High altitude rotors... I think I discovered a K-effect

[Guest]

I'm working on a reusable launch system for Eve, and am trying to work out the basics prototyping on Kerbin. Initial essays show that high-altitude rotors are surprisingly feasible. My test craft hauls a small payload up to 25 km altitude quickly and with little trouble, it climbs at 50-75 m/s on rotor power alone.

However I think I discovered a K-effect that kicks in somewhere past 25 km. The RPM on my motor goes off the scale, the meter reads well over 600 although the cap should be at 460. Around when it breaks 500 it starts climbing at a bonkers rate, well over 100 m/s and I think it went up to 160 m/s at some point. It's controllable a bit past that point, but there comes a point where even cutting power and hitting maximum brake won't stop the rotor (or, rather the freewheel under the rotor) -- the RPM just keep shooting up until the rotor tears itself apart. Picture related

Are there any others experimenting with pushing rotorcraft altitude? Any insights to share?

 

 

[AngrybobH]

I've noticed very similar behavior at ~21km with a significantly larger craft. Part of the problem appears to be the RPM limit does not do anything close to limiting the RPM. Lowering it seems to do the same as limiting torque(even before the limit is reached), max seems to be not a limit at all. The only way I have found to control rotor rpm when the air gets thin is to put in a huge amount of blade pitch. Which, in turn, usually stalls the rotor and sometimes over accelerates the freewheel. Alternating the brake and rotor power SOMETIMES bring the rotor and freewheel back into agreement. My plans for a similar Eve launcher is to abandon the rotor section at the pressure altitude that corresponds to 21km on kerbin. If I get lucky (not usually) I'll switch back to the rotor craft and attempt a landing for future re-use.

I really think the 460 RPM limit on rotors is because the game cannot deal with anything faster. And, it seems the Kraken starts to take notice ~300 RPM but is fully summoned by 500 RPM. What the Kraken does when he comes is, as usual, completely random. Rotors need some serious looking at by @SQUAD.

[KerikBalm]

I seem to recall seeing something like that on a pre-breaking ground thread on stock rotors, will try to find it tomorrow

[KerbMav]

12 hours ago, AngrybobH said:

Rotors need some serious looking at by @SQUAD.

No. Squad needs to give us actual propellers.

Preferably electrical and fuel/air powered ones.

 

[KerikBalm]

And some of the following posts.

My guess is that the same thing is going on here

Edited June 22, 2019 by KerikBalm

[Azimech]

1 hour ago, KerbMav said:

No. Squad needs to give us actual propellers.

Preferably electrical and fuel/air powered ones.

 

What, a model and an animation, with something that provides thrust due to the configuration file?

What's the fun in that?

Stock rotors and propellers are good enough, people have even been able to break the sound barrier, impossible in real life.

[KerikBalm]

Not impossible, impractical. On a related note: They made a supersonic propeller prototype, it was apparently incredibly loud (Sonic boom and all).

A prop engine part, like a jet engine, would help keep part count down, and be a lot easier to use. I know one person on this forum complains a lot about the rotors, and constantly understates what they can do... I guess because they can't figure out how to use them well.

I imagine that an easier to use option would be well received by the community in general

[Guest]

3 hours ago, Azimech said:

Stock rotors and propellers are good enough, people have even been able to break the sound barrier, impossible in real life.

There is one limitation with stock propellers that I would like to see addressed: efficiency with small diameters. I wouldn't object to a new "rotor/propeller blade" part that provided adjustable pitch out of the box, and implemented some fudge factors so that they worked at somewhat realistic efficiencies at low diameters. They wouldn't really change much other than cosmetics for big rotors, but would make it possible to make functional propeller aircraft that looked similar to RL ones.

As things currently stand, making a rotor involves a fair bit of rote, tedious fiddling -- placing servos, placing aero surfaces on the servos, adjusting servo min/max angle, binding the angle to an axis. There isn't anything particularly creative, challenging, or interesting with that process and I wouldn't mind a stock part that removed some of the tedium.

[Azimech]

On 6/22/2019 at 1:44 PM, Brikoleur said:

There is one limitation with stock propellers that I would like to see addressed: efficiency with small diameters. I wouldn't object to a new "rotor/propeller blade" part that provided adjustable pitch out of the box, and implemented some fudge factors so that they worked at somewhat realistic efficiencies at low diameters. They wouldn't really change much other than cosmetics for big rotors, but would make it possible to make functional propeller aircraft that looked similar to RL ones.

 

The only possible solution: an increase in max rpm. Time will tell if Unity will be able to do it.

 

Quote

As things currently stand, making a rotor involves a fair bit of rote, tedious fiddling -- placing servos, placing aero surfaces on the servos, adjusting servo min/max angle, binding the angle to an axis. There isn't anything particularly creative, challenging, or interesting with that process and I wouldn't mind a stock part that removed some of the tedium.

Just like real life ... the development of the propeller has been going on since 1901 and will continue ... probably for ever.

"Back to the drawing board, guys!"

By the way ... some actually like this design/test process.

Edited June 26, 2019 by Azimech

[Guest]

42 minutes ago, Azimech said:

By the way ... some actually like this design/test process.

I like the design/test process, that's the main reason I play KSP these days. 

I'm not a big fan of mechanical rote repetition however, and building a rotor with the current parts involves a lot of mechanical rote repetition.

42 minutes ago, Azimech said:

The only possible solution: an increase in max rpm. Time will tell if Unity will be able to do it.

It's not the only possible solution. An alternative is to fudge it -- create a propeller blade part with fudge factors that make it behave like a propeller would at the desired max RPM. That's the main reason I suggested it -- hard limits of the game engine are difficult to break through, but they can be worked around with fudge factors.

 

Edited June 26, 2019 by Guest

[Azimech]

52 minutes ago, Brikoleur said:

I like the design/test process, that's the main reason I play KSP these days. 

I'm not a big fan of mechanical rote repetition however, and building a rotor with the current parts involves a lot of mechanical rote repetition.

It's not the only possible solution. An alternative is to fudge it -- create a propeller blade part with fudge factors that make it behave like a propeller would at the desired max RPM. That's the main reason I suggested it -- hard limits of the game engine are difficult to break through, but they can be worked around with fudge factors.

 

I invested some time into that in the past. Place the CoM at the attachment point but place the CoL at the far edge ... I actually wasn't impressed with the result back then.

[Guest]

11 minutes ago, Azimech said:

I invested some time into that in the past. Place the CoM at the attachment point but place the CoL at the far edge ... I actually wasn't impressed with the result back then.

That sounds like a possible solution. You could even place the CoL outside the visible and physical radius of the propeller. What didn't you like about the result?

[Azimech]

Just now, Brikoleur said:

That sounds like a possible solution. You could even place the CoL outside the visible and physical radius of the propeller. What didn't you like about the result?

Ehm ...  I don't recall  atm. It's been ... three years?

 

[m4ti140]

On 6/22/2019 at 8:55 AM, KerbMav said:

No. Squad needs to give us actual propellers.

Preferably electrical and fuel/air powered ones.

 

Building props yourself out of rotors and aerodynamic surfaces gives you a more accurate simulation of props/lifting rotors than a jet posing for a prop engine would and therefore gives you more insight into what's happening in rotating systems like those - which is what we want from an educational game like this in the first place. Firespitter does fake rotors already and it's bad. Look up the Mi-26 craft someone made with Breaking Ground with fully articulated rotor hub and a swashplate - it accidentally got an almost X-plane tier flight model due to each blade section's aerodynamics being individually simulated. Of course they could make them procedural and simulate all of the effects (like center of lift shifting to the left in forward flight) instead, but I don't see them doing that.

What they should do is "just" fix the rotors themselves and change the way physics calculations are done for the parts attached to its nodes (i.e. use radial coordinates, constrain radial direction so that they don't "fly away" due to centrifugal force and perhaps do multiple physics loops per frame for rotating parts, basically give the rotating section its own physics grid that rotates with it - all of which might require circumventing how Unity handles stuff in one way or another, thought it's not like they haven't done it before).

Edited July 5, 2019 by m4ti140

[Azimech]

1 hour ago, m4ti140 said:

Building props yourself out of rotors and aerodynamic surfaces gives you a more accurate simulation of props/lifting rotors than a jet posing for a prop engine would and therefore gives you more insight into what's happening in rotating systems like those - which is what we won't from an educational game like this in the first place. Firespitter does fake rotors already and it's bad. Look up the Mi-26 craft someone made with Breaking Ground with fully articulated rotor hub and a swashplate - it accidentally got an almost X-plane tier flight model due to each blade section's aerodynamics being individually simulated. Of course they could make them procedural and simulate all of the effects (like center of lift shifting to the left in forward flight) instead, but I don't see them doing that.

What they should do is "just" fix the rotors themselves and change the way physics calculations are done for the parts attached to its nodes (i.e. use radial coordinates, constrain radial direction so that they don't "fly away" due to centrifugal force and perhaps do multiple physics loops per frame for rotating parts, basically give the rotating section its own physics grid that rotates with it - all of which might require circumventing how Unity handles stuff in one way or another, thought it's not like they haven't done it before).

It's not often I give someone a like on this topic.

[m4ti140]

OK, having massed around with rotor configs (I set the torque on the smallest rotor to something ridiculous, attached stuff to it and observed) I came to a realization that the RPM limiter is not actually an RPM limiter at all. It simply adjusts the max torque using a pre-calculated function of whatever you set on that slider and current speed, with no regard for rotor loading whatsoever. If there's literally anything attached to the rotor it will never reach the RPM value you set it to, because whatever method they use to calculate the slope of that function is only applied to an empty rotor, not to the actual craft upon lunch. It will apply the same torque at the same RPM regardless of what the actual torque requirements to maintain that RPM are. This is broken and is the reason both why everyone seems to have massive issues building props and the reason for the glitch in the OP. An RPM limiter has to be an active controller, the current solution is simply broken.

[Jesusthebird]

At the end of the day..i like the concept..but when i build a craft with a prop engine...i use airplane plus mod...less part count..manageable..easy..especally compared to the torque induced by a single prop rotor setup from stock. Dont think any airplane plus prop has a torque inducing property.