Control dynamics and physics

[Azimech]

Some ideas which could improve the game in lots of ways. What do you think, @JPLRepo?

 

1. Optional alternative control method.
Control surfaces only work with their assigned input depending on vehicle orientation. This means you can't have ailerons mounted at a 90 degree angle and expect them to function.

Why I believe an optional alternative method is interesting.
My largest helicopters are sluggish to control. Actually this is true for all large & heavy VTOLs. To improve control authority I use Juno's on the wingtips and tail. To have a fast response I use my invention, the Boost Flap, to block the thrust of the jets. But ... I have to use action groups. It would be terrific if ailerons would respond.

In these images, AG0 = roll to the right, AG8 = roll to the left, AG9 = cancel.

 

2. Changing control orientation is great for space but ...
Since no decent autopilot exists in stock I had to create my own. I use docking ports pointed at the sky and choose "control from here" and SAS setting "radial out".
This works reasonably except the game won't control the rudder, ailerons are reversed (countered with setting control authority to a negative value) and keys are reversed: A & D become roll, Q & E become yaw but reversed.

This picture shows the autopilot presets, controls pitch. Pitch down = forward flight. From left to right (in degrees) +1, 0, -1, -2, -3, -4, -5.

Why I would like an optional alternative method.
Using the docking ports or whatever part with the master control option, the user would finally have a good working autopilot without the hassle of the changed keyboard input and reversed ailerons.

 

3. Spool times of jet engines.
I used a stopwatch to measure the spool up time of the Juno. A full 38 seconds from zero to full thrust. There is no jet engine in existence, anywhere, with such poor and unrealistic values. Most engines require 4 to 10 seconds. Spooling down usually goes slower than spooling up but not in KSP.
Current values for the Juno, spool up: 0.12. Spool down: 0.5.
I recommend the following values, spool up: 1.12. Spool down: 1.0.

The Goliath is an even worse offender. Spooling up takes 2 minutes and 5 seconds, spooling down 38 seconds. A real engine company would get an FAA certification when hell freezes over.

Why I think this is important.
It would be a big step towards more realism and much better control while reducing the amount of frustration to some of the users, and it only takes a few minutes per engine to change the values.

 

4. Engine thrust vector physics, an idea for the future.
When the exhaust thrust of a jet or rocket engine hits an obstacle, two things happen:
1, a force is applied to that obstacle with the same vector as the exhaust thrust.
2, all energy is converted into heat.

Why I think this could be improved.
In real life the exhaust stream is deflected while applying a force to an obstacle depending on its angle. I propose the following: obstacle angle defines how much thrust is being deflected and the resultant force vector; how much is being converted into heat. Thrust loss depends on part temperature and ambient air temperature, the colder the engine the quicker deflected exhaust loses thrust.

For turbine builders this would mean an enormous increase in efficiency, using engine housing or stators to deflect thrust back onto turbine blades. This opposed to the current "turboshaft" engines being essentially "waterwheels". The users would start to get a feel for designing a more realistic gas turbine and it would mean a tiny step towards fluid dynamics. This requires extra calculations but a variable could be defined for the number of cascades and/or precision in calculation for low spec machines.
 

5. Easy way to increase the amount of available action groups.
I often have problems with assigning controls for my more complex machines. Using modifiers like CTRL, ALT and SHIFT the amount would be finally be sufficient.

[EpicSpaceTroll139]

1: Maybe some kind of compressor bleed air RCS would be the alternative control system? Anyways, when I get back from class, I can show you a way to set up boost flaps so you don't need action groups to control them. Even SAS can use them.

3: I would definitely appreciate faster spool times. Would probably be more likely to use Junos on my Helis if they spooled up realistically.

4. As much as I'd like realistic exhaust flow, I'd say to be careful what you ask for. I can't see much fun coming from spending a few hours optimizing a 300 part turbine with each blade care fully shaped from 12 thermometers.

Would it actually get greater efficiency though? Would be interesting to do some experiments to try to determine exactly how much energy we get from the exhaust and compare it to the engine's expected output.

5. Might want to stick to the ALT, considering the other two mess with the throttle.

[Azimech]

32 minutes ago, EpicSpaceTroll139 said:

1: Maybe some kind of compressor bleed air RCS would be the alternative control system? Anyways, when I get back from class, I can show you a way to set up boost flaps so you don't need action groups to control them. Even SAS can use them.

3: I would definitely appreciate faster spool times. Would probably be more likely to use Junos on my Helis if they spooled up realistically.

4. As much as I'd like realistic exhaust flow, I'd say to be careful what you ask for. I can't see much fun coming from spending a few hours optimizing a 300 part turbine with each blade care fully shaped from 12 thermometers.

Would it actually get greater efficiency though? Would be interesting to do some experiments to try to determine exactly how much energy we get from the exhaust and compare it to the engine's expected output.

5. Might want to stick to the ALT, considering the other two mess with the throttle.

 

Interesting :-)

1: It is true some boost flaps can be positioned differently. It doesn't change the fact the module that handles the actuation of control surfaces is a bit rigid in its implementation. When using my autopilot, the rudder still won't respond no matter what control input I give, except deploy. FAR does this better AFAIK.

4: Well ... maybe some other parts can be used than thermo's. The problem is that right now we have no way of knowing what could happen ... unless you know something I don't.

5: Good point ... what about the CTRL and SHIFT beneath the ENTER key?

[MiffedStarfish]

These seem like great ideas. I definitely agree with expanded action groups, maybe there could be a little "+" button after the usual 10, which when clicked comes up with option to bind it to custom keys. 

[EpicSpaceTroll139]

Ok so I forgot about this and so ran out of time to do a full demo, but here's the original craft.

https://kerbalx.com/jfrouleau/Jet-RCS-Demo

It's a VTOL that demonstrates a jet RCS system this guy thought would be useful on helicopters. While it doesn't fly perfectly out of the box (at least last time I checked it out), you can see some refinements I made in the comments. Basically it involves control surfaces set up to be controlled by steering which block/unblock the thrust of Junos as needed (similar to boost flaps, but no action groups needed). It even works with SAS (though this worked much better after the modifications I made).

Kudos to the guy who came up with this though. It's genius!

Edited June 20, 2017 by EpicSpaceTroll139
Oops

[qzgy]

9 minutes ago, EpicSpaceTroll139 said:

Ok so I forgot about this and so ran out of time to do a full demo, but here's the original craft. It's a VTOL that demonstrates a jet RCS system this guy thought would be useful on helicopters. While it doesn't fly perfectly out of the box (at least last time I checked it out), you can see some refinements I made in the comments. Basically it involves control surfaces set up to be controlled by steering which block/unblock the thrust of Junos as needed (similar to boost flaps, but no action groups needed). It even works with SAS (though this worked much better after the modifications I made).

Kudos to the guy who came up with this though. It's genius!

I don't see anything.....

[EpicSpaceTroll139]

28 minutes ago, qzgy said:

I don't see anything.....

Oops! Added the link now. 

[Azimech]

1 hour ago, EpicSpaceTroll139 said:

Ok so I forgot about this and so ran out of time to do a full demo, but here's the original craft.

https://kerbalx.com/jfrouleau/Jet-RCS-Demo

It's a VTOL that demonstrates a jet RCS system this guy thought would be useful on helicopters. While it doesn't fly perfectly out of the box (at least last time I checked it out), you can see some refinements I made in the comments. Basically it involves control surfaces set up to be controlled by steering which block/unblock the thrust of Junos as needed (similar to boost flaps, but no action groups needed). It even works with SAS (though this worked much better after the modifications I made).

Kudos to the guy who came up with this though. It's genius!

Seein' ya comment on antennae for the thrust vector.

Y'all mothers need Collide-o-Scope!

 

[qzgy]

1 minute ago, Azimech said:

Seein' ya comment on antennae for the thrust vector.

Y'all mothers need Collide-o-Scope!

 

Already have it!

[klond]

 Having customizable SAS would be fun and useful, especially if it saved the settings in the craft files.

 Not sure if action button modifiers would translate well to console controllers though.

[qzgy]

1 - I can see the use in that, although an intelligent bit of code that exists would have to be rewritten to accomodate for that fact. An override in advanced tweakables would work.

2 - Again, breaking a bit of code that exists. But I like it.

3 - These really need to be fixed. Not sure how they were made.

4 - The extra calculations would be a problem for potato owners. You've already mentioned good solutions, though you would have to consider also that the jet engines would heat up the air right? Also, IRL, I believe efficiency goes down with increased temperature and jet engines heat up the air, so having these in a closed environment would make efficiency suffer. Balancing out these effects would be hard. I think this would require a lot more discussion.

5 - What about using a number pad, if available? That would add an extra 10. Then if you include R-shift, R-ctrl, r-Ctrl+R-shift, each doubling 20, thats already 80. I don't think you can run out at that point. Only problem would be usability, and remembering them. A menu would be nice here.

 

In general though, I do like and agree with all the ideas.

[NSEP]

I never really use more than 3 action groups, but for a creator like you, it could be very usefull.

[Gman_builder]

I used to use elevons to block thrust to control yaw on tailless aircraft. Works basically the same.

[Azimech]

Since development of 1.4 has only just started, we might have a chance to make these items popular.

[NSEP]

2 minutes ago, Azimech said:

Since development of 1.4 has only just started, we might have a chance to make these items popular.

Sorry for being off topic, but, development of 1.4 has started?

[Azimech]

3 minutes ago, NSEP said:

Sorry for being off topic, but, development of 1.4 has started?

Yes :-)