[0.22-0.23.0] Payload Fraction Challenge

[Nao]

Please read the challenge requirements first Kourston . Jet engines are not accepted, and submissions require more data to be visible on screen. So unfortunately i don't think this is a valid submission

Also idk if the info panel is enough for the OP but at least we need resource panel open on when in orbit.

Please post your results after you make necessary adjustments, good luck

[mhoram]

21.1% payload fraction ! without any mod ! 80km orbit.. 60 ton payload

Welcome to the forum Kourston and thanks for your submission.

As Nao pointed out, Jet engines are excluded from the challenge, so sorry, but I can not accept your entry for the challenge. The reason for the exclusion is that they are way more efficient in atmosphere than the other engines. Using Jet engines, it should be no big problem to create payload-fractions beyond 30-40% and perhaps even 50% - I played around with them for a bit in the 0.21 days.

For your next submission please include screenshots as specified in the submission description. The altitude of the periapsis can be seen in the map-view while hovering with the mouse over the periapsis-marker on the orbit.

So, i took another bite at the "no asparagus idea", also lowering my previous stage count to only 3 stages.

Payload: 7,4 t (this time it's fully 0.2 drag)

Launch weight 39,45t

Payload fraction: 18,76%

You seem to be the only one pushing the non-asparagus-style to it's very limits - I have to try it sometime.

Edited November 11, 2013 by mhoram

[Nao]

Welp, there is no reason to use standard staging in KSP from efficiency standpoint, but i think it's easier to make good looking designs with it, as the center stage will have the most fuel and the first booster is quite small.

And since good payload fraction comes from actual engine TWR efficiency (having all engines 100% thrust all the time while keeping speed and TWR within efficient flight profile), id doesn't matter much if we use fuel crossfeed or not, as the most important is when and how much engines are we decoupling.

While designing this rocket I also tried to lower stage amount as much as i could, it turns out that getting the right TWR at launch (or just after accelerating to vertical ascent speed) can make the rocket TWR increase follow increased TWR demand from atmosphere changes. So even thou we know that having more stages means more dV, I can use the increasing TWR of long burning stage to lower dV required to orbit by flying more efficient profile. I think 18,5% would be possible with only 2 stage standard rocket.

Nevertheless it's fun designing rockets in a new way. Give it a try

edit: also i do it because tavert is a boss and i can't beat him and his mathematical simulations so i go for small victories

Edited November 11, 2013 by Nao

[numerobis]

The catcaller 2 -- it has 49.96t pad mass, 9.33t payload => 18.67% payload fraction.

There's obviously room for improvement there, I'm sad -- I have no more tricks to pull! But really this is just a fun entry for being such a ridiculous spacecraft.

MechJeb-compliant using my special build of MechJeb, which notably provides:

• Better tracking of terminal velocity -- it calculates how much thrust is needed, rather than using a proportional controller.
  
• Better tracking the intended flight path -- it does a gravity turn, controlling the flight path with throttle rather than with pitch.
  
• Better setting the apoapsis -- it calculates how much thrust is needed, rather than pulsing until it gets it right.
  

My boost stage ends with 20 m/s left to move the payload around, which means I'm burning the equivalent of 4440 m/s vacuum. If the booster had a probe on it, it could easily deorbit. Stock MechJeb could probably handle this spacecraft, but it would be tight.

I'm using the trick of putting a cube below an engine to allow it to fire through the spacecraft below -- that reduces the number of decouplers you need.

I'm using a bunch of cubes to make the spacecraft very tall, slightly increasing the Isp and reducing the amount of atmosphere we need to cross. I have a tripod made of cubes to stabilize the spacecraft on launch. I'm also using the long tanks rather than squat ones. All this length forces me to put a lot of struts around to keep the entire thing rigid.

I'm using a docking port instead of a decoupler to dump the boost stage. The docking port stays with the payload, so that's 200 kg of payload I get to carry, rather than 15 kg of booster. I disabled cross-feeding of course.

Seven stages, which are: a T30 on top, 5x 48-7S on the penultimate stage, three stages of 3x 48-7S, an RT-10, then a drop tank. Each stage has 4t of bipropellent, except the RT-10 has just 2t to make the timing work out. The stages below 10km have small control surfaces.

Javascript is disabled. View full album

Edited November 11, 2013 by numerobis

[Nao]

For a non LV-N entry 18,6% is really good, congrats. The TWR you got there also seem quite big, that must have been quite efficient flight path.

That design looks crazy, Is there any other reason than looking great for the probe to be so high up?

One thing of note, i think RT-10 *was* the best stage 1 propulsion, from my recent simple calculations 48-7S take the cake right now

Also controlling flight path with thrust seems to be a great idea for any ship that has thrust to spare. It would be interesting actually to use decoupling of stages on the base of required TWR and not on the amount of fuel in tank. This could increase effective engine use.

[numerobis]

That design looks crazy, Is there any other reason than looking great for the probe to be so high up?

The altitude of the spacecraft is based off the altitude of the root, which in this case was the probe. Every part in the spacecraft uses that as its estimate of the altitude of the part -- so you get better Isp at launch. The improvement is calculable -- a whole 0.2% for the 48-7S! Also, gravity is lower, coincidentally another 0.2% improvement. Oh boy oh boy! I'm not sure how to calculate how much deltaV it saves to be 60 meters higher.

One thing of note, i think RT-10 *was* the best stage 1 propulsion, from my recent simple calculations 48-7S take the cake right now

I'm not going to outright say that you're wrong, but adding a stage of equal mass powered by 48-7S engines wasn't improving the deltaV as much as the RT-10 did, and wasn't improving the TWR as much either. What I'm more curious about though is whether an aerospike would help. TWR is bad, but Isp is fantastic.

One issue with the RT-10 is that its thrust requires crazy scaffolding to transfer up the force; even with all the scaffolding, it gets hard to control just before it runs out.

As a testbed for mechjeb, I might want to add in two rules: First, when throttling back, turn off low-Isp engines, rather than throttling everyone back. That would help my design gain a bit of efficiency; more important with an aerospike (or an LV-N). Second, turn off gimbals above the center of mass. I think I got some waste that way on the early stages in my launch.

Also controlling flight path with thrust seems to be a great idea for any ship that has thrust to spare. It would be interesting actually to use decoupling of stages on the base of required TWR and not on the amount of fuel in tank. This could increase effective engine use.

The question I was trying to resolve today (again) was: how fast should you go when you start the gravity turn? I've given up on trying to compute the optimal flight path from first principles, so I'm just using basically your recommended trajectory: 8km start, 40km end, 65% shape. For the speed, instead of arbitrarily clipping to 23 m/s^2, I decided to try to follow that trajectory. When the actual flight path is below the target path (climbing too slowly), point at the target and full throttle. When the actual flight path is above (we're climbing too fast), burn prograde but throttle back to let the flight path fall back down. The policy I tried that works best is to throttle just enough to maintain vertical speed, letting horizontal speed build up. I also tried maintaining speed and letting both vertical speed fall and horizontal speed go up, but that didn't work as well.

My testbed was an SSTO like the ascent challenge from some time back. 1 capsule, 1 X200-16 tank, 1 T30, and a mechjeb unit. I could get 193 to 198m/s remaining after asking for a 75.1km orbit, about the same as your recomended mechjeb settings, and much better than default mechjeb settings.

Edited November 12, 2013 by numerobis

[Nao]

...snip

(going backwards on your post)

Figuring out optimal ascent profile for that SSTO is interesting (and a good challenge by itself), but recently I've been favoring low TWR instead of efficient use of throttle, and that adds even more variables to the formula. Having more engine mass to accelerate seems to gut the efficiency of whole rocket. And the savings of that mass seems to be greater than the losses from suboptimal ascents.

When doing gravity turn I feel like the most important variable is vertical speed, and how much i get it depends on burn time of later stages. In a way, I think of maximum vertical speed as a bank for accelerating sideways. Many different rockets and flight profiles that i tried (not including long burning, low TWR LV-N stage like the one posted here) i find that around 350-400m/s maximum vertical speed was giving me best results (500m/s for the LV-N). Maybe it would be possible to make some useful correlation with it.

And idk about RT-10, I've loaded simple ship in KSP, and counting atmospheric dV they have pretty much the same performance at 2 TWR. (using 9x 48-7S +FLT400+FLT200 in place of RT10) but i may be wrong of course.

I love the 0,2% improvements thou, such dedication. Now it needs a long launch legs to put CoM higher and get even more free Dv.

[Seanner]

Behold, the dumbest way to design a ship:

Javascript is disabled. View full album

The payload masses in at 6.745 and the total at 34.090 = 19.786% rounding up. I was slightly over the drag requirement because I started with a legit lander that was under it before I read the rules, so I had to delete stuff (a few goo pods and a few engines) and tack on random other stuff to bring the mass back up. It is not mechjeb compliant because I can't figure out how to set it properly (never used it before) and it requires a sustained ~35 degree nuke burn until the very very end.

[numerobis]

Nao: the RT-10 has 0.5t dry mass. The 48-7S and fuel tanks have 1.275t dry mass. So the TWR improves much more for the RT-10 as you burn, which gets you to terminal velocity faster.

That said, your comments and Seanner's ship are making me reconsider my slavish devotion to minimizing delta-V, which after all isn't the goal of this challenge. I could see a short high-TWR burn to jump-start and reduce the delta-V to get off the pad and through the first km, then lower TWR thenceforth. To be tested...

Edited November 12, 2013 by numerobis

[Kasuha]

Nao: the RT-10 has 0.5t dry mass. The 48-7S and fuel tanks have 1.275t dry mass. So the TWR improves much more for the RT-10 as you burn, which gets you to terminal velocity faster.

That said, your comments and Seanner's ship are making me reconsider my slavish devotion to minimizing delta-V, which after all isn't the goal of this challenge. I could see a short high-TWR burn to jump-start and reduce the delta-V to get off the pad and through the first km, then lower TWR thenceforth. To be tested...

If you can afford to coast during your ascent, you have too strong (therefore too heavy) engines. That especially holds for passing the atmosphere.

[numerobis]

That is not necessarily true: the Hohmann transfer is all about coasting. So there's a tradeoff to be made. Atmosphere definitely tilts towards not coasting: coasting means you're going faster than if you weren't coasting, and drag grows quadratically.

[Seanner]

I believe the RT-10 and BACCs are superior choices in practice for power/space, power/simplicity, and power/cost. They are inferior for power/mass (by a lot). Just look at the RT-10 wet mass and compare that to the sort of liquid fuel staying power it gets you.

Some actual test data, using minimal payload to set RT-10 to 1.87 takeoff ratio, resulting in 602 sea-level dV:

3,155m altitude achieved

Then swapping that for 8 rockos with 5 T100 tanks and some cubic strut nonsense giving 1.82 TWR and 599 dV:

3,337m altitude, no throttling needed

(Of course, outside of this challenge to convert a steel bicycle into a 300' tall unicycle made of balsa wood so that it's "more efficient", I prefer the RT-10 myself... In this regard, Nao's entries are impressive.)

[Kasuha]

That is not necessarily true: the Hohmann transfer is all about coasting. So there's a tradeoff to be made. Atmosphere definitely tilts towards not coasting: coasting means you're going faster than if you weren't coasting, and drag grows quadratically.

I hope you realize that liftoff from surface has absolute nothing to do with Hohmann transfer. You can only start to make Hohmann transfers when you get a periapsis you can actually reach.

[numerobis]

I'm merely pointing out that coasting = waste is not always true in rocketry; there are settings where it's false. So it's more complicated than your hard and fast rule. Yours is a good guideline on ascent, for sure.

[Kasuha]

So, i took another bite at the "no asparagus idea", also lowering my previous stage count to only 3 stages.

Payload: 7,4 t (this time it's fully 0.2 drag)

Launch weight 39,45t

Payload fraction: 18,76%

Just out of curiosity I have copied your design, changed it to asparagus, removed some fuel and got it to orbit. I believe with Mechjeb you could do that with even less fuel.

Payload: 7.42 t

Launch weight: 37.23 t

Payload fraction: 19.93%

Edit: that 0.02 t more is because I turned the last separator the other way.

Javascript is disabled. View full album

Edited November 12, 2013 by Kasuha

[Nao]

Just out of curiosity I have copied your design, changed it to asparagus, removed some fuel and got it to orbit. I believe with Mechjeb you could do that with even less fuel.

Payload: 7.42 t

Launch weight: 37.23 t

Payload fraction: 19.93%

Woha, nice! "I'm so good at designing ships that i don't know how good I am"

I'm merely pointing out that coasting = waste is not always true in rocketry; there are settings where it's false. So it's more complicated than your hard and fast rule. Yours is a good guideline on ascent, for sure.

That is true, but Oberth would like to point out that we want to burn as low as possible, and coasting to apoapsis is a waste of energy.

Nao: the RT-10 has 0.5t dry mass. The 48-7S and fuel tanks have 1.275t dry mass. So the TWR improves much more for the RT-10 as you burn, which gets you to terminal velocity faster.

If you use 9 rockomaxes the TWR is always greater for them, and with FLT400+FLT200 they are lighter by 30kg while having around 3m/s less atmospheric dV and above 50m/s more vacuum dv.

A simple test with one stage of 9x rockomax at 2,03 TWR got me maximum altitude of 3368m (617m/s dV expended), while RT-10 stage with 1,87 TWR got only 3169m (609m/s dV expended). So not only they save 30kg but give more dV.

@Seanner great results congrats .

The BACC is kind of bad thou, a simple LV-T30 with FLT800 give very similar performance.

For example 2x BACC at 1,9 Twr can give ~1000m/s and for the same weight 3x LV-T30 (with FLT400+200+100 tanks) will have slightly better TWR and ~100m/s more dV

[mhoram]

There is some serious thinking ongoing

The catcaller 2 -- it has 49.96t pad mass, 9.33t payload => 18.67% payload fraction.

MechJeb-compliant using my special build of MechJeb, which notably provides:

• Better tracking of terminal velocity -- it calculates how much thrust is needed, rather than using a proportional controller.
  
• Better tracking the intended flight path -- it does a gravity turn, controlling the flight path with throttle rather than with pitch.
  
• Better setting the apoapsis -- it calculates how much thrust is needed, rather than pulsing until it gets it right.
  

Interesting improvements on MechJeb!

I'm using the trick of putting a cube below an engine to allow it to fire through the spacecraft below -- that reduces the number of decouplers you need.

That method always felt a bit like cheating to me, but it is valid for the challenge.

I'm using a docking port instead of a decoupler to dump the boost stage. The docking port stays with the payload, so that's 200 kg of payload I get to carry, rather than 15 kg of booster. I disabled cross-feeding of course.

You are not the first in this challenge to use this trick.

Nice Job!

Behold, the dumbest way to design a ship:

The payload masses in at 6.745 and the total at 34.090 = 19.786% rounding up. I was slightly over the drag requirement because I started with a legit lander that was under it before I read the rules, so I had to delete stuff (a few goo pods and a few engines) and tack on random other stuff to bring the mass back up. It is not mechjeb compliant because I can't figure out how to set it properly (never used it before) and it requires a sustained ~35 degree nuke burn until the very very end.

You are taking the lead with this entry! Gratulations! You also take the lead with the most stages used in this challenge so far

Just out of curiosity I have copied your design, changed it to asparagus, removed some fuel and got it to orbit. I believe with Mechjeb you could do that with even less fuel.

Payload: 7.42 t

Launch weight: 37.23 t

Payload fraction: 19.93%

Edit: that 0.02 t more is because I turned the last separator the other way.

Scratching the 20% mark!

This is already the second rocket that was improved. I will add another separate list that contains rockets with multiple designers. Since these entries were not contributed by a single player, they are considered noncompetitive.

[numerobis]

Took me forever to get Nao's spacecraft into orbit. Looks like 8/70/16/40 work as settings for an unattended mechjeb run (this might only work in my own build, but I'm pretty sure I didn't change anything when you have too little thrust). I needed to add an RTG on top also, or else MechJeb runs out of electricity by going insane with the attitude control.

Maybe tomorrow I'll have time to improve it. 20% is definitely in sight!

[numerobis]

20.1% achieved...

Stages:

1: 2x 48-7S on 5t of fuel.

2: 2x 48-7S on 4t of fuel.

3: 3x 48-7S on 4t of fuel.

4: ditto.

5: 6x 48-7s on 4t of fuel.

6: nuke on 2t of fuel. Might be able to remove 100kg of fuel.

This is the same as Nao's and Kasuha's rockets, basically, just built vertically so that decouplers are cheaper for me. I split Kasuha's first stage into two. I also split Kasuha's second stage in two, and moved two of the engines earlier in the sequence, because otherwise the rocket has too much thrust around 10km. So I have all the same engines and fuel; how did I improve the payload fraction? I added 80kg of payload!

I'm curious to see how either Kasuha or Nao fly their rockets; MechJeb doesn't quite seem to succeed, though it's very close. You two could probably afford to remove some of my rocket's fuel. I end with about 30 m/s in the nuke stage, but that's after overrunning the apoapsis and thus having an extra ~50 m/s of drag to fight as I go back down.

I suspect an aerospike would nicely replace the 1t tank at the bottom, plus six 48-7S engines from lower stages. That would drop another 125kg from the launch.

Javascript is disabled. View full album

Edited November 13, 2013 by numerobis

[mhoram]

20.1% achieved...

This is the same as Nao's and Kasuha's rockets ...

In a combined effort from Nao, Kasuha and numerobis, the 20% Barrier has been broken!

Gratulations!

[rpayne88]

Can you add a category for people like my self who use FAR? I know it decreases the Dv needed to reach orbit, but it also make rockets less stable during their gravity turn. And if you include FAR, can you also include KW (for the fairings only, no engines or fuel tanks?)

[Kasuha]

I'm curious to see how either Kasuha or Nao fly their rockets

I don't use mods so it's always a bit of trial/error with these ships.

Your problem seems to be with the Nuke stage which is where I always struggle, too. I just switch several strategies which I believe you figured out, too:

- before reaching apoapsis, put a maneuver at the apoapsis and pull prograde. Then burn along the maneuver indicator to achieve maximum rise of periapsis while not moving apoapsis (my apoapsis after early stages is usually at 50-60 km)

- after reaching apoapsis (or sometimes before it by varying amounts), burn above prograde to push the apoapsis just a bit ahead of the ship

- if the angle is too steep, burn at horizon, and watch what the apoapsis behind you is doing. If it starts going too high, push it down by burning above horizon.

- if things go wrong, retry

I got the ship to the orbit on third attempt but I think there was some luck involved, too. When I finally got my apoapsis above 70 km it was on the other side of the planet and I was at 55 km or so, so I had to keep that apoapsis above atmosphere by occasional burns while the fuel level was already dangerously low.

Edited November 13, 2013 by Kasuha

[numerobis]

My "strategy", if you will, is to make sure I'm going ~1250 m/s orbit speed and 40km when the nuke stage starts, and otherwise revert to the VAB.

Otherwise I let mechjeb burn at about 25 degrees, down to 15 degrees as the spacecraft gets higher (the settings are in the screenshots).

[numerobis]

Data point: a high TWR rocket (using a T30) launched from 76m altitude burns 12 m/s more than the same rocket moved up on clamps to 128m. So it is a measurable improvement. I would expect greater improvement from the rockets here, given they are low TWR, and their Isp is more strongly dependent on pressure.

[700NitroXpress]

So, I want to start by saying that this darn thing made my computer angry with me, but it did the job.

Also, this is a refueling station so even tho there's RCS tanks on the station, there are no engines or RCS ports on the station so it complies with the rules.

This is the Ultra Max MK-I refueling station.

Ultra Max on the launchpad. Prelaunch weight: 4585.51 tons 1736 Parts

Ultra Max in flight. Next image is duel stage separation. Third image is orbital altitude stage separation.

Orbit achieved with a little fuel to spare. Second image is of the orbit and pre-separation weight.

Launch stage final separation complete, utilities deployed. Second image shows the final payload weight post-separation.

The top stage is the payload stage, and there's a dead stage with only launch clamps in it that never really get's activated because they hold the decoupler plates in place at the time of launch.

Payload Fraction: 15.33% 7 Stages

Edited November 14, 2013 by 700NitroXpress