Problems with spin stabilization

[cami]

I'm currently looking at Realistic Progression Zero, and due to lack of lightweight avionics parts, I'd really like to spin-stabilize (the last stage of) sounding rockets.

But they can only spin-DE-stabilize.

Sounding rockets usually have parts mostly stacked on top of each other, and it seems individual parts don't have any angular momentum. So the rocket behaves more like a hair-thin object, and even at the structural limits every rotation around the longitudinal axis apparently remains forceless and useless. The funny thing is, because the spinning axis is never perfectly longitudinal, there is always a small transverse rotation (attitude change). And this rotation does have momentum, because the parts aren't exactly on its axis. In the end what you get is a vehicle with its tip and its tail spinning around each other likeahelicopter rotor.

Now  I have seen videos with successful spin-stabilization so I'm wondering whether I'm doing something wrong, or whether there's some mod I need fixing this? Im currently using KSP 1.1.3 with Realism Overhaul and Persistent Rotation mods.

[Voyageur]

The best thing without a screenshot I can suggest is to make sure your CoM is as low as you can get it. Adding a set of 3 or 4 fins to the bottom that are slightly angled create the best spin-stabilization I've come across. If you hold Shift+W or Shift+S to move the fins just once, that'll be your best bet. Hope this helps.

 

Edit: If one set of fins at the bottom doesn't work, you can try adding a set of fins for each stage angled the same way. 

Edited February 13, 2017 by Voyageur

[Nathair]

5 hours ago, Voyageur said:

make sure your CoM is as low as you can get it.

Low? As in closer to the fins? How does that work?

[Voyageur]

3 minutes ago, Nathair said:

Low? As in closer to the fins? How does that work?

A low center of mass prevents inadvertent rolling and ensures the rocket stays straight/aligned with prograde during ascent. Not necessarily as low as the motors/fins, so I guess I could have worded it better. OP didn't provide screenshots so I threw it out there in case the CoM was high enough to make the rocket oscillate/roll/flip. 

[Snark]

5 hours ago, Voyageur said:

The best thing without a screenshot I can suggest is to make sure your CoM is as low as you can get it.

Actually, this is the exactly diametrically wrong thing for aerodynamic stability.

You want the CoM to be as high as you can make it.  You want the CoM to be right up there at the front of the rocket.

When a rocket flips during an ascent, the problem is that the CoM is too low-- specifically, it's below the center of dynamic pressure.  To be stable, the CoM has to be above the center of dynamic pressure.  There are two ways to accomplish this (not mutually exclusive):

• Raise the CoM.
• Lower the center of dynamic pressure.  (This is what "put fins at the back" does.)

 

Thus, the standard mantra for aerodynamic stability:

• Heavy, pointy bits at the front
• Light, draggy bits at the back

Definitely would like to see a screenshot, though-- that should quickly clear everything up and we can give you specific suggestions to fix the problem.

 

[Voyageur]

10 minutes ago, Snark said:

Actually, this is the exactly diametrically wrong thing for aerodynamic stability.

You want the CoM to be as high as you can make it.  You want the CoM to be right up there at the front of the rocket.

When a rocket flips during an ascent, the problem is that the CoM is too low-- specifically, it's below the center of dynamic pressure.  To be stable, the CoM has to be above the center of dynamic pressure.  There are two ways to accomplish this (not mutually exclusive):

• Raise the CoM.
• Lower the center of dynamic pressure.  (This is what "put fins at the back" does.)

 

Thus, the standard mantra for aerodynamic stability:

• Heavy, pointy bits at the front
• Light, draggy bits at the back

Definitely would like to see a screenshot, though-- that should quickly clear everything up and we can give you specific suggestions to fix the problem.

 

Whoops. I digress then. 

[DerekL1963]

8 hours ago, cami said:

Sounding rockets usually have parts mostly stacked on top of each other, and it seems individual parts don't have any angular momentum. So the rocket behaves more like a hair-thin object, and even at the structural limits every rotation around the longitudinal axis apparently remains forceless and useless. The funny thing is, because the spinning axis is never perfectly longitudinal, there is always a small transverse rotation (attitude change). And this rotation does have momentum, because the parts aren't exactly on its axis. In the end what you get is a vehicle with its tip and its tail spinning around each other likeahelicopter rotor.

Which is pretty realistic actually...  IRL it's hard to spin stabilize thing that are longer on the spin axis than across the spin axis, especially if they're low density.  Almost always they swap from spinning around the axis with the least inertial momentum (the long axis) to the axis with the most (the short axis).

[Nathair]

46 minutes ago, DerekL1963 said:

IRL it's hard to spin stabilize thing that are longer on the spin axis than across the spin axis, especially if they're low density.  Almost always they swap from spinning around the axis with the least inertial momentum (the long axis) to the axis with the most (the short axis).

I'm hard pressed to think of an example of that. A spiral thrown football doesn't suddenly change axes of rotation nor do arrows or spears turn into pinwheels half-way to the target.

[DerekL1963]

35 minutes ago, Nathair said:

I'm hard pressed to think of an example of that. A spiral thrown football doesn't suddenly change axes of rotation nor do arrows or spears turn into pinwheels half-way to the target.

A football isn't long and thin - and a spear isn't spin stabilized.   No wonder they're not examples of that.

[Nathair]

Just now, DerekL1963 said:

A football isn't long and thin

 

Are you suggesting that this effect only applies at a certain ratio of axes? Because a football is certainly long and thin compared to a soccer ball.

 

5 minutes ago, DerekL1963 said:

a spear isn't spin stabilized.

No? You sure about that?

This seems an odd suggestion because it seems to me that, if it were true, "spin stabilization" would not be a thing.

[DerekL1963]

3 minutes ago, Nathair said:

Are you suggesting that this effect only applies at a certain ratio of axes? Because a football is certainly long and thin compared to a soccer ball.

I'm not suggesting it - I'm stating it as what it is, a fact.  (Also see this.)  The ratio of axes matters, as does the distribution of mass (almost all of a football's mass in it's "rim"), the nature of the structure, and the time-of-flight (that is, it takes time for the the processes to happen).

[Nathair]

5 minutes ago, DerekL1963 said:

I'm not suggesting it - I'm stating it as what it is, a fact.  (Also see this.)  The ratio of axes matters, as does the distribution of mass (almost all of a football's mass in it's "rim"), the nature of the structure, and the time-of-flight (that is, it takes time for the the processes to happen).

I understand now. The problem is your example is an examination of how, eventually over a period of many days, a system with no active attitude control or spin maintenance can lose stability. That doesn't really apply here or to any of the examples given.

[Snark]

3 hours ago, Nathair said:

A spiral thrown football doesn't suddenly change axes of rotation nor do arrows or spears turn into pinwheels half-way to the target.

Spinning objects can be complicated.

Spears aren't a good example, because they're generally not spin-stabilized, and even if they were thrown with a spin around the longitudinal axis, it wouldn't be much.  I suspect that the spear's stabilization is some combination of simple inertia (i.e. a long skinny thing has a big moment of inertia relative to its mass, so doesn't suddenly start spinning easily) and perhaps some aerodynamic stability (i.e. a somewhat heavy spearhead at one end, which wants to be in the front of the object as it goes through the air).

I suspect that footballs also aren't a super great example-- I don't know what the ratio of maximum to minimum moment of inertia is, but I'd guess it's not that far removed from a sphere.  The long pointy ends of a football are relatively narrow-- I'd guess that the moment of a football wouldn't deviate all that much from that of a soccer ball of equivalent mass.

One example of a somewhat long, skinny object that is spin-stabilized around its longitudinal axis is a rifle bullet.  However... it's not that long and skinny (like a spear or a rocket), and also, I think they tend to be pretty carefully machined, which I would imagine would hold them steady long enough to do the job.  (I'm just hand-waving here, though-- I am not any kind of expert on firearms, I'm just speculating by applying general physics knowledge.)

I suspect that part of the problem of trying to spin-stabilize a very long, skinny object is that its skinniness causes it to have a very small moment of inertia, relative to its mass.  This means that unless you make it spin really fast, it simply doesn't have much angular momentum.  If it doesn't have much angular momentum, it doesn't take much torque to point that angular momentum in a different direction.  And given that it's long and skinny, that means any lateral forces applied at the ends have a big lever arm, meaning lots of torque to work with.  So you've got a system that, on the one hand, easily generates big torques trying to move it off-axis; and, on the other hand, doesn't need much torque to make it go off-axis.

And in the specific case of a rocket:  unless it's only going straight up, then its path needs to curve as it ascends, and a spin-stabilized rocket is going to have problems with that-- if it has enough spin & angular momentum to provide useful stabilization, that's going to make it resist staying prograde as the trajectory curves... and any torques that try to force it prograde (e.g. via fins on the back) are going to cause precession, like a wobbling top, which would be a Bad Thing.

Put that all together, and I suspect we're getting at most of the reason why rockets tend to be aero-stabilized rather than spin-stabilized.

[regex]

2 hours ago, Snark said:

And in the specific case of a rocket:  unless it's only going straight up, then its path needs to curve as it ascends, and a spin-stabilized rocket is going to have problems with that-- if it has enough spin & angular momentum to provide useful stabilization, that's going to make it resist staying prograde as the trajectory curves... and any torques that try to force it prograde (e.g. via fins on the back) are going to cause precession, like a wobbling top, which would be a Bad Thing.

This is probably the issue OP is facing. I've done many fin-induced spin-stabilized "straight up" sounding rocket launches, all pretty much successful and "as expected", but once I switch to trying to make an orbit I switch to aligned fins. During that sort of launch the place where spin-stabilization is helpful is when you're in or nearly in space, using slightly canted SRMs or something similar.

Spin-stabilized launch (ignore the navball, that's Principia's way of saying "You're at Vostochny". This thing is pointing straight up):

Spin-stabilized final stage (you can see the SRMs above the payload adapter):

[DerekL1963]

3 hours ago, Snark said:

One example of a somewhat long, skinny object that is spin-stabilized around its longitudinal axis is a rifle bullet.  However... it's not that long and skinny (like a spear or a rocket), and also, I think they tend to be pretty carefully machined, which I would imagine would hold them steady long enough to do the job.  (I'm just hand-waving here, though-- I am not any kind of expert on firearms, I'm just speculating by applying general physics knowledge.)

A rifle bullet is probably akin to the shells of big naval guns (something I am modestly knowledgeable about), while they're long and skinny - they're also both exceedingly dense.  That inertial mass makes them resistant (but not impervious) to swapping axes.  But both will, and do, tumble (in gunnery lexicon, a flat spin in aerodynamics and the problem the OP is having) in the 'right' circumstances.

As you say, it's complicated, and largely counter-intuitive.

[cami]

Wow, thank you for all the info, but I would like to remind of the OP. Im trying to stabilize sounding rockets (these are going straight up, not to stable orbit), and specifically high up inside, or entirely outside the atmosphere. I'm talking about the regime of less than 1 Pa dynamic pressure (FAR/RSS). I'm also using realism overhaul, so thrust is never perfectly on-axis, and I don't have thrust vectoring on the smaller engines.

I was spinning up the rocket as much as I could without tearing it apart, which is still rather little compared to real rockets (need to look at the numbers next time, but I'd estimate only around 1000 RPM). That might very well be the problem. The vehicles are about 80 cm (20") wide about 2,5m (8') long during that phase of the flight, and just about to leave atmosphere (70-100 km altitude on RSS Kerbin). I'm using actively angled fins (SAS off) at the tail of the vehicle for spin induction. This is ~halfway through the stage fuel and the CoM is rather close to the geometrical (circumsphere) center. I don't know how to figure out the center of drag, but it is definitely behind it: without thrust and without spin, the vehicle passively orients itself prograde. Edit: it is possible that the CoD moves in front of the CoM when the angle of attack is below 1° (so the air sees the fins head-on). Maybe that is enough to produce the effect?

For now I found a different solution, and put the Center of Thrust in front of the Center of Mass. Looks a bit Kerbal, but works x)

Edited February 14, 2017 by cami

[cami]

Concerning screenshots: I've already scrapped the design, but here's a photo of the current version of the relevant two stages. The lower stage has thrust vectoring and never needed the fins, the upper stage now has its CoT moved forwards and no longer needs to be spin stabilized by the lower stage.

Edit: oh, I did find a save with the earlier version:

Edited February 14, 2017 by cami

[cami]

14 hours ago, DerekL1963 said:

Which is pretty realistic actually...  IRL it's hard to spin stabilize thing that are longer on the spin axis than across the spin axis, especially if they're low density.  Almost always they swap from spinning around the axis with the least inertial momentum (the long axis) to the axis with the most (the short axis).

I did a bit of experimenting and added more parts off-axis (two long bars pointing straight outwards with a bit of mass at the tips). It didn't change anything.

While this thing is not nearly as sleek as successful designs I've seen in videos, I guess 1000 RPM is just not enough. Maybe I should try while the stage is still fully fueled.

Edited February 14, 2017 by cami

[Nathair]

On 2017-02-13 at 4:23 PM, Snark said:

Spears aren't a good example, because they're generally not spin-stabilized, and even if they were thrown with a spin around the longitudinal axis, it wouldn't be much.

Did you not see the links I posted about exactly this?

 

On 2017-02-13 at 4:23 PM, Snark said:

I suspect that footballs also aren't a super great example-- I don't know what the ratio of maximum to minimum moment of inertia is, but I'd guess it's not that far removed from a sphere.  The long pointy ends of a football are relatively narrow-- I'd guess that the moment of a football wouldn't deviate all that much from that of a soccer ball of equivalent mass.

No need to guess, really.

[John FX]

There`s also this for long, thin things with weight at one end.

EDIT :

Also depending on how well KSP models the liquid in a stage it may do this

 

Edited February 17, 2017 by John FX

[Snark]

2 hours ago, Nathair said:

Did you not see the links I posted about exactly this?

Yep, managed to miss those, thanks for pointing it out.

However, it's worth noting that in the context of spin-stabilizing a rocket, the spear analogy has quite a few differences.  Compared with a rocket, a spear is flying through the air for a much briefer time (thus less time for aero forces to have an effect), and at a much, much lower velocity (so aero forces will be a lot smaller, compared with the inertia of the projectile's mass).

That's all just hand-waving, of course, but given the significantly different numbers involved, I'd guess it would be a game-changer.

1 hour ago, John FX said:

Also depending on how well KSP models the liquid in a stage

KSP doesn't model liquid contents, at all.  The resource content of a container-- whether it's solid fuel, liquid fuel, ore, or anything else-- is just mass added uniformly.  For physics purposes, the fuel tank is simply treated as a solid cylinder of uniform density, and as the fuel is drained, it simply lowers the overall mass.

My evidence for the above assertion:

• Go to the VAB, create a "ship" consisting of a single fuel tank, turn on CoM display.  Use tweakables to adjust the fuel content of the tank from 0% to 100% to values in between.  CoM doesn't budge.
• Take a look at the config files for resource types (i.e. where it's defining the resource name, mass-per-unit, and so forth.  You'll find that there isn't any flag anywhere saying "is it a liquid or not".  So, the game has no way of even knowing whether a given resource is supposed to be a "liquid".

Aside from the above direct observations:  the hand-wavy argument (speaking here as a software engineer, though of course I don't work for Squad) is that trying to model liquid interactions inside the fuel tanks would be a great big fat hairy programming job, and would also add a lot of computational load to the physics model.  For not-very-useful-to-gameplay results.  So, I would guess that this would be the kind of feature that the game designers and programmers would stay away from.

[John FX]

8 hours ago, Snark said:

KSP doesn't model liquid contents, at all. 

I had literally no expectation that it would. Just showing examples of non-intuitive spin depending on object shape and contents.

[cami]

16 hours ago, John FX said:

EDIT :

Also depending on how well KSP models the liquid in a stage it may do this

 

Stock KSP doesnt, but maybe Realism Overhaul or Real Fuels adds this behavior. This actually matches what Im experiencing very well. Is there some resource on this effect? I would expect a hollow cylinder to be even more stable than a solid one of the same mass (of course the fuel container has much less mass than the fuel). So I'm wondering whether this also happens when the spin is much faster.

[John FX]

3 hours ago, cami said:

Stock KSP doesnt, but maybe Realism Overhaul or Real Fuels adds this behavior. This actually matches what Im experiencing very well. Is there some resource on this effect? I would expect a hollow cylinder to be even more stable than a solid one of the same mass (of course the fuel container has much less mass than the fuel). So I'm wondering whether this also happens when the spin is much faster.

I wouldn`t imagine than even RO or RF would go to the lengths of modelling liquid in a stage. It`s rather processor intensive.

It`s more likely to be along the lines of the T handle video where a mass further out from the centre line and towards one end causes instability.

On 14/02/2017 at 7:16 AM, cami said:

I'm using actively angled fins (SAS off) at the tail of the vehicle for spin induction

And here we in fact have extra mass at one end a little way out from the axis of rotation.

Try having some angled engines to provide thrust instead of the fins. My guess is that, like the T handle video your instability is caused by having the fins on the back end meaning the front and rear of your craft have different moments of inertia.

If you have an image of your craft it would help greatly in diagnosis.

[Zeiss Ikon]

Just to throw in another confounding factor...

From my reading, the main reason for spinning up a sounding rocket is so that imperfections in thrust alignment (caused by, for instance, deposits or erosion in the nozzle throat or bell) won't cause it to veer off course.  This doesn't require a high rotation rate (nothing like the tens of thousands of RPM applied to a bullet); a few revolutions per second, at most.   You'll probably find that the rocket isn't very prone to tumble anyway after engine burnout; at that point, even if its high enough for the fins to do nothing (which, depending on velocity, could be above 50 km in the stock Kerbin atmosphere), it'll only carry whatever angular momentum it had when thrust ended, so if it had near-zero tumble at that point it will keep that near-zero tumble until it comes back down far enough for the fins to give lift and stabilization again.

Sounding rockets usually do their engine burns early, where there's still enough air for fins to work well, and if the fins are stabilizing the rocket after burnout, you'll generally get a very good, stable coasting ascent.  Also, in the real world, a high spin rate will negatively affect the operation of many experiments (blurry photographs, completely wonky accelerometer readings, magnetometer output that's a sine wave), so it's usually preferred to keep spin low.

Plain static stability (COM ahead of COL) is what's needed for simple sounding rockets, just as it is for model rockets with Estes motors.