Atmospheric/Gravitational Drag Minimization?

[Nikolai]

After looking at some designs that look pretty small and still make it to Mun/Minmus on a different thread, I'm wondering if I'm not losing lots of potential delta-v to atmoshperic and/or gravitational drag. Any advice from the fuel-sippers out there on how to ensure that you're effectively minimizing these losses?

[PakledHostage]

Check out Closette's Mini-challenge: max altitude with this supplied spacecraft. The conclusion there seemed to be that climbing at terminal velocity is about optimal. Achieving this profile requires a fair bit of throttle modulation during the climb.

Also, check out the Optimal Ascent Profile for this spacecraft challenge.

Both of these challenges pre-date v0.16 so the stacks won't perform the same way anymore, but the conclusions are still valid.

[EndlessWaves]

One problem that causes large mun rockets is not realising just how little fuel and thrust you need. A big lander certainly looks cool but it dramatically scales up the rest of the rocket required.

Assuming you're not planning to drop off a rover, then If you're carrying more than 1000L of fuel for the 3 man pod then your lander design could definitely be slimmed down.

[Nikolai]

Thank you both for your advice (and the links to more info). It's clear that I have some work to do. Looking forward to it.

[Peppe]

What is the smallest craft that can make kerbin 100km orbit?

Looking at old threads nothing super small works anymore.

Did some tests with mechjeb auto asscent.

Built a simple rocket. Don't know if it has enough delta v to make 100km orbit -- haven't made it yet

Aero spike, 4 winglets, 4 T400 fuel tanks, small ASAS, and small 1 man capsule.

Fired the rocket straight up and jeb reported 2k m/s loss to gravity. 860m/s loss to drag. basically best case for drag and worst case for gravity.

Tried various other trajectories and closest I have gotten to 100km orbit is 100k apo ~ -62k per.

I believe those accent path settings were 6, 70, ~50% slider (1 sliver of black above the final red bar).

4416m/s delta v expended.

976 lost to drag, 1370 lost to gravity.

Edit: Got periapsis down to -40k with better turn finish altitude. Spend more time burning in orbital prograde.

Ran a bunch of launches and did a little manual binary search on end turn altitudes.

Found 48.5 provided the max benefit with the curve adjusted to have the last full red segment (upper right) shown fully. Moving slider right begins the flat line along the top. Sliding it left flattens the curve and induces a lot of steering error.

Starting altitude may note be ideal, but it feels pretty close.

Edited September 27, 2012 by Peppe

[EndlessWaves]

You're thinking far too large, you've got over 50% more weight that you need on that machine and the fins and ASAS are completely surplus, the command pod gyros have more than enough power. Something like the following at 7.58 tons will make it into a complete 100km orbit using the default mechjeb settings.

It uses slightly more DeltaV but I'd suspect that's down to mechjeb rather than rocket design.

If you want the simplest possible rocket then 3x400L fuel tanks, a pod and an aerospike should make it, although they're slightly heavier at 8.55 tons.

Edited September 27, 2012 by EndlessWaves

[Peppe]

For my testing i knew the design was fairly wastful. I just extended the design I saw the threads PakledHostage's linked to where they searched for optimal ascent paths. Adding a 4th tank gets it into orbit with ~66.7 liters of fuel in the current best path i have found. I think the goal of the design is to minimize error if you flew it manually, so you have asas, fins, and no staging.

I figured after finding an optimal flight path in the current version making the smallest lunar lander + return could follow. The super small designs the OP is looking at probably mostly come from older versions which had either different atmosphere for kerbin, fuel bugs, or just plan differently balanced engines.

Looks like with default ascent you finished about 1L fuel... I don't know what rockets it is true for, but for small/low drag rockets this path seems to be fairly optimal finishing with 5.4 L of fuel:

I think for most designs the 70K target for the end of the gravity turn should be brought down as you spend most of your ascent coasting in low drag regions of the atmostphere with a mostly vertical ascent. So you get low drag loss and high gravity loss.

[PakledHostage]

For my testing i knew the design was fairly wastful. I just extended the design I saw the threads PakledHostage's linked to where they searched for optimal ascent paths.

The configurations in the two threads that I linked to above were designed for "hand-flyability", rather than peak efficiency of design. That way, the challenges were more about testing ascent profiles and less about testing people's flying ability. Prior to v0.16, it was possible to reach orbit and even parachute safely back down using just 1.5 tanks of fuel.

[Person012345]

One problem that causes large mun rockets is not realising just how little fuel and thrust you need. A big lander certainly looks cool but it dramatically scales up the rest of the rocket required.

Assuming you're not planning to drop off a rover, then If you're carrying more than 1000L of fuel for the 3 man pod then your lander design could definitely be slimmed down.

Looking cool is far more important than efficiency and FPS.

[EndlessWaves]

Looking cool is far more important than efficiency and FPS.

It doesn't fly - yet - but it's got plenty of fuel for SSTO flights if I can work out how to launch it.

[khearn]

I just made it to a 100km orbit with a 1-kerbal capsule, 3 400l 1m tanks and an aerospike engine. No ASAS, no winglets. MechJeb handled the ascent and ended up with 19.1l of fuel left over. Probably not enough fuel to get back home from 100km, but that wasn't in the design specs.

Used a fairly default ascent profile. Started turn at 10km, ended it at 70km, final angle 0 degrees. Left edge of the turn shape slider was between the e and n of "Ascent Path".

Ascent stats:

Gravity losses 31.3%

Drag losses 22.1%

Steering losses 1.3%

Speed gained 45.5%

Launch mass 8.6 tons

Mass to orbit 2.6tons

Yes, that adds up to 100.2%. Why should we expect MechJeb to do math correctly?

I'm sure this could be improved, this was my first attempt and I haven't tried doing any tuning of the ascent path.

Edit: With the addition of a decoupler and a parachute, it can only get into a 100km x 88km orbit. Sorry, Jeb, you're not coming home. With some tweaks to the ascent path, it might be possible to get this build up and back down.

Edit2: With only 2x400l tanks, parachute & decoupler it can get the apoapsis up to 100km, but can't get periapsis above ground. But at least Jeb gets to come home. I don't think any tweaks will get this one to orbit and back. Probably not even just to orbit.

With 4x400l tanks, chute & decoupler, it makes it to 100x100 orbit with 91.9l remaining. That's plenty to do a retro burn for re-entry.

With 3x400l and 1x200l, parachute & decoupler, it gets to a 100x100 orbit with 44.9l left over. After a retro burn to bring the periapsis back down to 39km (sufficient to deorbit), he still has 36.6l to spare. That was more burn than needed.

Edited September 28, 2012 by khearn

[Grizzlol]

Does anyone know of an add-on or spreadsheet that shows the current terminal velocity (for a given altitude) ? It would help save some fuel knowing exactly when and how to manage the throttle seeing as trying to push the ship beyond terminal velocity is a waste.

[antbin]

It doesn't fly - yet

Oh god, is that 35 tanks with 3200L of fuel each? What does that weigh? 600 tons?

I think you will need eight wings per tank, so if you can fit 250 wings on there you should be good to go.

[EndlessWaves]

Oh god, is that 35 tanks with 3200L of fuel each? What does that weigh? 600 tons?

I think you will need eight wings per tank, so if you can fit 250 wings on there you should be good to go.

650 tons or so full as there are 33-odd aerospikes on the back.

Lifting off on the wings is likely to be almost impossible but given that 560 tons of the weight is fuel I'm hopeful it'll be able to glide back down because tri-couplers really don't like the shock of parachutes opening.

It's possible to fit 250 wings on ~10 tanks because I've done it before but as you can imagine it pushes the part count up well into laggy territory. I've concluded that above around 200 tons you need more lift that you can practically get from the current little wings but I'd love to be proved wrong.

[phoenix_ca]

This...well let's just say I'll be happy when some more realistic aerodynamics are implemented. Long-and-pointy design will win in KSP at that point. Right now it's still...flat? Those things shouldn't fly. Everything about them says intuitively to my brain "this thing is like trying to push a massive sail directly up; it shouldn't work!" >.<