit: fixed horrible typo, also crunch kills people, i've lost a friend to it lately, he doesn't even have time at Sunday evening >_<.

Edited November 23, 2013 by Nao
Gawd, wrote week, thinking hard not to make a "calendar" mistake >_<

[numerobis]

I always needed 8h, or I was less effective overall.

I'm at crunch time at work; lots of ideas I don't have time to test... Looking forward to cracking the 21% point with y'all.

[Seanner]

There's definitely some upper limit (well obviously it's not infinite). One could set up a recursive program or equation if so inclined to figure out how much each additional pair of best-sized rocko stage adds to the overall solution's mass, and then map this equation onto a range of payloads and take the max. Maybe I'll work on that when I'm lucid.

[mhoram]

@Seanner your comment goes in a direction, I was thinking about.

Finding a ship with the best payload fraction consists of several tasks.

- finding the optimal engine and fuel distribution for the propellant

- performing an optimal gravity turn

I believe that the following task is also a part of the process

- finding the optimal payload mass

The reasoning for this idea is that different payload masses require different propellant configurations.

If we take two different payload masses and construct for both the optimal propellants, then I am quite sure that both rockets have different payload fractions.

This means that there are sweat spots for payload masses that allow better payload fractions than other payload masses.

[Kasuha]

Assuming we have some utility which will find optimum ascent for any ship, the process could be iterated using something like the following rules (I am assuming asparagus staging):

- get some engines and fuel, no payload. Put all engines on the last stage (staging only empty fuel tanks)

- whenever the ship doesn't run at full thrust, move some engine power from the last stage to just staged fuel tank

- after the ship gets to orbit, exchange all remaining fuel for payload

[Seanner]

How do you decide proper fuel size tank?

edit: And I agree, max fraction depends on the payload size.

Edited November 23, 2013 by Seanner

[Kasuha]

If you can lift 5 t with a ship. you can lift 10 t if you put two such ships side by side. So technically any fuel amount is ok. You only decide on initial TWR, i.e. ratio of fuel and engines. So that's the next thing over which you may need to iterate.

[rpayne88]

I like this, but can you please add a category for people who use FAR?

[Nao]

I like this, but can you please add a category for people who use FAR?

I didn't use FAR for quite some time, but that would be pretty interesting. If anybody else is up for test's ill chip in. (In far the required TWR to obtain lowest dV to orbit is really high, i remember needing around 3000 and something m/s dV to orbit, starting gravity turn off the pad )

[numerobis]

As payload increases, the payload fraction should improve, because of integer effects blocking you from implementing the best designs. But eventually part count will bite you.