Tips on undersizing engines?

[Streetwind]

I'm generally fine with low-TWR orbital stages that don't need to land anywhere. So often enough, when hoosing an engine for them, I'm tempted to downsize - mount a Terrier instead of a Poodle, for instance. But at the same time, I feel kind of pressured into mounting the Poodle anyway, because I know that the stack will start to wobble a lot if I suddenly go 2.5m -> 1.25m -> 2.5m somewhere in the middle. And I've only had limited success trying to fight it with struts.

Do any of you regularly take smaller engines than the stack size would dictate? And if you do, how do you deal with the wobble?

Edited October 30, 2015 by Streetwind

[FancyMouse]

Not too regularly but I do it sometimes (most notably hitchhiker+909 and then land on places like Mun). I don't feel bad about it, and I haven't had any problem with it - 4 strut setup is usually enough for me. I always *feel* put the struts with a X setup is more stable than + when doing maneuver towards one of the + direction. No evidence, though.

Or maybe my payload is just not heavy enough so that struts can easily win.

[Geher]

MOAR STRUTS!

When I have a really long thin rocket that's very wobbly I put Cubic Octagonal Struts on the Side of the rocket and connect them with normal struts.

But for the case of the engine being smaller than the tank, 3 struts from the lower tank to the upper tank (the small engine is inbetween) should be enough.

[Streetwind]

But for the case of the engine being smaller than the tank, 3 struts from the lower tank to the upper tank (the small engine is inbetween) should be enough.

Well as I said, my results with struts have been mixed. How would I attach them for consistently good results? Straight up? Slanted inwards? Slanted outwards? Slanted sideways and crossed?

[sgt_flyer]

I generally have good results with slanted sideways and not crossing (triangles), but sideways and crossing works equally well (but you need to place them as close as the stage's side as possible)

Using adapters can also help (even the flat ones) - joining nodes of two different sizes is generally more wobbly

It's the designs of open lattice russian designs anyway. (Sideways Triangles on soyuz, sideways crossings on proton and N1 rocket)

I've used with good results this sideway crossings / triangles within my chaos star rockets, and they are quite stiff despite the payload's weight

http://forum.kerbalspaceprogram.com/threads/115732-V1-04-Chaos-Star-Series-1500-1200-and-500-tons-to-LKO-launcher-maximum-fairing-diameter

Edited October 30, 2015 by sgt_flyer

[Superfluous J]

I frequently put a 48-7s on a 1.25m stack and never have problems. Of course, the payloads are usually small too. I have used 3 or 4 vertical struts with good results.

I've used a terrier (or even a nuke or a swivel/reliant, though I haven't used those in space outside of early career in a while) on the larger tanks, too, and in those cases if the vertical struts don't work, 6 or 8 struts in sets of 2 in an X pattern does the trick.

It also helps sometimes to disable all control devices on your payload during launch, and put a probe core somewhere on the lifter and control from there on liftoff. Then if your payload wobbles, the lifting engines don't gimble all crazily trying to "correct" a problem that's not there.

[nothingSpecial]

That was really my first problem with KSP. I tried to simulate Ariane-like wide payload fairing with 2.5m payload on 1.25m lifter back in 23.0, and it didn't work well.

First of all - yeah, cube-strut-cube setup can help you.

Shortening the stack and investing in boosters (I believe) also can help you deal with wobble.

Generally speaking, try to not put rocket under wobble at all - low angle of attack, no phys-warp.

If you have some extra payload, not just fuel-heatshield-capsule, consider to put it behind your Terrier in stack and once in orbit just rotate the craft and dock to it, the way Apollo CSM did it with LEM.

Try to increase number of struts. I usually go for six or eight struts total.

If you feel like engineering, you can do something crazy like skycranning and multidocking.

[Snark]

Attaching engines radially works well for me. For example, if I have a 2.5m stack, radially attach a couple of the 4-ton-capacity 1.25m tanks, and put the engines on those. Then the butt end of my 2.5m stack can just attach directly to a decoupler. Nice and stiff (better, in fact, than if I had gone with a Poodle).

It's a win all around. It reduces the vertical height, which makes it less floppy. Shorter means lower moment of angular inertia, which means my reaction wheels get more bang for their buck. Having a couple of 1.25m stacks off to the side that aren't part of the main structural integrity core of the ship gives me a really convenient place to pack on stack-mounted components (reaction wheels, batteries, ore tanks, whatever) without boosting either the height or the part count of the ship's central core, which really helps stiffness a lot. The side-mounted engines mean I can use long, awkwardly shaped engines if I want to (such as LV-Ns), and they give a convenient mount point to stick on a couple of SRBs below them for launch. Having the outriggers out to the side also gives a convenient mounting location for laterally attached solar panels, since it moves them out from the center of the ship so they're less likely to get shaded.

[sgt_flyer]

Another possibility, is to use the translation tool technique to create a gap between the fuel tank and the decoupler, then put your terrier on a radial attachment point and radially attach it to the fuel tank

http://forum.kerbalspaceprogram.com/threads/43086-Open-Source-Construction-Techniques-for-Craft-Aesthetics?p=1709894&viewfull=1#post1709894

Basically, place your upper stage fuel tank, add a 2.5m decoupler directly under it. Use the rotation tool to turn the decoupler vertically by 90° (In 'snap' mode, then translate the decoupler downwards as far as you need (in smooth mode, to avoid unwanted sideways movements). Go back to the rotation tool in 'snap' mode to put back the decoupler horizontally, and you'll have a gap between the tank and the decoupler. Finally add your radial attachment point + terrier underneath the tank. No more 2.5 -> 1.25 -> 2.5 problem (you can add a few struts to connect the lower stage to the upper stage if you want, to keep a 'realistic' connection between the two

[cantab]

What I've always done is, from the top: 2.5m tank, 1.25m engine, 1.25m decoupler, 2.5m fairing base, 2.5m tank. Then take the fairing straight up to the tank above. Then three or four struts down from the bottom of the upper tank into the fairing base.

What I also do is put a probe core somewhere on my launcher, usually on either the first or second stage, and control from it during launch. A Stayputnik on a small nose cone does well and looks kinda neat. SAS responds to the movement of the controlling part, so this means that if the top of the rocket wobbles a bit it doesn't matter as much.

[Streetwind]

Using adapters can also help (even the flat ones) - joining nodes of two different sizes is generally more wobbly

I have done this before. The Kerbal President sent me a sternly worded letter on manners and propriety. I do not do this anymore.

It also helps sometimes to disable all control devices on your payload during launch, and put a probe core somewhere on the lifter and control from there on liftoff. Then if your payload wobbles, the lifting engines don't gimble all crazily trying to "correct" a problem that's not there.

Interesting idea... I shall investigate.

Basically, place your upper stage fuel tank, add a 2.5m decoupler directly under it. Use the rotation tool to turn the decoupler vertically by 90° (In 'snap' mode, then translate the decoupler downwards as far as you need (in smooth mode, to avoid unwanted sideways movements). Go back to the rotation tool in 'snap' mode to put back the decoupler horizontally, and you'll have a gap between the tank and the decoupler. Finally add your radial attachment point + terrier underneath the tank. No more 2.5 -> 1.25 -> 2.5 problem (you can add a few struts to connect the lower stage to the upper stage if you want, to keep a 'realistic' connection between the two

Now THAT'S an intriguing trick! Definitely going to try that one. Although it'll probably keep me from forming an interstage around the gap... or maybe not... must experiment.

[SpaceplaneAddict]

I just did. One tweakscaled jet fuel tank and a 1.25 nuke engine. Trying to fight a wobble? Try using just one tank then the engine. So Tank-Engine-Whatever

[Snark]

What I've always done is, from the top: 2.5m tank, 1.25m engine, 1.25m decoupler, 2.5m fairing base, 2.5m tank. Then take the fairing straight up to the tank above. Then three or four struts down from the bottom of the upper tank into the fairing base.

One issue with that approach is that since you're starting the struts from the top stage, their mass will stay with it after stage separation. That would be .2 tons if you have four struts. (Maybe not much of an issue if your top stage is really massive).

Adding the short, squat 2.5m-to-1.25m adapter on top of the fairing base allows placing the struts there and extending them upwards, so they'll get dropped she. You stage.

Or just omit the interstage fairing and eat the aero hit you take from the mid-stack roughness.

[Tex_NL]

My Mun/Minmus landers are usually pretty simple: Mk2 Lander-can, Rockomax X200-8 and powered by just a single LV-909 "Terrier" More than enough to land and get back to orbit.

The Poodle is slightly more efficient engine but weighs more than three times as much. The weight difference more than makes up for the slightly lower ISP.

Edited October 30, 2015 by Tex_NL

[celem]

I will say that if you are experiencing wobble across a weak joint like this then one of my main steps is to disable all the torque and RCS stuff from above the join. It doesnt help that on launch your main core is gimballing at the back and the pod up front is trying to torque, leads to aggravation of the flexing. Once out of atmosphere its less harmfull but imparts a lot of stress during atmopsheric flight if your rocket keeps assuming a banana-like profile.

[Van Disaster]

Pretty much re-iterating what's said before: put a core on your launcher - it helps if you build launchers seperately & payloads as subassemblies - and disable reaction wheels on the payload, make sure you "control from here" on the launcher's core. If you're running mods then iirc procedural fairings will auto-strut the payload, although I may have a faulty memory & mistaking that for auto-strutting the fairing. You might try rebuilding the launcher so the initial high-power boost stage is actually two boost stages attached radially, and the upper atmosphere/orbital insertion stage is the one under the payload - that way you can strut the payload to all stages.

As a last resort you could radially mount even smaller engines:

[Sovek]

My suggestion is using Procedural Parts (fuel tanks and fairings) and make the rocket the way you need it. I find clustering/smaller engines for upper stages to be very effective. My current 3 kerbal rocket uses 6 H-1 engines for the first stage (1.25m engines, on 2.5m tanks) and 3 RL-10s for second stage/transfer. CSM engine is like a .625 part.

[Norcalplanner]

Three letters - KJR