PDA

View Full Version : Propwash/smoke interactions


phrank
12-06-2011, 02:56 AM
Picking up where flexibones left off and what Jim described about the new RF6 smoke behavior........
Jim states that RF6 smoke is calculated once when it exits, and then again when it hits the ground.

Pic1) Steam bar ahead of the prop - only inner trails are affected by the prop. (This is weird, wouldn't the smoke be affected behind the prop and not ahead?) :confused:
Pic2) Steam bar at CG - 6 inner streams affected by prop wash, 2 partial, outer ones unaffected.
Pic3) Steam bar behind aircraft - all steam bars affected.

This would suggest that the propwash comes out as a cone, not a cylinder.

I thought this was kind of interesting to demonstrate when in hover and the new aggressive torque roll effect in RF6.
A very small amount of the ailerons are being touched by the prop wash.

Jim, any possibility of getting more smoke calculations beyond the 2? ;)

td9cowboy
12-06-2011, 03:27 AM
Interesting Phrank. That explains why that small change to the yak to extend the ailerons to the fuse made such a big improvement. That's where all the prop wash is. We need a virtual wind tunnel. I guess you kind of made one. :p

flexible
12-06-2011, 08:24 AM
Sorry, Frank, I do not see any thing to support a (cone shaped) prop wash.

flexible
12-06-2011, 08:32 AM
Notice the smoke pods next to the fuse are being blown back hard, while the smoke pods on the outer wing are just being emmited.

Boof69
12-06-2011, 10:12 AM
I think phrank's experiment makes it obvious that the prop wash is cone shaped.

td9cowboy
12-06-2011, 10:32 AM
Phrank, did you have to lower the back torque to make that foamy hover? You're still holding a lot of throw on some mighty big ailerons there.

flexible
12-06-2011, 11:03 AM
You are both wrong. Did you even bother to do the prop wash test that I set up? All the proof is there. When there is no blow back on the smoke from the pods on the wings, except around the prop wash cylinder. I will even go one step futher, I think the blow back is stronger at the center of the prop wash cylinder center, than at the outer part of the cylinder. I did the test first with streamers, but found that streamers have no value, so I used smoke.
I have done this test in real life using pieces of create paper about 6" long, stuck on the trailing edge of the Ailerons on my 35% 260. The first piece was next to the fuse, and the others a couple inches apart, all the way to the end of the ailerons.
I tied the back of the plane down. With the engine idling, all the streamers hung straight down. As I added power, the streamers started to blow back. At full power, the streamers next to the fuse were straight back, while the streamers at the ends were only moving a little. Showing that in a hover, you have more aileron control next to the fuse, and less at the ends.
If you have a cylinder shaped prop wash, as RF does, you loose all ail control, from mid wing out. It is like having half an aileron.
In RF, the only way to get any control at all is by adding more prop wash. This is not real, for two reasons. First is the obvious, no prop wash over half of the ailerons. The second thing is, increasing prop wash, does not increase volume, but increases speed causing erratic feel, and the look of flying on a stick.
Conclusion, hovering in RF will not be even close to real world in these conditions!

jbourke
12-06-2011, 11:45 AM
Prop wash, both in the real world and in RealFlight, is essentially conical in shape. Within this cone the air moves in a slightly spiral pattern.

We can imagine the airflow as consisting of an inner and outer cone. The inner cone has a uniform velocity. The outer cone has a velocity inversely proportional to the difference between the inner and outer cone's radius.

The actual calculations are more complicated than that, since the fuselage and airfoils disturb the propwash, but the above is a good visualization. Also, presuming the aircraft is moving through space, the cone shape becomes elongated, skewed, twisted, or curved in response.

We made a concentrated effort to improve this aspect of the simulation for RealFlight 6.

Jim

jbourke
12-06-2011, 12:15 PM
Pic1) Steam bar ahead of the prop - only inner trails are affected by the prop. (This is weird, wouldn't the smoke be affected behind the prop and not ahead?) :confused:


The sim calculates the airflow due to the prop, including the airflow being sucked into the prop disk.

Keep in mind that the airflow is only calculated one time for each particle. The smoke will not necessarily match the propwash.

Also, for performance reasons the airflow is only calculated in a small area in front of the prop. I think if you keep mucking around with the smoke you'll probably notice a cylindrical area in front of the prop where there is airflow. This isn't realistic but it's a needed simplification.


I thought this was kind of interesting to demonstrate when in hover and the new aggressive torque roll effect in RF6.
A very small amount of the ailerons are being touched by the prop wash.


The smoke particles are not a reliable way to visualize the prop wash, for many reasons. We have tools only available to us internally which we're using to evaluate the prop wash in regards to the torque roll behavior.

The smoke physics are a visual effect.


Jim, any possibility of getting more smoke calculations beyond the 2? ;)

Yes, of course there is a possibility. :)

We have everything we need to do smoke physics each tick, except that it is just way too expensive to be done in real time. Perhaps we can work out a few optimizations for a future release.

Jim

flexible
12-06-2011, 12:15 PM
Jim, I am sure you know how the prop wash in you're softwear is made. So why can't that be proven with the prop wash test I did, where I put two poles in the runway to hold the plane during run up. With smoke pods along the trailling edge of the ailerons, there is proven prop wash out a little ways from the fuse, then it abruptly stops. Starting about 40% out, the prop wash stops moving the smoke back in a streme. From 40% out to the end the smoke is just being emmited, with no push. This shows no air flow over the ailerons, which I conclude is why the aileron controll is so week in a hover.
RC 3D, has more to do with prop wash, than wing lift. If there is semingly no wash over more than half the ailerons.
This effect can allso be proven in the sim. Set up a 3D plane to hover well, then make a copy, and cut off 60% of the outer ailerons. In a hover, you want even miss the part thatis gone.

jbourke
12-06-2011, 12:32 PM
Jim, I am sure you know how the prop wash in you're softwear is made. So why can't that be proven with the prop wash test I did, where I put two poles in the runway to hold the plane during run up.


You don't need to use two poles. Just use the "lockPosition" console command. I like to bind this to the "O" key and I bind "lockOrientation" to the "I" key.

This will work better because you can fly the plane up into the air and then press both O and I, locking the aircraft in space, out of ground effect, etc.


With smoke pods along the trailling edge of the ailerons, there is proven prop wash out a little ways from the fuse, then it abruptly stops. Starting about 40% out, the prop wash stops moving the smoke back in a streme. From 40% out to the end the smoke is just being emmited, with no push.


Yes, that's intentional. The prop wash cone expands but typically not fast enough to cover the whole wing.


This shows no air flow over the ailerons, which I conclude is why the aileron controll is so week in a hover.


Yes, certainly if there is no airflow over the ailerons they can't contribute anything.


RC 3D, has more to do with prop wash, than wing lift. If there is semingly no wash over more than half the ailerons.


Have you contrasted these RealFlight findings with a test in the real world? We have, using an anemometer (wind gauge) to calibrate the prop wash in the sim.

We're reviewing all of our data now as we revisit the torque roll behavior.

With all due respect, we already know how the sim works, and chances are if you are making a casual observation about r/c planes, we're already aware of that, too. If you want to contribute then take a video showing a plane hovering perfectly vertically with zero crosswind. Keep it in a perfect hover while deflecting the ailerons to the left for 10 seconds, then to the right for 10 seconds. Post the video on youtube or post a message here letting us know the rate of rotation both ways. We will also need an rfvehicle file which exactly matches the real world aircraft. All of the masses need to be calculated, the CG must be in the correct position, expo values and throws must match, as well as airfoils, engine and prop. That's the sort of work we're engaged in right now.

Jim

phrank
12-06-2011, 01:10 PM
Thanks Jim,
It would be great if the user could configure the amount of pain they wish to administer to their rigs if we do get additional smoke resolution.

This new version is proving to be a great visual aide for my goat, I'm learning quite a few things myself.

TD9,
Yes, there is some a mild amount of cheating to allow the foamy to have a fighting chance against the torque.
It is also Gyro assisted so my boy doesn't get frustrated.

** I'm asked to share this, so if you have the Flatout from the EPs, this should work out for you.
Otherwise, it's really easy to make, I have symmetrical smoke emitters spaced out .1 on a movable pod, so that it can be moved forward/backward or rotated along an axis. (pic)

Knob varies the amount of smoke,
Smoke switch to enable smoke.

Enjoy.

flexible
12-06-2011, 07:04 PM
I did this test with my 35% TEXAS EXTREME 260. I have only EPP foamies now, so I can not do the other test vedio.

flexible
12-07-2011, 02:06 PM
Jim, I am a designer of RC planes, and most recently of Sim models. One thing all designers have in common, we all want what we do, to be the very best it can be.
I am not here trying to make you're blood pressure go up.
I am here trying to shed light. I am not a numbers guy. I am real world guy, with a lot first hand problem solving experience.
I feel very strongly that the Prop Wash model, in it's present form, is a large part of the flight problem in 6. I can not write the code for a new Prop Wash model, and give it to you for a gift. All I can do, is try to put out there a thought, that you in turn, will mull over in you're mind.
Her is the thought; I know from my own test that prop wash leaves the tip of any prop blade at an angle, maybe some where around 45 degs. That means that the influence that the blade has on the Environment, is many times larger than the blade it's self. This means built up models, foamies, and also Helices, are impacted by the PW design. If the PW model is to small, the wrong shape, and or the wrong velocity, all the things that are done to fix flight problems, will fall short. Foamies fly on prop wash. A poorly performing PW will, you see how this goes. All 3D planes use PW to give lift, less lift, here we go again. Helices, if you don't have that Prop Wash leaving the tip of the blades, giving a large wash, it will change everything.
I love the new environment in 6. Please, all I ask, is keep an open mind, consider the size, shape, and velocity of the Prop Wash, and how it will impact the overall performance of 3D flight in 6. Have a nice day!

flexible
12-07-2011, 09:33 PM
This is the way it is now, it's broke and I can't fix it.

td9cowboy
12-08-2011, 06:39 AM
Just for the record, I'm finding I have to set the back torque to around 70 and the prop wash around 115 on the 3D swap content I've worked on. The increased prop wash on the tail feathers is also helpful for flat spins. All have been large or giant scale aircraft. Flex, I think Jim is working on it. He has to figure out exactly what the problem is before he can fix it. It could be a combination of problems. If they do tweak it, they have to do some testing to make sure they didn't adversely affect another flight mode. There is no EASY button.

opjose
12-21-2011, 11:50 PM
The problem is you should NOT need to make that adjustment AT ALL.

I am due to make a presentation at our February club meeting, so I elected to show people how to model a plane's physics in RF6.

To that end I selected this little foamy which I recently purchased...

Click me (http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=19962)

I got the specs right and adjusted the everything in the sim as per the plane.

I got surprisingly accurate behaviour ( more so than before BTW ) EXCEPT for ONE THING... prop/wash-torque roll.

If you check out the video in the above link you'll see that the plane can easily be launched into a harrier/torque roll by hand and the pilot has absolutely NO problem holding the plane in a prop hang as he flies it around completely stalled.

The sim behaves well, except the plane cannot hold against wash/torque the same way without my performing adjustments similiar to what you said.

Ergo the upgrade improved wing lift stuff greatly but damaged the modeling of prop/wash-torque roll.

In effect I side with Flexible on this...


However Flexible, please be aware that this is a TABLE DRIVE simulation.
No fluid dynamics type of modeling is done for the physics per-se, what you see on the smoke is mostly eye candy and should not be taken as indicative of what is actually going on... these are just numbers after all, not an "ether" through which the plane is actually flying...

Sometimes the emulation is so good we forget that some things simply have no meaning as a way of looking for deviations from behavioural reality off other things.

The sim may be modeling prop wash exactly as Jim indicates, but the smoke particles could move at an angle 90 degrees to that, and the physics emulation may still be correct....

flexible
12-22-2011, 09:57 AM
I have made this test in real life. The red shows the Prop wash of my 104" wingspan Texas Extreme 260. There is some amount of Prop wash moving over the wing tips.
This gives you full control, even in a hover, or in Harrier roles.
Go back and read Jims remarks about there Prop Wash cone. It becomes elongated, and does not expand out over the entire wing. I don't care about the inner and outer rings of RF Prop Wash, that is not the issue. The Prop Wash, MUST cover the Wing Tips, or it is not real world, and compromises performance. I proved it in RF using the plane that I cut the ailerons off just be yon the prop Wash line, it made no deference.
If I still had a big plane, I would post the test. Some one do the very simple test, using a big plane. Use a 35% or larger aerobatic plane. Use masking tape to stick 6" pieces of crate paper along the trailing edge of the wing. First piece just outside the fuse. Put them a few inches apart, all the way to the wing tip. Run a tie down around the enfonage, and anchor it to the ground. Run the power to full. Take a pick and post it. The prop wash is full next to the fuse. There is still some prop wash at the tips.
Prop Wash has everything to do with extreme flight, even the helies are deficient.

opjose
12-22-2011, 06:21 PM
Prop Wash has everything to do with extreme flight, even the helies are deficient.

There IS NO SPOON! :D

That the simulated plane does not behave as we would expect is a given, particularly since you've pointed and isolated things so well.

Jim mentioned last year that the simulation subdivides the wing area/effect calculations into four or so segments. It may well be that the tips fall outside of the segments that are included in the math.