Automatic Orbiter

[PolishSpaceProgram]

Hey all, I've been unsuccessfully attempting to make an automatic orbiter the past few days. IE, a rocket that will take off and get itself into orbit without any human assistance (besides staging). The one I made either overshoots or undershoots orbit, depending on what throttle I have set.

Anyone up for the challenge of creating an automatic orbiter?

[FatalFlux]

Oh god. This sounds like a living nightmare. Good luck, brave kerbonaut.

[Bobsama]

The closest I've gotten to automatic orbiter is launching straight up, manually nosing over to the horizon, and applying thrust until I'm at an orbital velocity. At the apogee (or perigee), adjust to a low-eccentricity orbit. That can take a while, though. My bigger concern is how do you reliably nose-over the craft? It'd be relatively easy to exceed orbital velocity and enter into an eccentric orbit. If your aim is only to stage it, perhaps you can use symmetrical ships and have one booster or engine engage moments before the rest, to nose over the craft. But then timing is still up to you.

[Letum]

Lots of my larger craft orbit with only the SAS being turned off for a short while to tip the rocket over and then being reapplid

[PolishSpaceProgram]

I would imagine the radial booster pack would be very helpful

[DoctorEvo]

So I'm allowed to control staging myself?

...I think I can pull it off. Still far from 'automatic' though - carefully-timed staging and ignition can give you a surprising amount of control over a well-designed rocket.

*edit*

Major progress made, but I haven't succeeded just yet. I built a nice, stable 2STO rocket, and after a few iterations (adding another tank to the upper stage; adding winglets to both stages; and adding an SAS to the lower stage), I seem to be nearly there. I've finally got my flight profile nailed down though: throttle to 95-100%, turn on stability augmentation, ascend at a slight angle (under full SAS control) to ~14 km on first stage and jettison immediately; wait 5-12 seconds as the upper stage arcs down to about 75 degrees pitch, and fire upper stage. It is at about 26 km and 1500 m/s that I run into trouble; the fuel burning from the front tank makes the upper stage unstable and it quits arcing towards the horizon, then flips over into a nose-up attitude until about 65 km and 2500 m/s (still going up too steeply for a stable orbit). More inert noseweight is probably needed, but I'm gonna call it a night right now.

[Rocketeer]

You can have a SRB stage firing radially from just one engine, giving asymmetric thrust and tilting the rocket. Or you can just carry an asymmetry in the rocket that makes it tilt slightly (but not too much).

Though, the simplest possible orbiter is just a 3 tank rocket (no parachute or SAS). It will slowly tilt by itself and reach near horizontal velocity just about when it runs out of fuel.

[DoctorEvo]

You can have a SRB stage firing radially from just one engine, giving asymmetric thrust and tilting the rocket. Or you can just carry an asymmetry in the rocket that makes it tilt slightly (but not too much).

Bad idea.

I tried it by mounting just ONE winglet on the side of the rocket, and even that was too much.

For the most part, you don't need to force your rocket to arc over - even the SLIGHTEST disturbance will be enough to make the rocket arc over on it's own, simply due to the pull of gravity.

http://en.wikipedia.org/wiki/Gravity_turn

Also, it's interesting to note that, if properly executed, this is actually the most efficient way to get a normal rocket to orbit.

Though, the simplest possible orbiter is just a 3 tank rocket (no parachute or SAS). It will slowly tilt by itself and reach near horizontal velocity just about when it runs out of fuel.

That will still tilt over too quickly. Using a two-stage arrangement with an SAS in the first stage (NOT in the second stage) works well for boosting your upper stage with enough velocity so that you can make a nice, gentle, well-timed gravity turn after you drop the first stage. The SAS fights the gravity turn, but once you drop it, your second stage will simply arc over with gravity on it's own (if it's stable, that is); all YOU have to do is pick the right moment to fire the engine and it should be nearly level when you run out of propellant.

[Warringer]

i Have run a few test and have to say that the effectiveness of a Gravity Turn depends on the rocket.

If you got a nice slim rocket without any boosters strapped to it, the Gravity turn works nicely. As soon as you have thrust that is not centered however the rocket turns so rapidly that you crash back to the launch pad. In that case you need the SAS to fight the wobble. But once you're over 10km you can turn off the SAS and the gravity turn does the rest.

[DoctorEvo]

i Have run a few test and have to say that the effectiveness of a Gravity Turn depends on the rocket.

If you got a nice slim rocket without any boosters strapped to it, the Gravity turn works nicely. As soon as you have thrust that is not centered however the rocket turns so rapidly that you crash back to the launch pad. In that case you need the SAS to fight the wobble. But once you're over 10km you can turn off the SAS and the gravity turn does the rest.

You still need SAS on a symmetrical rocket. Even the SLIGHTEST disturbance at launch will cause the gravity turn to occur much too soon, and you will inevitably crash.

I used a 2-stage arrangement because that meant I could 'turn off' the SAS at about 130km by simply staging and still be following the rules of the challenge. The top stage, without the SAS on it, performs the gravity turn just fine.

Oh yeah, and winglets. They help.

[foamyesque]

You can actually perform some interesting shenanigans with winglets. For starters, the two different kinds of winglets have differing masses (one is 0.03, one is 0.04, if I recall correctly) and different drag profiles. This means that even when arranged exactly symmetrically, you can shift a rocket's CoG slightly.

What's more interesting, though, is that you can in fact get the winglets to cause rotation in the winglet's plane. Simply place one a bit further forward than the other. The difference in where the drag is applied will create a torque. Unfortunately, this is such a minor effect that it is easily overwhelmed by other influences and therefore not practical.

Me, I've been working on doing this as 'throttle-to-max, hit space, achieve orbit' challenge, but SAS appears to be a necessity, which is going to mean more inputs. =(

[DoctorEvo]

You can actually perform some interesting shenanigans with winglets.

That you can.

For starters, the two different kinds of winglets have differing masses (one is 0.03, one is 0.04, if I recall correctly) and different drag profiles. This means that even when arranged exactly symmetrically, you can shift a rocket's CoG slightly.

Their drag values are the same (0.5). Their LIFT values are not.

What's more interesting, though, is that you can in fact get the winglets to cause rotation in the winglet's plane. Simply place one a bit further forward than the other. The difference in where the drag is applied will create a torque. Unfortunately, this is such a minor effect that it is easily overwhelmed by other influences and therefore not practical.

There's no reason why this should work. Many a high-performance sailplane use arrangements which place one wing a few inches in front of the other, to no ill-effect. The net lift shouldn't have any imbalance left and right.

Me, I've been working on doing this as 'throttle-to-max, hit space, achieve orbit' challenge, but SAS appears to be a necessity, which is going to mean more inputs. =(

You mean I have competition?

It's on, broski.

*edit*

SO FRIGGIN CLOSE

2,568,351 meters of ground covered... gotta be near 2/3 of the way around. I've got the velocity, it's just fine-tuning the attitude I ignite at...

[PolishSpaceProgram]

the rocket I'm working on for auto-orbiting looks very similar to that. Then again I imagine simple rockets would work best with the approach of using a gravity turn.

[DoctorEvo]

the rocket I'm working on for auto-orbiting looks very similar to that. Then again I imagine simple rockets would work best with the approach of using a gravity turn.

This rocket is about average complexity for me. It uses SASs, which the majority of my designs do not, but it is only two stages. The absence of complicated staging sequences certainly simplifies things.

Overall, for a good gravity turn, all you need is decent positive aerodynamic stability (which this rocket stage only has until the first tank empties, but by then it's outside the atmosphere, so that's good enough) and a good, intermediate acceleration (i.e. moderate amount of thrust). From there, it's all just about managing a proper entry - and this is the part I'm still perfecting. My first stage performs great, tracking at a slight 4-5 degree angle under SAS control, right until burnout at 15.5 km, at which point I stage immediately and wait for a target velocity to fire the engine. That last attempt was at about 92 m/s, but we're talking a fraction of a second opportunity to light the motor and get it right. If I do it, there's no way I'll be able to do it consistently, and it's certainly not completely automatic by any measure.

[foamyesque]

Evo, it works like this:

It gets more noticeable the more vertical seperation you have.

[DoctorEvo]

Riiiiiight...

WHAT effect am I looking for, exactly? As far as I can tell, it's not doing anything unusual.

[foamyesque]

This effect:

I'm not entirely sure where the rotation around the vertical axis is coming from.

[FoC400]

I don't understand how torque can be generated inwards in a planar surface.

I think I understand what you're getting at, but remember to use the right hand rule when defining your torque vectors.

Evo: The drag on top of the wing causes a resultant force perpendicular to drag and acting on the ship. When the two resultant forces aren't head-to-head (co-linear), they cause a moment about the center of the craft, thus creating a torque acting Into or out of the page.

Used a lot of mechanical terms there. Might need to google a bit.

[DoctorEvo]

I don't understand how torque can be generated inwards in a planar surface.

I think I understand what you're getting at, but remember to use the right hand rule when defining your torque vectors.

Evo: The drag on top of the wing causes a resultant force perpendicular to drag and acting on the ship. When the two resultant forces aren't head-to-head (co-linear), they cause a moment about the center of the craft, thus creating a torque acting Into or out of the page.

Drag is, BY DEFINITION, opposite to the direction of motion. There's no way for two drag vectors to be non-parallel without an (already-present) ridiculous rate of spin.

Used a lot of mechanical terms there. Might need to google a bit.

I ought'a slap you...

[andydouble07]

the drag produces torque, sheesh

[FoC400]

Drag is, BY DEFINITION, opposite to the direction of motion. There's no way for two drag vectors to be non-parallel without an (already-present) ridiculous rate of spin.

The drag vectors are not causing torque. The resultant forces are. (resultant forces are sometimes refered to as normal forces)

The drag forces are parallel and both acting in the same direction. The resultant forces from the drag acting on the craft are also parallel, but not acting in the same direction and are not co-linear. Therefore, by calculating the sum of moments about a point anywhere on the craft, you'll see that you'll have an angular acceleration by using Euler's trusty equation.

Now, because the torque creates an angular acceleration, and not a constant angular velocity, I would think offsetting the wings would take a lot of time to perfect the turning. But it would be interesting to see if anyone could get it to work.

Also, I didn't mean to insult you, just inform you.

[DoctorEvo]

the drag produces torque, sheesh

True, each winglet produces a moment, but they SHOULD cancel out, regardless of their longitudinal offset.

I'm starting to think that there may be a slight error in the radial position of the winglet when not using the symmetry tool, which would cause a small moment like the one Foamy's illustrating. Maybe.

The drag vectors are not causing torque. The resultant forces are. (resultant forces are sometimes refered to as normal forces)

The drag forces are parallel and both acting in the same direction.

Yes...

The resultant forces from the drag acting on the craft are also parallel, but not acting in the same direction and are not co-linear. Therefore, by calculating the sum of moments about a point anywhere on the craft, you'll see that you'll have an angular acceleration by using Euler's trusty equation.

WHA-

Go learn the difference between resultants and components; between coplanar, colinear and parallel; between lift and drag. Then MAYBE we can talk. Flipping through a glossary of physics terms and picking them at random does not constitute a cogent argument.

Also, I didn't mean to insult you, just inform you.

Well, aren't you a gem....

[FoC400]

We'll have to move this to a PM.

Sorry to everyone who had to read through the flames.

[DoctorEvo]

Alright, now that we've cleared things up a bit, here's what we've established:

-FoC400's theory (now that I understand what he was trying to say) is out.

-Empirical testing shows that the turning moment seems most prominent at launch and diminishes quickly as acceleration subsides.

-Neither one of us have any clue why it is happening.

In any case, it seems the effect may not be aerodynamic, judging by how it goes away once you're up to speed.

*edit*

In other news,

I'VE DONE IT!!

After fiddling with my previous rocket design, I realized that a properly-timed engine ignition would in fact put my burnout point right at level flight at a low altitude, but unfortunately the extra drag from the relatively-flatter ascent meant I no longer had enough velocity to make a full orbit.

This in mind, I shaved a bit of weight off my rocket (by replacing the nose-mounted LFT with an LFE) and began fishing for the optimal second-stage ignition point yet again. Eventually, after lighting my engine at a velocity of 67 m/s...

You see, after ignition, my rocket would make a steady gravity turn through the upper atmosphere, but once above 34.5, it would often keep pointing upwards as there was no longer any force on the fins to prevent this. THIS time, it happened to keep rotating... past level, and back towards Kearth. The rocket stopped climbing at about 45 km and was gently pushed DOWN again with the last bit of fuel.

This was FANTASTIC for me, as it meant that my perikee had been pushed IN FRONT of me, and I only had to wait a minute or two to cross it and determine whether my orbit was stable or not. As it so happened, my perikee WAS above the atmosphere, and since I didn't have enough velocity to escape, I think I can safely say that my automatic rocket has achieved a stable orbit!!

Will I be able to reproduce it? Hard to say. With this setup, luck and timing both played a massive part. Thus, even if my rocket still constitutes an 'automatic' orbiter, it's certainly not a reliable one.

[PTwr]

Wow! Just wow. I cant get to orbit with anything still