Jump to content

Wet spaghetti rockets


Scarecrow88

Recommended Posts

Crafts in ksp are built like trees. Everything is branched off of the root part of the craft. So your root part is the command pod up top for instance. Imagine holding a penicil at the tip and trying to control it. Now imagine holding the same pencil at the center and now control it. Much easier and more stable. No splitting of the ship and im not sure what you mean by manual page. With the root part gadget or whatever its called tool in the VAB. After your finished construction flip on the CoM indicator to see where its at. ( will most likely be in the first stage fuel tank so thats the one you want rooted ) Hit the root part button, click on a part next to the part you wish to be the new root then select the part. Now your craft is rooted from the center out.

Im not sure of the technicalities and forgive me if that explanation is a bit foggy. But it works really well. Combine that with KJR and its perfect. As for the people who dont like KJR because it makes things too strong, well then you havent tried FAR yet..

I don't get your point with the pencil. On a the rocket, the force that is moving it comes from the engine, and it makes the rocket turn when the thrust vector is not oriented towards the center of mass.

Maybe there's something with SAS, the root part and where it is located. But it's really not intuitive and certainly not documented. Moreover there's only 1 root part, then you cannot use the trick to balance ships with multiple stages.

Link to comment
Share on other sites

In older versions of KSP, some of my larger rockets used the big docking ports in place of decouplers, just in case I wanted to re-dock something. Setting 'control from here' on these mid-stack docking ports did lessen the wobble, because they weren't swinging around as much as the top of the rocket. Perhaps having a mid-stack root part has the same effect.

Link to comment
Share on other sites

I don't get your point with the pencil. On a the rocket, the force that is moving it comes from the engine, and it makes the rocket turn when the thrust vector is not oriented towards the center of mass.

Maybe there's something with SAS, the root part and where it is located. But it's really not intuitive and certainly not documented. Moreover there's only 1 root part, then you cannot use the trick to balance ships with multiple stages.

My analogy has nothing to do about how a rocket performs, but how a rocket in KSP is built. Again I'm not sure what you mean by multiple stages. That has nothing to do with the root part. With that tool in the VAB you can make any part of the craft the root part at anytime.

If by chance you mean making the part closest to the CoM initially isn't any good because your CoM moves as you burn fuel and stage depleted tanks then... It doesn't matter I'd say. I'm not sure what exactly goes on after you stage away the root part. At that point I don't know what would become the new root part, but whatever it is it works.

Regardless I find it makes a night and day difference by initially changing your root part. It may be even wiser to switch it to where the CoM will be after staging if it occurs really early, but wherever you change it too it'll be better then having the root part all the way at the top.

Edited by Motokid600
Link to comment
Share on other sites

Again I'm not sure what you mean by multiple stages. That has nothing to do with the root part. With that tool in the VAB you can make any part of the craft the root part at anytime.

"In the VAB" != "anytime"

How can you help with 2nd stage stabilization when you dropped your former root part as debris of the 1st stage?

- - - Updated - - -

Most of what I've seen with wobbly rockets as of 1.0 is because of the gimbal of the engines when SAS is on. I set the gimbal down to like 30-40 percent now mostly, and viola, no wobbling.

Usually it's still slightly wobbling, even if you can barely notice. And you lost 60% of your gimbal angle for when you need it, because of a buggy SAS.

Link to comment
Share on other sites

No idea. Not sure what then becomes the root part after staging. I'd love to know how that works. But nine times out of ten your in the clear by time your upper stage fires. And whatever it is it still works better then having the root part at the top.

Or make your root part the second stage tank if that works out for your particular craft. If it's a serial staged rocket that drops early it might be a good idea. But it's a bad idea to drop the first stage early because you then are in danger of flipping. Which is a different discussion.

Edited by Motokid600
Link to comment
Share on other sites

The root part doesn't matter. It's the Control From Here part that matters. SAS will try to correct based on the control part's current orientation. If it's far from the CoM, and the ship is flexing, the control orientation and the overall craft orientation don't line up. So, the SAS tries to correct based on bad data, leading to potentially destructive harmonics.

Option 1, Lots of struts.

Option 2: Turn SAS off and fly manually. The flex will dampen itself out in a lot of cases

Option 3: Put probe cores or docking ports on each stage. Control from the one closest to the CoM. I like this option because you can get by with a lot fewer struts. I built a rocket last night that had the top flexing 10 degrees from prograde. I manually set control to the probe core at the top of stage 2, and it got to space just fine.

Link to comment
Share on other sites

When I get wobbly rockets, it's usually in the lower stages of a launcher.

To solve it, I usually use 4 cubic octagonal struts surface attached to the biggest part on each side of the wobbling connection, and strut the bottom to the top (start struts on lower stage so their mass and drag is discarded with the lower stage)

Haven't had a wobble problem fail to be solved by this. Of course, the drag it creates might cause instability, and if that happens I either add more fins to the bottom of the rocket, or switch out the fins for bigger ones (the spaceplane rudder makes a great fin for 3.75m rockets).

Of course, it would be nice if I didn't have to add 12 parts to my rocket to make it stop wobbling, but at least it's usually on the launcher instead of the payload.

And if you're getting wobble across a docking port connection, that's because docking port connections are comparatively very weak. The only solution that I trust for that problem is KAS struts. Other solutions that I know of, I haven't tried, or know to have certain glitches that would cause me problems (Q-Struts hitting physics-less parts, for example)

Edited by SciMan
Link to comment
Share on other sites

The root part doesn't matter. It's the Control From Here part that matters. SAS will try to correct based on the control part's current orientation. If it's far from the CoM, and the ship is flexing, the control orientation and the overall craft orientation don't line up. So, the SAS tries to correct based on bad data, leading to potentially destructive harmonics.

Option 1, Lots of struts.

Option 2: Turn SAS off and fly manually. The flex will dampen itself out in a lot of cases

Option 3: Put probe cores or docking ports on each stage. Control from the one closest to the CoM. I like this option because you can get by with a lot fewer struts. I built a rocket last night that had the top flexing 10 degrees from prograde. I manually set control to the probe core at the top of stage 2, and it got to space just fine.

Changing the root part will help with wobbling. Significantly. I've been doing it since SelectRoot was a thing.

Link to comment
Share on other sites

If it's a serial staged rocket that drops early it might be a good idea. But it's a bad idea to drop the first stage early because you then are in danger of flipping.

Come on, wobbling happens when burning with your 4th stage right in the middle of space, it's not just a matter of flipping when leaving Kerbin.

Edited by gogozerg
Link to comment
Share on other sites

I didn't have any issues with this till I unlocked the 3.5m engines and tanks. I don't use KJR and the 2.5 m parts by themselves are just fine even without struts, there is definitely some issue with the 3.5m parts.

Link to comment
Share on other sites

Come on, wobbling happens when burning with your 4th stage right in the middle of space, it's not just a matter of flipping when leaving Kerbin.

Nevemind man. Your thinking waaay too much into it. Do whatever suits you best.

Link to comment
Share on other sites

I didn't have any issues with this till I unlocked the 3.5m engines and tanks. I don't use KJR and the 2.5 m parts by themselves are just fine even without struts, there is definitely some issue with the 3.5m parts.

Hum...

wob.jpg

Link to comment
Share on other sites

That is not true. Struts were introduced by Squad as a lazy fix to their wobbling joints problem. This "problem" was supposedly fixed in latest iterations of the game but like every other problem with the game, obviously not fixed at all. If anything struts are there as a crutch for people with ridiculous designs to somehow get them to fly, not the other way around.

That's not really true.

The "wobble" is an effect of using spherical joints instead of tri-joints. That's mechanical engineering 101; if squad can add in struts they can add in tri-joints. Often, using struts to create trijoints WILL fix most of the "wobble". So yes, it is "that easy." (It won't fix joint stress, tri-joints only restrict movement), (Now, yes, there's exasperation and things that make it worse, that's faults with the physics engine and how the rocket is represented in said engine.)

But "in reality" attempting to create a bigger rocket requires stronger and weightier infrastructure. "Space Stages" can be little more than tinfoil and be perfectly fine (so long as you keep forces correct). KSP doesn't have a good method of creating this effect, hence "wobble"

Make no mistake, rocket wobble is not a bug; the exasperations are (i.e. small mass against large mass stuff), but the wobble itself is by design.

Edited by Fel
Link to comment
Share on other sites

I'm talking about average rockets ascending under normal conditions, not a mainsail pushing a weak joint at 4x physics warp.

This is the rocket I had trouble with. Prior to this rocket (my first one in this career with 3.5m parts) I had dozens of 2.5m rockets including this one, which originally had orange tanks and was much taller. This isn't a strange obscure contraption in the slightest, but it would not launch well even after strutting it. I maintain there is some issue here, it shouldn't be that way. I can't put my finger on exactly what, but the only thing I did was retrofit the lifter.

2C82EA68440F4EF0ED7164723198A3BF89D484BB

Edited by Alshain
Link to comment
Share on other sites

I'm talking about average rockets ascending under normal conditions, not a mainsail pushing a weak joint at 4x physics warp.

This is the rocket I had trouble with. Prior to this rocket (my first one in this career with 3.5m parts) I had dozens of 2.5m rockets including this one, which originally had orange tanks and was much taller. This isn't a strange obscure contraption in the slightest, but it would not launch well even after strutting it. I maintain there is some issue here, it shouldn't be that way. I can't put my finger on exactly what, but the only thing I did was retrofit the lifter.

http://images.akamai.steamusercontent.com/ugc/39750633467608768/2C82EA68440F4EF0ED7164723198A3BF89D484BB/

"Wobble" on launch or "Wobble" on maneuvering? You have a top heavy second stage, but given the size of your first I'd bet you are trying to maneuver solely with thrust vectoring and reaction wheels.

Just think about the forces you're talking about applying to those joints from the base... think about the momentum involved. RCS helps quite a bit with those to keep all parts having the same vectorized momentum. (Do note: I am assuming you're using the classic "High TWR" launches once you leave the atmosphere. If it is actually wobbling with a constant 2 TWR or less all the way to space then it's buggy behavior)

(And, in the case of large mass differences between parts, momentum gets very screwy... or so I've been told when building super-heavies)

Edited by Fel
Link to comment
Share on other sites

Whats that pic suppose to show? Your at full physics warp with a mainsail at 100% throttle. I'd be surprised if it didn't wobble.

Of course this is an extreme case. Just showing it happens with 2.5 m.

More interesting:

- On this rocket, wobbling is very prononced at warp x1 when heading prograde with stock SAS (breaks at x3).

- When using Smart ASS, no wobble at all until warp x4 (or maybe slightly at x3). That was my point. In my experience, smart ASS always looked superior.

- I changed the root part -> no difference.

- I put another probe core in the middle, and controlled from here -> no difference.

Edited by gogozerg
Link to comment
Share on other sites

The wobble is intentional. Squad apparently likes it, so we have it. I kind of agree, as it makes rocket design more challenging.

There's nothing particularly mysterious in the way SAS works. It's just a simple PID controller. If the craft is rotating, SAS tries to apply torque to stop the rotation in a way that the craft ends up pointing in the right direction. The controller is so simple that it only looks at one part, assuming that the rest of the craft rotates in the same way. I'm not sure whether KSP keeps this control part separate from the root part during flight, but at least the initial control part depends on the root, and you can also change the control part during flight.

Consider the relatively common mistake, where you dock a heavy (e.g. 20-tonne) lander to a large (100-tonne) ship by a single 1.25 m docking port. If you control the ship from the lander, many things can go wrong. The docking port isn't rigid enough for the heavy lander, so the ship bends between the lander and the main body. As the bend is between the lander and the center of mass, any torque from the lander rotates the main body of the ship in the wrong way, while the torque from the main body rotates the lander in the wrong way. The result is a wobbly ship, and because SAS takes its measurements from the lander, it often ends up making things worse.

The standard solution to situations like this is to select a control part from the main body of the ship and to disable torque from the lander. The lander will still wobble, but because the main body is much more massive, because SAS takes it's measurements from the correct section of the ship, and because the lander just hangs passively from the docking port, the ship usually remains controllable.

Changing the control part doesn't always help. In my second tall rocket test, I disabled the reaction wheels in the payload and controlled the ship from the probe core in the lower stage.

tall_rocket_2.jpeg

The control part, all torque, and the center of mass were all in the same rigid section, so everything was supposed to be fine. Unfortunately the upper stage and the payload were too massive for the bendy joints (Skipper, decoupler, and the probe core). When I started the gravity turn, the rocket bent a bit. Then aerodynamic forces bent it more, and because the upper section was relatively heavy compared to the lower stage, the lower stage could not steer the rocket in the correct direction, and the rocket started to wobble. A few struts would have fixed the situation.

- - - Updated - - -

This is the rocket I had trouble with. Prior to this rocket (my first one in this career with 3.5m parts) I had dozens of 2.5m rockets including this one, which originally had orange tanks and was much taller. This isn't a strange obscure contraption in the slightest, but it would not launch well even after strutting it. I maintain there is some issue here, it shouldn't be that way. I can't put my finger on exactly what, but the only thing I did was retrofit the lifter.

http://images.akamai.steamusercontent.com/ugc/39750633467608768/2C82EA68440F4EF0ED7164723198A3BF89D484BB/

My guess: the rocket had too many joints. It would probably have worked better with 2 large fuel tanks instead of 8 small tanks.

Link to comment
Share on other sites

"Wobble" on launch or "Wobble" on maneuvering? You have a top heavy second stage, but given the size of your first I'd bet you are trying to maneuver solely with thrust vectoring and reaction wheels.

Just think about the forces you're talking about applying to those joints from the base... think about the momentum involved. RCS helps quite a bit with those to keep all parts having the same vectorized momentum.

(And, in the case of large mass differences between parts, momentum gets very screwy... or so I've been told when building super-heavies)

If you have to use RCS to stabilize your rocket, then that is a problem. RCS is for manuevering in space (as are reaction wheels), not launch. I've tried it with engine gimbals and maneuvering control surfaces and both but in either case it doesn't respond well. I'm not sure what you mean by a heavy second stage, you want your rocket a little top heavy.

My guess: the rocket had too many joints. It would probably have worked better with 2 large fuel tanks instead of 8 small tanks.

Just tried the bigger tanks, no noticeable difference

Edited by Alshain
Link to comment
Share on other sites

Of course this is an extreme case. Just showing it happens with 2.5 m.

More interesting:

- On this rocket, wobbling is very prononced at warp x1 when heading prograde with stock SAS (breaks at x3).

- When using Smart ASS, no wobble at all until warp x4 (or maybe slightly at x3). That was my point. In my experience, smart ASS always looked superior.

- I changed the root part -> no difference.

- I put another probe core in the middle, and controlled for here -> no difference.

Smart ASS reads universe data and directly manipulates the ship in ways that you normally would be unable to do. If you want to use MechJeb, use it; but it isn't realistic and it isn't vanilla so don't bring it up.

You are applying an absurd amount of force to a stage that has more mass, space or no, that is bad design. That you can use software to issue micro-corrections before the problems get out of hand is besides the point.

- - - Updated - - -

If you have to use RCS to stabilize your rocket, then that is a problem. RCS is for manuevering in space (as are reaction wheels), not launch. I've tried it with engine gimbals and maneuvering control surfaces and both but in either case it doesn't respond well. I'm not sure what you mean by a heavy second stage, you want your rocket a little top heavy.

I don't see it as a problem, why space x clearly used RCS in the atmosphere to stabilize their landing... into the ocean. Most people don't even disable reaction wheels on launch (they're used in the atmosphere all the time with KSP.) Having control surfaces on your third stage would likely help, just that RCS usually "looks better" and always provides forces despite the atmospheric density.

Also, the second stage is lighter than the third stage, hence "top heavy". Topheavy is not good design, topheavy means that your engines are applying a force to a "light" object that has to transfer to a "heavy" object (re: inertia). Keeping low TWR keeps this from being too much of a problem.

Link to comment
Share on other sites

Also, the second stage is lighter than the third stage, hence "top heavy". Topheavy is not good design, topheavy means that your engines are applying a force to a "light" object that has to transfer to a "heavy" object (re: inertia). Keeping low TWR keeps this from being too much of a problem.

No, Top heavy is a good design. You want a rocket to be top heavy. Regardless I tried it without the center stage. It didn't change anything, and it wouldn't have made it to the Mun... also I did mention that it worked fine with a different lifter, didn't I?

Link to comment
Share on other sites

Smart ASS reads universe data and directly manipulates the ship in ways that you normally would be unable to do. If you want to use MechJeb, use it; but it isn't realistic and it isn't vanilla so don't bring it up.

I didn't notice Smart ASS was doing some illegal controls. Got a link please?

But I don't want MechJeb, I want convincing stock controls. And in the way SAS is working, I often feel like there's something wrong, like oversteering leading to resonance.

Link to comment
Share on other sites

Just tried the bigger tanks, no noticeable difference

I tried building the rocket, and found a lot of problems. The primary problem is that there are too many joints in the upper stage and the payload. Other problems include too much control authority, too powerful engine for such a small rocket, and drag from the upper parts feeding the wobble once you start the gravity turn. I got the rocket behaving better after a major redesign, which included payload fairings, removing the separate reaction wheels, a 3.75 m upper stage (with inline fairings and a few struts to reinforce the Poodle), removing the fins, and tweaking the main engine down to 70% thrust / 50% gimbal. It behaved even better, after I replaced the inline battery with a few radial batteries.

Link to comment
Share on other sites

I didn't notice Smart ASS was doing some illegal controls. Got a link please?

But I don't want MechJeb, I want convincing stock controls. And in the way SAS is working, I often feel like there's something wrong, like oversteering leading to resonance.

As far as I know, Smart ASS is applying the same inputs that the player does, it's just smarter about when.

Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...