Roll control and RCS

[LostOblivion]

When designing upper stages, I usually put four quad RCS ports on the bottom near the engine for attitude control. Having them there provides higher torque than having them up near the center of mass. This works great for pitch and yaw, however, when doing rolls, the RCS ports barely thrust at all, giving little or almost no roll control. For roll control, having them down away from the CoM instead of up near it shouldn't really matter...had they fired at full thrust as they do when placed near the CoM it would provide the exact same roll, right?

What gives?

[Retread]

Could we see a picture of your craft? There's no reason what you're describing shouldn't work as expected.

[Snark]

+1 for seeing a picture, would be very helpful.

Any particular reason you're using RCS for attitude control, instead of reaction wheels (which are "free", don't require spending monopropellant)? I generally use RCS only for lateral thrusting, e.g. when docking, and turn it off when rotating in order to save monopropellant.

[AlextheBodacious]

Can I just ask why you want this to roll?

[LostOblivion]

Don't have pics right now, but I'll get them. It's probably a bug/design flaw in how control inputs are translated by the game into RCS thrusts for the different ports, as it really makes no sense.

It's easy to replicate, just put four RV-105s on the bottom of a rocket without reaction control, and try to do pitch and yaw, works fine (in upper stages, you normally don't care about the incurred translation side-effect), but when doing rolls almost no thrust happens. Move the blocks up to the Center of Mass and you'll see more thrusts when doing rolls again (however now getting less pitch and yaw authority, as expected).

- - - Updated - - -

Can I just ask why you want this to roll?

I've modded reaction wheels to only have very little effect, forcing me to create working RCS setups for most of my craft. I don't have any at all on the current rocket, though, since its probe is the OKTO2.

[LostOblivion]

Here are some images. The RCS ports use are essentially the same as LV-105s.

https://www.dropbox.com/s/joijplosh8ep08g/screenshot553.png?dl=0

https://www.dropbox.com/s/wzxqprd44cvxwoc/screenshot574.png?dl=0

[GeneralVeers]

Can I just ask why you want this to roll?

Maybe he's using it as a lander that moves on a planetary surface by rolling....? Don't knock it, I accidentally did that with one of my landers once--it dropped on its side, the widest radius was the wide 360-unit fuel tank, and I discovered I could get moving at a pretty good clip, without burning any fuel, by using the SAS to roll the ship around.

Anyway, instead of RCS quads, my preferred setup is to use the SAS system for roll control and Vernier engines for translation. I like it when all the engines on a ship use the same fuel.

[DeMatt]

Assuming you're using fine controls (turn Caps Lock on, the control input indicators in the lower-left turn blue), then the RCS thrusters really aren't that far from your roll axis.

I'd also suggest putting a second ring of RCS quads on the opposite end of your rocket. It'll improve your handling SO much to have a balanced setup, instead of just the one ring on one side of the CoM.

[Thegamer211]

Try to place RCS thrusters on X4 symmetry.

[LostOblivion]

Assuming you're using fine controls (turn Caps Lock on, the control input indicators in the lower-left turn blue), then the RCS thrusters really aren't that far from your roll axis.

Doesn't really matter if I use fine controls or not, the ports still thrust very little.

I'd also suggest putting a second ring of RCS quads on the opposite end of your rocket. It'll improve your handling SO much to have a balanced setup, instead of just the one ring on one side of the CoM.

I do this for SMs and such, but for upper stages where all I care about is rotation, I don't need it.

- - - Updated - - -

Try to place RCS thrusters on X4 symmetry.

I do that already.

The problem is not the setup, but the fact that the game makes RCS ports far from the CoM thrust very little. Had they been thrusting at maximum, I'd have roll authority.

- - - Updated - - -

Like I said, it's probably a bug/algorithm design decision in the game. Try it yourself, you'll see that the game will reduce thrust for these scenarios, which removes any roll authority, while they could do full thrust and maintain roll authority.

[Junebug]

HeyLostOblivion,

I did a quick experiment, I have used the same RM-105-45 RCS thrusters as you, and put them on a rocket that approximates yours.

The problem is the Mini thrusters have not enough thrust to roll the rocket quickly, during pitch and yaw the thrusters are far away from the center of mass, this means they have a long lever.

For roll however the only half the diameter of your fuel tank can be used as leverage, to test this I also put a second set of RM-105-45 thrusters on booms (5 cubic octagonal struts).

I got the following results for a 60° turn in seconds.

Stock RCS: 7

RM-105-45: 22

On boom

Stock RCS: 2

RM-105-45: 6

So you could use more RCS thrusters or put them further away from the central axis.

P.s. Sorry if that sounds confusing English is not my first language.

[DancesWithSquirrels]

Moar thrusters.

Junebug is entirely correct - at the lever arm your thrusters are located they aren't going to have a great deal of authority, especially with that big tank. I end up stacking 3-4 sets of thrusters up the sides of the tank in cases like this - and either placing the groups in the vertical center of the tank or using CapsLock for pitch/roll since lots of thrusters at the ends of the tank are a bit OP.

Distance = leverage.

Either add more thrusters (and take the hit in monoprop consumption) or find a way to place them farther from the radial CoG.

[cantab]

While it's true that the "lever arm" for the arrangement in roll is short, the moment of inertia in roll will also be small, compared to pitch and yaw. The effects should at least partly cancel, so I'm not sure that can be the problem.

If the engine has a gimbal and it's enabled that will give substantial pitch and yaw control authority when the engine is thrusting. It can easily be more potent than any sensible amount of RCS or reaction wheels. However the engine will not give any roll control unless it's the rapier or maybe the mammoth. That might account for the discrepancy.

[Brainlord Mesomorph]

I've had some strange issues with roll and SAS.

Mine was with a large space plane. The plane would start a barrel roll and the SAS wouldn’t even bother to try to fix it.

You can see the control and SAS inputs in the lower left hand corner of the screen. I had plenty of role control manually, but as soon as I would take my finger off the Q and E buttons will roll indicator would just center, regardless of what the plane was doing. I could watch the sas trying to control the yaw and pitch. But it wasn't doing anything to correct roll.

So are you sure SAS is working?

[Junebug]

While it's true that the "lever arm" for the arrangement in roll is short, the moment of inertia in roll will also be small, compared to pitch and yaw. The effects should at least partly cancel, so I'm not sure that can be the problem.

That would be true for a sphere, LostOblivions's vessel is a cylinder the FL-T800 fuel tank alone is 2.4 times longer than wide, the rest adds even more weight while the diameter stays the same.

Edited November 3, 2015 by Junebug
typo

[LostOblivion]

That would be true for a sphere, LostOblivions's vessel is a cylinder the FL-T800 fuel tank alone is 2.4 times longer than wide, the rest adds even more weight while the diameter stays the same.

Thank you all for your inputs. While this is true, the problem here is not that the RCS ports lack leverage in roll, I know they lack leverage. The problem is that, for some reason, they don't thrust at 100 % when placed this far down on the longitudinal axis from the CoM and doing rolls. Move them up to the CoM, and they thrust at near 100 %, which is weird.

MY POINT, is that this shouldn't matter. They SHOULD have the same leverage in roll by thrusting 100 % no matter where on the longitudinal axis they are placed with respect to the CoM, however, the game seems to reduce thrust when placed far away from the CoM, resulting in lower than expected roll authority.

[Junebug]

Sorry for the misunderstanding, I could replicate the bug I put only 4x of the mini thrusters at the end of my test vessel, they fire if fine control was activated but it is not, toggling doesn't help. In my previous attempts I had multiple sets of 4x RM-105-45 and stock RCS on the vessel, the bug did not occur. That is very strange.

[DancesWithSquirrels]

What happens if you use stock RCS only? Might this be a problem with the mod?

[LostOblivion]

Thanks, you understand now. :-)

I've tried with the LV-105s too, same result. (Though, being angled 90° to the surface on which they are attached compared to the ones on the pictures, which were 45°, they have slightly more leverage.)

- - - Updated - - -

Sorry for the misunderstanding, I could replicate the bug I put only 4x of the mini thrusters at the end of my test vessel, they fire if fine control was activated but it is not, toggling doesn't help. In my previous attempts I had multiple sets of 4x RM-105-45 and stock RCS on the vessel, the bug did not occur. That is very strange.

I think Squad is trying to do something fancy when doing the calculations to yield thrust for the different RCS ports, and normally it works, but this is probably an edge-case scenario they hadn't thought of.

[Renegrade]

Any particular reason you're using RCS for attitude control, instead of reaction wheels (which are "free", don't require spending monopropellant)? I generally use RCS only for lateral thrusting, e.g. when docking, and turn it off when rotating in order to save monopropellant.

Some mods will basically require you to use RCS for any decent control (BTSM, my own Horrible Nerf). Also, although I'm not a realism nut, I do generally oppose things that are excessively un-realistic (or "#lolfake' as I like to call 'em), and the three-orders-of-magnitude-too-strong reaction wheels (leaving aside saturation as being a 'realism nut' thing) are definitely in that category.

MY POINT, is that this shouldn't matter. They SHOULD have the same leverage in roll by thrusting 100 % no matter where on the longitudinal axis they are placed with respect to the CoM, however, the game seems to reduce thrust when placed far away from the CoM, resulting in lower than expected roll authority.

Yeah, this is an old bug/design limitation of the RCS system. They should work as you describe, but they don't. It's been in there for as long as I can remember.

To the people who aren't understanding the issue:

Note how the thrusters at the end are producing almost no thrust during roll (tiny thrust vectors).

I suspect it's related to a stopgap measure to balance thrust (not entirely successful obviously ).

[pincushionman]

i suspect what's happening is the RCS blocks' thrust is being scaled by the magnitude of their distance to the CoM (a single factor), rather than scaling in each direction relative to the distance to the relevant axis.

That is, it must be a single "ThrustScale" value, whereas ideally there would be six separate factors for pitch, roll, yaw, and the three translations.

This makes me think of Scott Manley's "Tiony Probe" videos where he abused the hell out of the thrust-scaling mechanic by putting a single linear port at the probe's CoM and got some ungodly-high TWR because the game engine was trying to compensate for the nonexistent moment arm.

Edited November 5, 2015 by pincushionman

[LostOblivion]

Ah, good to know I'm not going nuts here. Good explanation on the stopgap measure.

I don't know if it's even possible to create a generic algorithm for RCS, but are there any mods that replace the stock RCS system?

[Renegrade]

i suspect what's happening is the RCS blocks' thrust is being scaled by the magnitude of their distance to the CoM (a single factor), rather than scaling in each direction relative to the distance to the relevant axis.

That is, it must be a single "ThrustScale" value, whereas ideally there would be six separate factors for pitch, roll, yaw, and the three translations.

I think you're right there -- that does seem to line up with what I've seen.

Ah, good to know I'm not going nuts here. Good explanation on the stopgap measure.

I don't know if it's even possible to create a generic algorithm for RCS, but are there any mods that replace the stock RCS system?

I'm actually pretty sure a more robust solution could be implemented, but until then, there's always ModuleRCSFX. I've not used it myself (it was mentioned in passing in the RCS Build Aid thread), but it sounds like it fits the bill.

[LostOblivion]

Was it fixed in 1.0.5?

From the changelog:

"Fixed some issues in RCS"