"Mammoth's" NOT-SO psychotic gimbal wobble

[Dire_Squid]

I've seen that the Mammoth seems to be more powerful now, in 1.0.2 (trying to create a Duna Rover, inside of a fairing, attached to a stack of super-heavy tanks, and two jumbotank "boosters" (using the Mainsails as engines, gimbal turned off).

However, I see that the gimbal for the Mammoth engine eventually begins to wobble, STRONGLY: making the rocket wobble, and I eventually see the rover eventually clip outside of its own fairing (it stops the moment I turn the gimbal off). I haven't gotten it down to a science, but the wobbly gimbal seems to be a BIT more controllable once I set the gimbal limit to 45, or less. I am unsure how bad the wobbly gimbal acts in a vaccum, atm.

Edited May 9, 2015 by Dire_Squid

[zekes]

It's "zero- BANG- zero- BANG"

But I'm okay with it....

[Dire_Squid]

It's "zero- BANG- zero- BANG"

But I'm okay with it....

Liiiiiike setting the Gimbal to 0 and then shooting it up all the way?

A shame because I LOVED using the Mammoths for my "power engines" for super-heavy craft. Rhinos USED to be good, but now they seem to have become underpowered (took about 20 seconds for the craft to SLOWLY lift... like "2 meters a second" lift. rofl).

[Reddeyfish]

a lot of engines seem to have a wobble problem. Also, >Dire_squad: the rhino is NOT a lifting engine; it's an upper stage engine (it has twice as much thrust in orbit as it does on the pad)

[zekes]

Liiiiiike setting the Gimbal to 0 and then shooting it up all the way?

A shame because I LOVED using the Mammoths for my "power engines" for super-heavy craft. Rhinos USED to be good, but now they seem to have become underpowered (took about 20 seconds for the craft to SLOWLY lift... like "2 meters a second" lift. rofl).

They're fine for me - In fact I built a 300-ton-to LKO rocket using just 4 of them (plus some SRBs to help)

[Dire_Squid]

They're fine for me - In fact I built a 300-ton-to LKO rocket using just 4 of them (plus some SRBs to help)

Well, I most likely put too many tanks or the Rhino to take, but a good amount for the Mammoth.

[Streetwind]

You can adjust the gimbal strength - even in flight.

Just turn it down to 20% or so. That should get rid of the overzealous wobble while still letting you steer with it. If it still wobbles... turn it down even further.

[Kuu Lightwing]

Well, I most likely put too many tanks or the Rhino to take, but a good amount for the Mammoth.

rhino is a vacuum engine. It loses half the thrust at ASL

[Renegrade]

rhino is a vacuum engine. It loses half the thrust at ASL

The atmosphere effectively ends between 4km and 12km for 'vacuum' engines, FYI. My Terrier-based spaceplane is like 0.1 Isp off of max Isp at 12km when the rockets kick in as an example.

[Jaeleth]

Wobbling seems to occur in 2 scenarios:

1. when the vehicle's ability to correct position is limited, either because is too heavy or because it is too long, i don't know the precise reason, thus making it respond slowly to the engine's inputs, thus overcontrolling, due to the PID controller response, and creating a positive feedback which increases the amplitude of the oscillation over time (resonance frequency achieved, in real life you can bring down a bridge with a large number of feet stomping it at an exact frequency), destroying the vessel or simply flippling it over. And the PID in KSP excels at finding that frequency

2. When the vehicle is not rigid enough, ending in a situation like 1.

Solutions:

1. The Boeing 737 pilot way: turn it all off and fly it like a man! meaning you must turn off KAS and fly manually, it is not that hard with enough flight hours under the hood and on big rockets. It will bump around a bit but you will get it to orbit. (Worked with me)

2. If it is wobbly, add struts, and struts... AND struts... Make sure the struts are separated by stages and the stages are connected with another set of struts, if needed. ( always works, but doesn't so,ve the whole of the problem)

3. Maybe increase the number of reaction wheels, this will minimize, theoretically, the vessel's response time to the engine inputs, thus reducing the frequency thus preventing the PID from finding that wretched resonance frequency. (I never verified this one)

4. Lock gimbal on all or some of the bigger engines and add some side stabilization engines if needed (cant remember the name of those, they are the ones who resemble saturn 5 first stage stabilization engines). (Worked with me)

Edited May 8, 2015 by Jaeleth
Misspelling

[rkman]

Makes one wonder why Squad does not use the gimbal response speed parameter that they have created.

In the mean time i'm using this modulemanager tweak:

@PART[*]:HAS[@MODULE[ModuleGimbal]]
{
	@MODULE[ModuleGimbal]
	{
	 useGimbalResponseSpeed = true
	 gimbalResponseSpeed = 5
	}
}

[Red Iron Crown]

Wobbling seems to occur in 2 scenarios:

1. when the vehicle's ability to correct position is limited, either because is too heavy or because it is too long, i don't know the precise reason, thus making it respond slowly to the engine's inputs, thus overcontrolling, due to the PID controller response, and creating a positive feedback which increases the amplitude of the oscillation over time (resonance frequency achieved, in real life you can bring down a bridge with a large number of feet stomping it at an exact frequency), destroying the vessel or simply flippling it over. And the PID in KSP excels at finding that frequency

2. When the vehicle is not rigid enough, ending in a situation like 1.

I would add:

3. You are using SAS in a mode other than stability assist.

Seriously, the other modes look really useful but are close to unusable because every time I activate them the ship starts to oscillate like a bowstring.

[Temeter]

I would add:

3. You are using SAS in a mode other than stability assist.

Seriously, the other modes look really useful but are close to unusable because every time I activate them the ship starts to oscillate like a bowstring.

Stability assist can get pretty wobbly too, especially if you use an engine like the Mammoth and don't lower the gimbal.

[Snark]

One potential issue which I raise because A) the OP mentions the lander clipping through the fairing, and I've been bitten by this myself:

Fairings don't stop physics inside. Or even give any physical support.

We all know by now that floppy rockets don't fly well, and we use all the usual design techniques to avoid such problems. However, much as I love the fairings, they behave in a way that I find counterintuitive, and would not be surprised to see others have trouble with as well.

The first issue is that the stuff inside a fairing undergoes full physical simulation, even though you can't see it. It will wobble. This is made worse by the fact that the fairing doesn't provide any physical support at all for its contents (which seems wrong to me). When they do start the first subtle swaying, it's invisible, and you don't notice until it gets so bad that it makes the whole rocket wobble, at which point it's easy to misattribute the problem (e.g. "My engine is psychotic" rather than "My rocket is floppy").

Making things still worse: I find that I'm considerably more likely to build a floppy rocket when a fairing is involved. The fairing looks big and solid when it's not, giving a false sense of security. Also, having the fairing there discourages me from using struts in some of the places I'd normally put them, such as where the strut would penetrate the fairing (the game does let you do that, but the behavior is buggy/inconsistent and it just feels wrong to me).

Moral of the story: Be careful with fairings, and firmly secure their contents.

Edited May 8, 2015 by Snark

[mikegarrison]

One potential issue which I raise because A) the OP mentions the lander clipping through the fairing, and I've been bitten by this myself:

Fairings don't stop physics inside. Or even give any physical support.

We all know by now that floppy rockets don't fly well, and we use all the usual design techniques to avoid such problems. However, much as I love the fairings, they behave in a way that I find counterintuitive, and would not be surprised to see others have trouble with as well.

The first issue is that the stuff inside a fairing undergoes full physical simulation, even though you can't see it. It will wobble. This is made worse by the fact that the fairing doesn't provide any physical support at all for its contents (which seems wrong to me). When they do start the first subtle swaying, it's invisible, and you don't notice until it gets so bad that it makes the whole rocket wobble, at which point it's easy to misattribute the problem (e.g. "My engine is psychotic" rather than "My rocket is floppy").

Making things still worse: I find that I'm considerably more likely to build a floppy rocket when a fairing is involved. The fairing looks big and solid when it's not, giving a false sense of security. Also, having the fairing there discourages me from using struts in some of the places I'd normally put them, such as where the strut would penetrate the fairing (the game does let you do that, but the behavior is buggy/inconsistent and it just feels wrong to me).

Moral of the story: Be careful with fairings, and firmly secure their contents.

Yep. My bigger rockets with fairings were behaving exactly this way. So I took a little more care with designing them for stiffness and I added the mod Kerbal Joint Reinforcement. It was like I waved a magic wand, and all that crazy wobbling disappeared.

The problem is not the engine. The problem is that the SAS is trying to steer at the cockpit, but it does so by using the engines (and/or fins) all the way at the other end of the rocket. So if your rocket isn't structurally stiff, the SAS gets out of phase with itself and your heading goes divergent.

[Snark]

...I added the mod Kerbal Joint Reinforcement. It was like I waved a magic wand, and all that crazy wobbling disappeared.

What I'd like would be a KJR that only applies to fairing contents (with reinforcement that disappears when you pop the fairing). Better yet, if stock just behaved that way.

I don't like to use KJR myself, because I like the design problem of making the rocket stable enough-- it's fun for me. But fairings are where that particular fun hits its boundary.

[mythbusters844]

Everyone's using these new engine names but I only know the technical names so I'm left in the dark .

EDIT: Seems the wiki has updated the engine pages with the new nicknames. Yay.

[Dire_Squid]

You can adjust the gimbal strength - even in flight.

Just turn it down to 20% or so. That should get rid of the overzealous wobble while still letting you steer with it. If it still wobbles... turn it down even further.

Well, that I know. I've manually shut the gimbal off, and turned the limiter to about 45%, but you've got a more solid point on brining it own to 20%.

rhino is a vacuum engine. It loses half the thrust at ASL

The atmosphere effectively ends between 4km and 12km for 'vacuum' engines, FYI. My Terrier-based spaceplane is like 0.1 Isp off of max Isp at 12km when the rockets kick in as an example.

OMG, THESE MAKES ALL THE SENSE!!! I could use it in 0.90 just fine, but now it makes SO MUCH SENSE that this is a vaccuum engine (I know it reads those stats in the VAB/Hangar, but I've just gotta get it in my head to intuitively know these things, like how I LOVE how nuclear engines worked excellently on Minmus, in terms of push and fuel efficiency, when getting to that low-grav hunk of mint ice cream).

Wobbling seems to occur in 2 scenarios:

1. when the vehicle's ability to correct position is limited, either because is too heavy or because it is too long, i don't know the precise reason, thus making it respond slowly to the engine's inputs, thus overcontrolling, due to the PID controller response, and creating a positive feedback which increases the amplitude of the oscillation over time (resonance frequency achieved, in real life you can bring down a bridge with a large number of feet stomping it at an exact frequency), destroying the vessel or simply flippling it over. And the PID in KSP excels at finding that frequency

2. When the vehicle is not rigid enough, ending in a situation like 1.

Solutions:

1. The Boeing 737 pilot way: turn it all off and fly it like a man! meaning you must turn off KAS and fly manually, it is not that hard with enough flight hours under the hood and on big rockets. It will bump around a bit but you will get it to orbit. (Worked with me)

2. If it is wobbly, add struts, and struts... AND struts... Make sure the struts are separated by stages and the stages are connected with another set of struts, if needed. ( always works, but doesn't so,ve the whole of the problem)

3. Maybe increase the number of reaction wheels, this will minimize, theoretically, the vessel's response time to the engine inputs, thus reducing the frequency thus preventing the PID from finding that wretched resonance frequency. (I never verified this one)

4. Lock gimbal on all or some of the bigger engines and add some side stabilization engines if needed (cant remember the name of those, they are the ones who resemble saturn 5 first stage stabilization engines). (Worked with me)

I may need to have to add struts to the rover, inside the fairing. That COULD be my integral issue, because the rest of the rocket is strutted enough where wobble is unseen, otherwise.

Perhaps if I unlock the gimbal of the Mainsail engines attached to the "boosters", and lock the gimbal of the Mammoth... and then there's the issue with adding more reaction wheels. I have two Medium-sized wheels on the rover, itself, and one big one located inside of the fairing, underneath a Rockomax decoupler... I know in 0.90, too many reaction wheels would ruin my day because the internal torque would build up and make the ship wobble "internally" (for those who've used Roverdude's Starlifter reaction wheels... you'd know EXACTLY what I'm talking about).

One potential issue which I raise because A) the OP mentions the lander clipping through the fairing, and I've been bitten by this myself:

Fairings don't stop physics inside. Or even give any physical support.

We all know by now that floppy rockets don't fly well, and we use all the usual design techniques to avoid such problems. However, much as I love the fairings, they behave in a way that I find counterintuitive, and would not be surprised to see others have trouble with as well.

The first issue is that the stuff inside a fairing undergoes full physical simulation, even though you can't see it. It will wobble. This is made worse by the fact that the fairing doesn't provide any physical support at all for its contents (which seems wrong to me). When they do start the first subtle swaying, it's invisible, and you don't notice until it gets so bad that it makes the whole rocket wobble, at which point it's easy to misattribute the problem (e.g. "My engine is psychotic" rather than "My rocket is floppy").

Making things still worse: I find that I'm considerably more likely to build a floppy rocket when a fairing is involved. The fairing looks big and solid when it's not, giving a false sense of security. Also, having the fairing there discourages me from using struts in some of the places I'd normally put them, such as where the strut would penetrate the fairing (the game does let you do that, but the behavior is buggy/inconsistent and it just feels wrong to me).

Moral of the story: Be careful with fairings, and firmly secure their contents.

... and then there's that. I'm thinking the payload MAY be the culprit.

Edited May 9, 2015 by Dire_Squid