Power and RCS are drained very fast when using Pilot controls

[katateochi]

If you use the pilot controls to hold the craft steady "stability assist" then power drain is like it was in 0.25 and before. But if you use the pilot to hold the on the other headings resources are consumed at an alarming rate. I noticed the power being drained v fast in my career so I tried it out in sandbox so I could check how RCS was used too.

Using anything other than stabilty assist drains the power as if you where holding the craft in a constant spin (full drain in other words). With RCS switched on, the problem is more obvious. All RCS ports are constantly firing and the craft vibrates.

We've seen this before though!! This was something that MechJeb used to do, but that was fixed in mechjeb a long time ago.

Is this known about or should I make a new bug report?

[edit]

This has now been confirmed as a bug in the bug tracker. http://bugs.kerbalspaceprogram.com/issues/3861

Thanks for your input guys.

[/edit]

Edited December 27, 2014 by Claw
bug tracker update - confirmed

[Tw1]

Strange that. Seems like something non-fully trained might do, though.

P.s. Thou is a pronoun, not the word that should be there. not tho either. "Though" is what you're looking for. ,

Edited December 17, 2014 by Tw1

[katateochi]

I wondered if it was to do with training, but its happens in sandbox just the same thou (doh!), though!

[Brotoro]

Yeah...five star pilots in sandbox are lousy at pointing ships, wandering back and forth past the target. This whole thing seems to be a huge step backwards to me. I was so happy when KSP finally got SAS working well several versions ago and we quit bleeding RCS fuel into space. Now I'm unhappy again.

[katateochi]

Yeah...five star pilots in sandbox are lousy at pointing ships, wandering back and forth past the target. This whole thing seems to be a huge step backwards to me. I was so happy when KSP finally got SAS working well several versions ago and we quit bleeding RCS fuel into space. Now I'm unhappy again.

You mean this is a feature!!? oh. I thought it was a bug (and a pretty silly one at that)! If that is intended thou, it makes no sense; a 1 star pilot in career can hold the craft steady at any arbitrary heading, but asked to go pro or retro grade they make it twitch like crazy. And to have the sandbox pilots doing it too, that's really daft. Yeah this is a huge step back, as in; its totally useless!

[Claw]

No, this isn't intentional. It's not intended to bleed the ship dry of resources by being "untrained."

Thanks for the feedback. There's been a standing SAS problem with lightweight ships being over controlled by excess torque. Maybe try using smaller torque (which I know is not possible with just a Mk1 pod) and turning off the RCS to see if that helps. You could also right-click disable the pod torque.

Cheers,

~Claw

[pacbard]

I got the same issue and I was waiting for someone to report it too (I didn't want to sound more whiny than necessary ).

Anyways, I noticed that the pilot jiggles the controls very fast left-right and up-down (rotation seems not to be an issue). Sometimes, however, it's just fine (for example when returning from an EVA and activating the SAS control) for the same ship. I suspect that there is a floating-point kraken hidden somewhere in the code (similar to the one that jiggles the apoapsis/periapsis when the orbit is very close to circular).

The electrical power drain is the most bugging issue, especially with the mk1 pod. When you have 50 power and you spend 0.25 trying to hold prograde, your flight becomes very short unless you deactivate SAS (which at that point defies its purpose).

Hopefully the issue will be taken care of pretty soon.

[Majorjim!]

If you use the pilot controls to hold the craft steady "stability assist" then power drain is like it was in 0.25 and before. But if you use the pilot to hold the on the other headings resources are consumed at an alarming rate. I noticed the power being drained v fast in my career so I tried it out in sandbox so I could check how RCS was used too.

Using anything other than stabilty assist drains the power as if you where holding the craft in a constant spin (full drain in other words). With RCS switched on, the problem is more obvious. All RCS ports are constantly firing and the craft vibrates.

http://i.imgur.com/DX3AOill.jpg

We've seen this before though!! This was something that MechJeb used to do, but that was fixed in mechjeb a long time ago.

Is this known about or should I make a new bug report?

Yup hold course wobbles and shakes and is generally unhelpful.. Bug report, its shoddy coding.

[katateochi]

Yup hold course wobbles and shakes and is generally unhelpful.. Bug report, its shoddy coding.

Done. Here's the bug report I submitted - http://bugs.kerbalspaceprogram.com/issues/3861

[Xaver]

As i said in a similar other thread: If you want you can call it a bug. But i am sure thats the law of controll system Engineering. Like when you have cruise control in your car. You will need an well-fitting equation to get those things stable.

[katateochi]

As i said in a similar other thread: If you want you can call it a bug. But i am sure thats the law of controll system Engineering. Like when you have cruise control in your car. You will need an well-fitting equation to get those things stable.

Are you referring to something like "Hunting oscillation"? Sure systems oscillate around a point, but that doesn't mean you can't engineer a solution to account for it.

Mechjeb used to have this same problem, but it was fixed and it no longer suffers that oscillation. The default SAS (stability assist) also doesn't suffer this problem.

I'm calling this a bug because the behavior is just the same as something that was considered a bug in mechjeb and was subsequently fixed.

If this was in by design then why doesn't stability assist have the same "feature"?

[Xaver]

Are you referring to something like "Hunting oscillation"?

exactly! When the possible torque-moment of inertia of area-ratio is to high the equation leeds to oscillation.

I don´t say that it isn´t possible, cause the new SAS has solved this problem to. But remember that the old SAS had this problem to.

Planes with SAS was sooo wobbly

Edited December 17, 2014 by Xaver

[Ziff]

As i said in a similar other thread: If you want you can call it a bug. But i am sure thats the law of controll system Engineering. Like when you have cruise control in your car. You will need an well-fitting equation to get those things stable.

It's definitely a bug. The Kerbals have no problems holding a steady heading in SAS mode with RCS. If I point him prograde and turn on SAS, he holds it just fine. If I tell the Kerbal to hold prograde he suddenly starts using RCS and jittering like he had 12 cups of coffee. It's a bug.

Edit: I read your other post, this isn't the same thing we are talking about. The other post is asking why autopilot sways past the target and back to it. That's not what this thread is discussing.

Edited December 17, 2014 by Ziff

[TriggerAu]

Anyone able to confirm whether it's happening for larger vessels as well, or just the smaller ones?

[MartGonzo]

Yeah...five star pilots in sandbox are lousy at pointing ships, wandering back and forth past the target. This whole thing seems to be a huge step backwards to me. I was so happy when KSP finally got SAS working well several versions ago and we quit bleeding RCS fuel into space. Now I'm unhappy again.

I'm glad someone said it, I love KSP to bits and don't want to criticize but it's really hard to see this kerbal experience thing as a good idea, to me it seems like trying to make the game into something else, some kind of sims in space.

[RocketPilot573]

The pilots seem to be using the original ASAS logic here...

[pacbard]

Anyone able to confirm whether it's happening for larger vessels as well, or just the smaller ones?

Just smaller ones.

In fact, a very wobbly small vessel (i.e., command pod, tank, engine, and rcs) that "fixes" itself when docked to a larger ship.

Also, I noticed that the problem worsens when you try to hold a moving market (e.g., retrograde marker when docking).

[Brotoro]

Anyone able to confirm whether it's happening for larger vessels as well, or just the smaller ones?

I asked my five-star pilot to point a very large ship in a specified direction... He had the thing swinging all over the place trying to get it pointed...overshooting...coming back...overshooting. This was using torque (of which the ship has plenty) and not RCS. After a while being disgusted by this performance, I manually pointed the ship (which didn't take long...it DOES have a lot of torque). After that, five-star pilot could hold the heading.

[sal_vager]

This can happen to large craft if you have too much reaction wheel power, SAS seems unable to settle and I suspect it's just a tuning issue.

Regex made a PID controller some time ago that might be able to help with this, other than that all we can do is wait for 0.91, and remember to change the SAS mode from directionhold to stability assist.

[Claw]

Anyone able to confirm whether it's happening for larger vessels as well, or just the smaller ones?

I'd say it's less of a "small vs. large craft" problem and more of some magic ratio of torque for the weight.

It's very easy to see with small craft because there are a lot of cases where even just the Mk1 pod has too much control authority. Planes are also suceptible to this because of all the flight controls. Larger ships tend to suffer less because the user has to add that much torque.

I haven't dug into it, but it seems like when the SAS is "really" close, it just puts too much input in (when there is a high torque/mass ratio) causing a small overshoot, where the process repeats itself. Simply damping down the input rate doesn't really fix it though because planes already suffer from a SAS unlock problem, where under some circumstances if you provide even the slightest input, the SAS is to slow to put its input back to keep the nose up, and the craft never comes back under control. Some of that is aircraft design, but the SAS will fail to keep up even if there is enough control authority available.

Cheers,

-Claw

[Ed Gauss]

I'd say it's less of a "small vs. large craft" problem and more of some magic ratio of torque for the weight.

It's very easy to see with small craft because there are a lot of cases where even just the Mk1 pod has too much control authority. Planes are also suceptible to this because of all the flight controls. Larger ships tend to suffer less because the user has to add that much torque. -snip-

This is what I have observed, as seen in these videos: http://gfycat.com/IndolentThinFerret http://gfycat.com/BareMaleJanenschia http://gfycat.com/PeskyWeeDuckbillcat

I've been tracking a similar issue in another thread. I'm convinced that this issue causes the autopilot to malfunction in cases of extreme torque-to-mass ratios. I've noticed some in this thread have observed the same thing as well.