Problems with off axis torque in artificial gravity centrifuges (not torque effect)

[ajburges]

Having issues were my rotor pieces on stations. The counter-rotating arms are producing torque on the rest of the station that is misaligned with the center of rotation. Is there any way to eliminate this? Could this be caused by slight mass variations between arms?

Note this isn't from uncanceled torque. I have visually verified the arms are rotating WRT the station in opposite directions. Concepts that worked in 1.7.1 without this issue get rotation around 1 RPM. Net torque is not consistent and varies. I would say Kraken like, but the forces are tame and disappear when rotors are stopped. This uncontrolled attitude change makes it impractical for a station. Furthermore SAS refuses to correct for it (I have a control point orthogonal to the spin axis I intended to point in the normal heading).

Took a break around 1.7.1 when I got sidetracked and stagnated designing BG craft. Came back recently. I've been toying with designs waiting for mods to update and one of the items in design is concept artificial gravity habs for stations.

[Kerbart]

Simple solution: get rid of the counter-rotating arms. You really don't need them.

[Wobbly Av8r]

So the short story suggestion is to add reaction wheels sufficient to overcome the gyroscopic wobble but the SAS mode of those reaction wheel(s) must be one which is aligned with some non-vessel-defined reference, such as Radial Out or Normal, which references the body being orbited.

Venturing a guess as to why (longer story) is that, because rotating any mass creates an axis along which the torque is directed, if that axis does not align *perfectly* with coincident forces (direction of travel, for one), it will impart a "divergent" force.  Aligning TWO rotating axes would create a situation where EACH rotating body axis creates a feed-back loop with the other, as well as both influencing the direction of travel. So even if your rotational mass balance is identical, my guess is that round-off error might be enough to destabilize things; most folks call this the Kraken because of the way Unity resolves connected, harmonically destabilizing forces: Sudden Unplanned Disassembly.

So the bottom line is you may be able to salvage the situation if you have enough reaction wheel force to overcome the imbalance (whether by mass or round-off error) that is externally referenced to the vessel - otherwise you'll just give it yet another feed-back mechanism to summon you-know-who...

[Vanamonde]

Pics of the station in question? 

[ajburges]

On 1/3/2021 at 12:06 PM, Vanamonde said:

Pics of the station in question? 

Prepped an album with the latest contra rotating mechanism concept. Frustratingly, this is the the tamest one so far so it's less illustrative. I was able to eventually trim the oscillations out by redistributing ore mass as ballast (as intended for this design). With 1.11 adding Kerbal Mass, I think redistributable ballast is now the way to go.

I also have a single ring countered by a rotating ore tank in-line, but the significant difference in moments of inertia further complicate things as spin-up is unbalanced. Also have no idea how to approximate the balancing of the opposite forces of the ring mass vs the counterweight.

Edited January 5, 2021 by ajburges

[Vanamonde]

I don't know what the issue might be but I have a couple of suggestions.

In flight if you spin the camera around a craft, the default point the view pivots around is the center of mass of the craft. If that is not also the geometrical center then you know something is unbalanced with the masses of the parts.

Also, in the latest version of the game the mass of kerbals is added to the mass of the parts they're in. It's not clear from your screenshots, but are you seeing the problem if no kerbals are aboard or if there's a kerbal in one cabin but not its mirror? 

[ajburges]

Good thought on camera focus. I forgot about it because I use camera aiming to access the rotors easier on the prototypes.

Did the crew cabin mass get rebalanced as well? The asymmetric arms were carefully balanced in 1.7.1, but one holds over 3x as many kerbals. if empty mass was adjusted, that would explain how they got so out of whack. Regardless, the variable mass of cabin parts seems to necessitate a ballast system these days.

[Vanamonde]

I don't think there's been any change to the cabin masses except to reflect the presence or absence of kerbals. 

[ajburges]

Looks like they did change cabin masses while adding occupant mass. The one arm became 1 ton lighter fully loaded while the lab arm actually got a few dozen kg more massive fully loaded. Removing kerbals would increase the difference further.

Looks like the root cause was the part rebalancing completely broke the station, but it did expose that ballast (or careful berthing management) would be needed for future artificial gravity projects.

[ajburges]

Disclaimer: I was comparing values between 1.7.1 and 1.11.0. It is possible there was a change pre 1.11.0 that I missed. Ran both versions and compared parts in SPH with vessel mass readout.