Rocket whipping about like a slinky.

[qoonpooka]

Rocket consists of 3xX200-32s atop a Mainsail, stack decoupler, Poodle with additional X200-32, another decoupler, station consisting of 3 crew cans and one mobile lab + attachments (PV, docking clamp, etc) separator and then 2t fairing + probe package.

Mission is a minmus orbital station + probe does flybys of some temperature sites then lands on minmus and sends back data.

Shortly after launch the thing whips around like a slinky before finally whipping itself into a tumble. No apparent cause, no amount of fins seems able to stabilize it.

Rocket length is 38.2m, total vehicle weight is 269t.

Struts don't seem to help, either.

FIXED: Joint Reinforcement seems basically necessary once you step up to 2.5m parts. With that mod, this thing flies well. Pics on 2nd page.

Edited November 2, 2015 by qoonpooka

[Nothalogh]

I suspect you're pushing into your gravity turn too hard

[qoonpooka]

I suspect you're pushing into your gravity turn too hard

The whipping around begins immediately after launch, before I even touch the controls.

The only thing I can think of is that I do have two probe cores (OKTOs) and I turn SAS on by habit. Maybe they're engaged in tug-of-war? But it can't be my grav turn because it happens even when I don't grav turn. (The slightest touch of the controls will, however, make it much worse.)

[Superfluous J]

Are the probe cores really far apart, or are they possibly right on top of the rocket with your mainsail at the bottom? What's your TWR at launch (You can tell by reading the G meter after you hit spacebar. Each tick is 1G and right at the surface, 1G=1TWR)? Do you have RCS or reaction wheels up at the top there as well?

What I think is happening is you're getting some resonance in your ship on liftoff. The probe cores sway to the right, and both any reaction wheels up at the top AND The mainsail try to compensate, causing it to sway to the left. Then they compensate again, causing it to sway again right. Repeat until your ship breaks up. It's not super common but it does happen. When it does, you have a few options:

1) Fly without SAS. It's a pain IMO and not worth it usually.

2) Disable gimbaling on the Mainsail. That's the main problem. It's far from your probe cores but assumes if they're turning, it's turning. So it tries to correct a problem that's not there.

3) Put a probe core on your main stack, and "control from here." on it. The closer to your Mainsail the better, though anywhere on the lifter is fine probably. This will cause the mainsail to consider "forward" where THAT probe core is, regardless of whether or not the upper stack is wobbling.

You can also look into the Kerbal Joint Reinforcement mod but I've been playing for 2.5 years and have never felt the need, myself, so I know it's not necessary to fix a problem like this.

[Snark]

Yah, +1 for engine gimbal being the likely culprit.

It's worth noting that gimbal isn't all or nothing. You can leave it on, but tone it down; I've found that for tall skinny rockets, a Mainsail is over-gimbaled, but often works well if the gimbal is toned down to 25% or 30%.

[Venusgate]

best rocket design is locating things that use torque closest to the center, and keeping fins farthest out. I know some people go one or the other, but i find using some of both keeps things from going too nuts one way or the other.

Also, in general, build out at the same rate you build up.If you squint your eyes, you should see an equilateral triangle on the launch pad

[qoonpooka]

Are the probe cores really far apart, or are they possibly right on top of the rocket with your mainsail at the bottom? What's your TWR at launch (You can tell by reading the G meter after you hit spacebar. Each tick is 1G and right at the surface, 1G=1TWR)? Do you have RCS or reaction wheels up at the top there as well?

What I think is happening is you're getting some resonance in your ship on liftoff. The probe cores sway to the right, and both any reaction wheels up at the top AND The mainsail try to compensate, causing it to sway to the left. Then they compensate again, causing it to sway again right. Repeat until your ship breaks up. It's not super common but it does happen. When it does, you have a few options:

1) Fly without SAS. It's a pain IMO and not worth it usually.

2) Disable gimbaling on the Mainsail. That's the main problem. It's far from your probe cores but assumes if they're turning, it's turning. So it tries to correct a problem that's not there.

3) Put a probe core on your main stack, and "control from here." on it. The closer to your Mainsail the better, though anywhere on the lifter is fine probably. This will cause the mainsail to consider "forward" where THAT probe core is, regardless of whether or not the upper stack is wobbling.

You can also look into the Kerbal Joint Reinforcement mod but I've been playing for 2.5 years and have never felt the need, myself, so I know it's not necessary to fix a problem like this.

Answering the questions: Launch TWR is 2.85, main lift stage is 2.51, orbit stage is 1.57, transit stage is 0.7. Reaction wheels exist, but have been disabled. RCS is equipped, but disabled. Rocket has fins near the tail.

Flying without SAS gives me a stable ascent, but the natural gravity turn (probe core riding on the side of the rocket, ultimately destined to be the station's maneuvering pilot) is too steep.

The probe cores are pretty far apart, one riding 2/3rds of the way up the station stack, and the other as part of the terminal probe payload. I'd fly without the probe core on the stationstack, but I'm worried the decoupling will knock the station out of minmus orbit - I plan to re-use the station as a refueling outpost for outbound interplanetary voyages.

It sounds like there's no way (short of unlocking better fuel tanks to reduce the joint count in the stack) to get this rocket off the ground.

Disabling gimballing leaves me with a rocket that can't really fly into the gravity turn and has to circularlize nearly a vertical ascent.

Edited November 1, 2015 by qoonpooka

[Harry Rhodan]

Launch TWR is 2.85,

With the new aerodynamic system after 1.0 you usually want this value to be between 1.5 and 2.0 to not hit the sound barrier too early and reduce aerodynamical stress.

[Snark]

With the new aerodynamic system after 1.0 you usually want this value to be between 1.5 and 2.0 to not hit the sound barrier too early and reduce aerodynamical stress.

Yah, 2.85 is way high. I'd contend that even 2.0 is a bit on the high side-- I generally aim for 1.5 myself, and the numbers I've heard other folks recommend most often are in the 1.3-1.5 range.

Aside from aerodynamic issues as Harry mentions, having a big TWR like that will greatly magnify any rocket-floppiness problems like you've been experiencing; the high acceleration puts a lot of strain on the joints. Floppy deformations will go up with the square of your TWR, I think.

Edited November 1, 2015 by Snark

[Superfluous J]

Agreed on TWR being too high. I'd replace that Mainsail with a Skipper with some solids to get back to 1.5 off the pad. Also, if your fins are control surfaces swap them for regular fins.

I disagree about building out "as fast as you build up." Your rocket should look like a pencil, not a lampshade.

[qoonpooka]

The problem was delta-v. I wound up cutting the station in half, and plan to launch the second half once I finish the probe's mission. It was still floppy but I wrestled it into orbit with 200m/sec delta-v left over after I parked the main station body. It sounds like this is a problem without much solution, but I'll mess with the lower TWR launch profiles. The boosters are released before I go supersonic and the next stage can be throttled, but I can tweak it a bit.

[Waxing_Kibbous]

If you aren't opposed to mods give Kerbal joint reinforcement a try, http://forum.kerbalspaceprogram.com/threads/55657-0-24-2-Kerbal-Joint-Reinforcement-v2-4-3-7-25-14

[qoonpooka]

If you aren't opposed to mods give Kerbal joint reinforcement a try, http://forum.kerbalspaceprogram.com/threads/55657-0-24-2-Kerbal-Joint-Reinforcement-v2-4-3-7-25-14

I will give that a check-out just on principle. This was my first 2.5m project, but the 1.25m rockets I had started to get an unnerving amount of flex to them too.

[Nothalogh]

Yah, 2.85 is way high. I'd contend that even 2.0 is a bit on the high side-- I generally aim for 1.5 myself, and the numbers I've heard other folks recommend most often are in the 1.3-1.5 range.

Aside from aerodynamic issues as Harry mentions, having a big TWR like that will greatly magnify any rocket-floppiness problems like you've been experiencing; the high acceleration puts a lot of strain on the joints. Floppy deformations will go up with the square of your TWR, I think.

Yep, I use the MechJeb acceleration control feature and set it at 1.5G, do a correct gravity turn with that and it just about falls into orbit

[Sanic]

There's a simple answer...

MOAR STRUTS!!!

(A picture would be really nice, as proper strut placement can make them exponentially more useful)

[qoonpooka]

There's a simple answer...

MOAR STRUTS!!!

(A picture would be really nice, as proper strut placement can make them exponentially more useful)

I tried struts, but they all seem to clip inside the tank walls and it's a VERY vertical stack. When I get home I'll try getting a picture, but it's so tall it exceeds the viewing area.

[Ripper2900]

I agree that the cause is a resonance. The various controls fights each other. Move the sas controls to other places, turn down or off gimballing on the motor, and use other fins.

You could also try to scale back the thrust of the motor, or change to a lower powered one if that still gives you 1.2 - 1.6 twr.

[uncle_jew]

I tried struts, but they all seem to clip inside the tank walls and it's a VERY vertical stack.

You can add some beams protruding outside to the middle of the stack. Then add struts between the ends of these beams and rocket elements above/below them. So you would get some latticework surrounding your rocket.If beams are aesthetically unappealing - place them on decouplers and throw them away once in orbit.

[qoonpooka]

Joint reinforcement did VERY nicely. Now it only flips out if I oversteer during the gravity turn. I was mistaken about the TWR. Those were after I chopped the station payload in half, and decided to send the 2nd half of the station up as an add-on.

Anyway, Joint Reinforcement seems to have solved the slinkyness. Here's photos for further commentary, though. (That probe payload worked like a dream, it's parked on Minmus with 2.8k dV to spare for future temperature scan missions.)