Do we all reduce gimbal in our engines?

[JackBush]

Or not? I've read a number of times that taking the Gimbal number down to 20-25 is a good way to reduce wobble-tipping-over-spinning. Do we all do it? Only a few? And to what levels?

[MagicFireCaster]

I almost never do it, not even with the vector 

[Evanitis]

Really rarely. A well designed rocket doesn't even thinks about tipping. Though if I'm lazy to design something well and just slap some random tanks, engines and fins on my payload, usually doing the gravity turn slowly and carefully is enough. If I fail the first time and can only be bothered to do a few clicks before the next launch, seting 50% gimbal is an option.

[ineon]

I only ever reduce the gimbal if I am following the prograde, but the rocket is really shaking. No idea if it really helps, but it stops it shaking as much, and Scott Manley did it...

[Gaarst]

Same as above: I rarely limit the gimbal on my engines, unless they are in big clusters which tend to be quite violent when touching the controls.

Using KJR and being gentle on the controls help a lot too.

Edited January 22, 2016 by Gaarst

[FungusForge]

Only with the vector on plain ol' rockets.

[RizzoTheRat]

Occasionally if I've got a fairly tall rocket and it's wobbling a bit.

[Dilli]

Only occasionally on liquid boosters stuck to the sides of the rocket with chewing gum (Radial Decoupler). Since upgrading to duct tape (Kerbal Joint Reinforcement), haven't bothered doing so anymore.

[Geschosskopf]

1 hour ago, JackBush said:

Or not? I've read a number of times that taking the Gimbal number down to 20-25 is a good way to reduce wobble-tipping-over-spinning. Do we all do it? Only a few? And to what levels?

I have never found any desirable outcomes for rockets using gimbal.  The effect of gimbal is to make the running engine act as a boat's rudder, causing a curve in the trajectory while under thrust.  This is normally counterproductive because burns are all vector addition and efficiency is maximized by burning in a single direction at the appropriate time.  Any off-line burning results in "cosine loss".  Thus, the only possible benefit of gimbal on rockets is to counteract a curve in the opposite direction (such as due to mounting the engine out of line with the rocket's CoM) so that the net thrust is in the desired direction, with the cosine loss accepted as the price for making the lop-sided design work at all.

I never make lopsided rockets, and I don't make structurally wobbly rockets, either.  As a result, for ROCKETS, I find gimbal useless for its intended purpose and have no need to turn it down to avoid structural wobbles.  So most of the time I just ignore it completely  I pick engines based on thrust and Isp, and either they have no gimbal or the amount of gimbal has no noticeable effect either positive or negative.  Thus, I never think about gimbal at all most of the time.  There are, however, a few engines that I might use for their combination of thrust and Isp, but which have stupid amounts of gimbal.  Like the Vector.  If you don't actually NEED the extreme gimbal to balance a lop-sided shuttle or something, it will throw even a solid, monolithic rocket around all over the sky during a burn.  So with such engines, I always disable the gimbal entirely.

The only other time I mess with gimbal is for SSTO spaceplanes.  I find that allowing gimbal to work in the atmosphere makes it pretty much impossible to for the spaceplane to hold the desired heading and AoA, and even harder to force it back into the desired attitude once it gets off it.  So I always disable the gimbal of spaceplane engines.

Now OTOH, if I'm making an atmospheric stunt plane, I want lots of gimbal (aka thrust vectoring) to make the plane more maneuverable.  But with these, I'm not concerned with holding a specific orientation (usually I'm not on a steady course at all except during takeoff and landing).

 

 

 

[SSgt Baloo]

I have found that educing gimballing to 30% prior to launch quells most of the ills excessive gimballing can cause, while keeping the virtues (having directional control is handy, at times).

[Kelderek]

It usually depends on how many gimballed engines I am using on a rocket.  Particularly in the 2.5m engine size where they all have gimbal, I will only have one engine at full gimbal and will dial the others down.

It also matters if you are using other control surfaces like tail fins.  Use one or the other for steering your rocket, but not both.  I often have better luck with immobile fins and a gimballed engine since the engine can still turn your ship once it is in a vacuum or in the upper atmosphere.  But what can really screw things up is when you have moving fins and also active gimbal on your engine, especially with SAS turned on.

[swjr-swis]

I only ever look at the gimbals when a craft is clearly uncontrollable due to it, which is not often, not even when I use vectors (and despite some pretty out there designs). And even then I tend to just disable it completely rather than tweak it.

I'd say about 95% of my engines stay on defaults, <5% get their gimbal disabled completely, <0.1% their gimbal gets tweaked to a lower but non-zero value.

[Snark]

1 hour ago, Geschosskopf said:

This is normally counterproductive because burns are all vector addition and efficiency is maximized by burning in a single direction at the appropriate time.  Any off-line burning results in "cosine loss".

True, but the cosine loss due to gimbaling is microscopic (unless you're talking about something with a huge gimbal, like the Vector)-- the cosine loss due to a 2-degree deflection is just 0.06%, and even that much is only when it's actually deflected and not pointing straight.  It's so tiny as to be utterly negligible.

My own rockets are almost always three stages to orbit:

1. SRBs, to lift off the pad
2. High TWR liquid-fuel engine to get out of the atmosphere and start the circularization burn
3. Low TWR, vacuum-efficient liquid-fuel engine to complete orbit.

#1 can't gimbal at all, 'coz SRBs.

#3, I usually disable gimbal completely, since it's a slow, low-TWR burn with no atmosphere to contend with and plenty of time to get pointed in the correct direction before I start the burn, and reaction wheels are plenty for doing any orientation adjustments.  (I've been sorely tempted to use ModuleManager to tweak the Terrier and Poodle to have locked gimbal by default, since I almost always want it locked and equally frequently forget to do so in the VAB.)

But I do like to have some gimbal in #2.  It's a handy adjunct for maintaining stability during the high-speed atmosphere transition from around 10-15 km up out of the atmosphere, like having steerable fins.  (And my 2nd stage is often constrained from having much in the way of fins, because it's up higher on the rocket when it's sitting on the launchpad, and putting fins would destabilize the rocket during stage #1.  Engine gimbal is a way of gaining some stability without fins.)

Emphasis on "some".  Too much, and it thrashes madly and is really irritating, regardless of practical consequences.  So I often find myself reducing gimbal, but it depends on the ship design-- specifically, very tall ships need less gimbal, because the engine is farther behind the CoM and therefore the gimbal produces more torque.

A common stage-2 for my larger ships will be a Mainsail under a Big Orange Tank, and in those cases I almost always cut gimbal down to around 25-30%.  For smaller ships (like a Swivel under the 4-ton LFO tank), I generally just leave gimbal as-is because it doesn't seem to hurt anything.

The primary case where I deliberately leave gimbal cranked way up is if I'm launching something that has a really awkward payload that's impractical to streamline, so I know that my aero stability will be crap, in which case full-power gimbal helps to make up for the ugly ship design and just muscle the ship onto prograde.

Edited January 22, 2016 by Snark

[Plusck]

Once I'm sure that my rocket is relatively stable in the ascent, I often disable gimbal manually to scrimp a few extra m/s dv from steering losses.

Yes, I could "design [my] rockets properly" from the start but where would the fun be in that?

[GDJ]

4 hours ago, JackBush said:

Or not? I've read a number of times that taking the Gimbal number down to 20-25 is a good way to reduce wobble-tipping-over-spinning. Do we all do it? Only a few? And to what levels?

I usually drop the gimbal down to 25%, sometimes just to 50%. Anything past that is not really needed. Between the engine gimbal and some aerodynamic fins on the bottom of the rockets spinouts are very rare, and once you get out of the 50Km altitude, one tends to have far less issues with instability.

In short, no more gimbal than one needs.

[FancyMouse]

I have two cases.

1. When I get to orbit and I need RT flight computer burn execution. Holding maneuver while burning will make the gimbal sort of "orbiting" the origin. With full gimbal it's even visible in ship view. I don't like that visually so I'll lower it.

2. When I have a lander using high gimbal engines (e.g. Thuds) for non-atmospheric landing (e.g. on Mun), when in orbit I'll lower gimbal and only restore full gimbal upon landing.

Edited January 22, 2016 by FancyMouse

[Geschosskopf]

8 hours ago, Snark said:

True, but the cosine loss due to gimbaling is microscopic (unless you're talking about something with a huge gimbal, like the Vector)-- the cosine loss due to a 2-degree deflection is just 0.06%, and even that much is only when it's actually deflected and not pointing straight.  It's so tiny as to be utterly negligible.

Agreed, but 2^ worth of gimbal isn't enough to justify having in the 1st place for a symmetrical rocket, and is too little to help with something as asymmetric as a shuttle.

The thing about gimbal is that it only has any effect when the engine is burning.  IOW, it's only useful in situations where you need to turn while under thrust.  Such situations are very infrequent for rockets (apart from permanently asymmetrical designs).  The vast majority of the time with rockets, turning and thrusting are separated.  You orient the ship while it's coasting, then thrust while not turning.  So apart from the initial launch and ascent, and maybe an especially tricky landing on a low-gravity world, gimbal does you zero good at all.  The rest of the time, you have reaction wheels and/or RCS to orient the ship while it's coasting.  And if you've got enough of that, a symmetrical rocket doesn't need gimbal even in the situations where it's potentially useful (asymmetric rockets are another story).

Quote

My own rockets are almost always three stages to orbit:

1. SRBs, to lift off the pad
2. High TWR liquid-fuel engine to get out of the atmosphere and start the circularization burn
3. Low TWR, vacuum-efficient liquid-fuel engine to complete orbit.

#1 can't gimbal at all, 'coz SRBs.

#3, I usually disable gimbal completely, since it's a slow, low-TWR burn with no atmosphere to contend with and plenty of time to get pointed in the correct direction before I start the burn, and reaction wheels are plenty for doing any orientation adjustments.  (I've been sorely tempted to use ModuleManager to tweak the Terrier and Poodle to have locked gimbal by default, since I almost always want it locked and equally frequently forget to do so in the VAB.)

But I do like to have some gimbal in #2.  It's a handy adjunct for maintaining stability during the high-speed atmosphere transition from around 10-15 km up out of the atmosphere, like having steerable fins.  (And my 2nd stage is often constrained from having much in the way of fins, because it's up higher on the rocket when it's sitting on the launchpad, and putting fins would destabilize the rocket during stage #1.  Engine gimbal is a way of gaining some stability without fins.)

Steerable fins on the 2nd stage are no problem, provided they're smaller than the fins on the 1st stage.  But some Vernor engines on the 2nd stage are even better, in which case having zero gimbal on that engine is no problem.  I have become a huge fan of Vernor engines.  They provide rock-solid stability during launch and might consume a total of about 5m/s if left running all the way from liftoff.

My own philosophy is to build the payload, then build the smallest-diameter single-stack lifter under it that can have at least 2400m/s and 1.5 TWR.  Then I use SRBs with 1000m/s and 1.5 TWR to get that off the ground, so the sustainer goes the rest of the way to LKO.  It has both fins for in the atmosphere and Vernors for orienting to circularize.  Once this stage is gone, the payload has enough reaction wheels or RCS to fend for itself.

 

[cantab]

I frequently do limit the gimbals. Now there are a couple of things to know about my flying. Firstly I use FAR, which as a side-effect increases gimbal range - Ferram reckons it helps control rockets in FAR. So often I'm effectively just undoing that. Secondly I use the hold prograde during my ascents; this is notoriously jittery, and reducing the gimbals keeps it under control.

I rarely outright lock gimbals on launchers though. If the engine has it then hey, (mass) free control! And they can offer *considerable* control authority, far exceeding what any reasonable mass of reaction wheels or sensible number of Vernors can give. My big Tylo lander relied primarily on the gimbals on its four KR-2Ls for control.

[Snark]

1 hour ago, Geschosskopf said:

Agreed, but 2^ worth of gimbal isn't enough to justify having in the 1st place for a symmetrical rocket, and is too little to help with something as asymmetric as a shuttle.

Actually, it can come in very handy.  For a marginally-stable rocket, those 2 degrees of gimbal have far more effect than, say, reaction wheels.  And yes, fins are good, but sometimes placement can be problematic.

It's not a silver bullet, nor is it always relevant.  But it's a handy tool in the toolbox and I'm glad it's there.

1 hour ago, Geschosskopf said:

But some Vernor engines on the 2nd stage are even better, in which case having zero gimbal on that engine is no problem.

Yeah, but I just don't like 'em.  I have no rational reason why.  They just irritate me, for some reason.  I'm not saying that using them is stupid, just that I'm irrationally prejudiced against them.  About the only time I use them is on the underside of airplane-style landers for worlds with very thin atmospheres, to help with stall speed on landing.  (Rare for me, but it occasionally comes up.  Duna.  Thatmo, in Outer Planets Mod.  Lave, in New Horizons.)

And if one is using Vernor engines for stability during launch... then they're serving exactly the same purpose as a gimbaling engine does.  Yes, it's not quite apples to apples, but I'd say that anyone recommending Vernors for launch stability needs to allow that gimbal helps with that, too. 

I suppose Vernors could be handy if you're using some engine that can't gimbal, like the aerospike.  Or maybe if the rocket is unstable enough that gimbal alone can't help it and it needs some additional nudges.  But I'm having trouble picturing a reason why one would turn gimbal off, and then add Vernors.  The gimbal is free, effective, and more fuel-efficient than Vernors.

I do use fins on the 2nd stage, to the degree that I can, but I can't go nuts with 'em, and I find that some engine gimbal really helps.

[Geschosskopf]

57 minutes ago, Snark said:

Yeah, but I just don't like 'em.  I have no rational reason why.  They just irritate me, for some reason.  I'm not saying that using them is stupid, just that I'm irrationally prejudiced against them.  About the only time I use them is on the underside of airplane-style landers for worlds with very thin atmospheres, to help with stall speed on landing.  (Rare for me, but it occasionally comes up.  Duna.  Thatmo, in Outer Planets Mod.  Lave, in New Horizons.)

And if one is using Vernor engines for stability during launch... then they're serving exactly the same purpose as a gimbaling engine does.  Yes, it's not quite apples to apples, but I'd say that anyone recommending Vernors for launch stability needs to allow that gimbal helps with that, too. 

But if a few little Vernors solve the problem, then gimballing the main engine is certainly overkill

I used to distrust them myself, fearing they'd rob too much fuel from the main engine, but this turns out not to be the case.  Besides, when I was very small, TV stations signed off at night and signed on again early in the morning playing the national anthem and showing clips of rockets launching.  And the main rocket featured was the Atlas, which always had these little flames coming out the sides down at the bottom, and I thought that looked cool.  So now that I trust Vernors, I use them to get that cool Atlas look

[Evanitis]

2 hours ago, Geschosskopf said:

trust Vernors

You just made me realize my logic:

'ahh, I don't want those sleek little engines eat away my precious fuel, let me just carry very heavy monoprop separately with ugly exhausts around..'

[Geschosskopf]

7 hours ago, Evanitis said:

You just made me realize my logic:

'ahh, I don't want those sleek little engines eat away my precious fuel, let me just carry very heavy monoprop separately with ugly exhausts around..'

This is, in fact, exactly what I used to do prior to the invention of Vernors.  And I did that a while afterwards, until I actually started messing with Vernors and discovered they're actually both extremely useful and benign.

Now, a lot of my payloads have mono aboard because they're intended to dock at some point.  To avoid depleting the payload mono on ascent while having RCS on so the Vernors will work, I lock the mono tanks in the VAB.  Then it sometimes happens that the payload rendezvous with its target and the RCS doesn't work.  D'oh!   Unlock the mono tanks