Jump to content

Comparable Flight Dynamics


Wetted Area
 Share

Recommended Posts

Hello, I'm new to X-Pilot and to X-Plane.  I ran a couple searches for the terms "Flight Dynamics" and only found one thread dealing with a different issue on the Duchess aircraft, so here goes.

 

I am currently evaluating both FSX and XP.  In FSX, you can attempt to optimize some of the aircraft's flight control effectiveness (among other things) by going through both the .cfg and .air files. 

 

Question:

 

- What are the equivalent file types for doing the same flight dynamics optimization in X-Plane 10 Global, and what's the file directory path to those files?

 

I would like to compare the aircraft optimization capabilities available in both flight simulator platforms.  I am using the default C-172S in both platforms as an initial evaluation starting point for determining which flight simulator offers the most "realistic" flight characteristics.  I do not current have any physical flight controls installed at this time, but I do plan to use a yoke, rudder pedals and a TPM, going forward.  Right now, I am doing all tests with a Logitech mouse, using its track-ball as the "yoke."

 

My first out-the-box experience with X-Plane was that I was very disappointed with the overall look, given what I had just seen in the company's promotional video.  However, when comparing the out-the-box flight experience of the default Cessna 172SP in FSX Deluxe, with the default Cessna 172 in X-Plane Global, I noticed a significant difference in the flight characteristics. 

 

The XP default aircraft, using the mouse track-ball as the "yoke," flew far more true-to-form than the default Cessna 172SP in FSX Deluxe.  The FSX default aircraft has roll characteristics that are unlike the real C-172SP, while the X-Plane Global default aircraft was more true-to-form.  A Yoke, Rudder Pedals and a TPM are coming but not yet installed.  Still, I was very much able to control the X-Plane default with a mouse track-ball in a way that reminded me of the actual C-172 in some respects (not perfect).

 

Much appreciated.

Link to comment
Share on other sites

- What are the equivalent file types for doing the same flight dynamics optimization in X-Plane 10 Global, and what's the file directory path to those files?

This would be the .acf files. In X-Plane, the geometry in most cases defines the actual flight model, unlike MSFS which uses tables. That said, you would use the Plane Maker program located in your X-Plane directory to open an aircraft and edit the parameters for it.

 

 

My first out-the-box experience with X-Plane was that I was very disappointed with the overall look, given what I had just seen in the company's promotional video.  However, when comparing the out-the-box flight experience of the default Cessna 172SP in FSX Deluxe, with the default Cessna 172 in X-Plane Global, I noticed a significant difference in the flight characteristics.

I believe the promo video used some add-on aircraft, but don't recall on scenery. It certainly can look as is seen so long as your specs are bumped up! It's also important that you keep updated with X-Plane updates. It's come a very long way since version 10.0.0. 

Link to comment
Share on other sites

Thanks, Cameron.  I appreciate the help.

 

Ok, I've cracked PlaneMaker 10.10 open and took a look at all of its file menu options.  I've found the Control Geometry page and noticed something that I did not quite understand:

 

2ly39cp.jpg

 

If you don't mind, why are there two (2) configurations for the Rudder?  The only movable device that I know of on a C-172, attached to the vertical stabilizer, is the Rudder.  In 1977, Cessna first introduced the Rudder Trim, as a mechanically actuated device.  The X-Plane 10 Global default is based on the 1998 TCDS (Type Certification Data Sheet), which is the "S" variant, given the product name SP.  That 1998 Skyhawk also has a check-list that includes a check of the "trim" - though it does not specify Rudder Trim Tab.

 

Austin, did not clearly label the Control Geometry columns, nor did he clearly label the values for control surface deflection.  However, I was able to figure out what he meant by lables such as: "control surface down then up" or "control surface left then right."  You just have to know that he's talking about Control Surface Deflection Angle.  He does a pretty good job of giving a bubble explanation for Control Surface Chord Ratios, so I had no problems figuring that out.

 

When you take a look at the FAA TCDS number 3A12 (Cessna 172S Page 30), it specifies the maximum type certification Rudder Deflection as: Right 17-degrees and Left 17-degrees (+/- 1-dgree).  Resultant Yaw control being symmetric, that makes sense.  When looking back at the X-Plane Control Geometry page at settings for "Rudder 1", you can see that default aircraft model is using the correct specs for the Rudder Deflection.

 

However, this does not explain the settings for Rudder 2.  And, since the default aircraft is not the Cessna 336 or 337, having two instances of Rudder, that are actually being used is a bit confusing for me.  There are two instances for Aileron as well, but only one instance is actually being used - appropriately.

 

Any ideas on what the Left 30-degrees and Right 30-degrees Rudder instance number 2, is all about?

 

This is my first time looking at PlaneMaker and I think I like it!

Edited by Wetted Area
Link to comment
Share on other sites

Most likely these values are not used.   X-Plane is a bit of a "what if" software and it is perfectly feasible to have two rudders each with different chords and movements ala the Rutan boomerang.  Austin generally likes to give folks options and usualy two rudders are enough to experiment with.     The way x-plane works is it looks to see if there is a vertical stabilizer with a length greater than zero... IF there is, then it will look to see if there is a rudder (the chord ratio) and if so, how much the rudder moves.   You can have zero for the vertical stab length yet x-plane will retain values in the rudder deflection fields if you populate them....so could be a author accidentally put the values in the wrong field...(done that myself) and just went to rudder 1 to input it OR it could be that the author is trying to fine tune the yaw behavior in some fashion with a 2nd rudder.   Contrary to many folks believing that such a technique is cheating....the truth is an x-plane aircraft is a numerical / mathematical model and if a 2nd rudder used in some capacity gets the final results closer to the real thing...then that is a viable option.    3x4=12  just like 6x2 does...not the only way, but you never know what people are going to try.

 

Tom K

Laminar / IXEG

Edited by tkyler
Link to comment
Share on other sites

Most likely these values are not used.   X-Plane is a bit of a "what if" software and it is perfectly feasible to have two rudders each with different chords and movements ala the Rutan boomerang.  Austin generally likes to give folks options and usualy two rudders are enough to experiment with.     The way x-plane works is it looks to see if there is a vertical stabilizer with a length greater than zero... IF there is, then it will look to see if there is a rudder (the chord ratio) and if so, how much the rudder moves.   You can have zero for the vertical stab length yet x-plane will retain values in the rudder deflection fields if you populate them....so could be a author accidentally put the values in the wrong field...(done that myself) and just went to rudder 1 to input it OR it could be that the author is trying to fine tune the yaw behavior in some fashion with a 2nd rudder.   Contrary to many folks believing that such a technique is cheating....the truth is an x-plane aircraft is a numerical / mathematical model and if a 2nd rudder used in some capacity gets the final results closer to the real thing...then that is a viable option.    3x4=12  just like 6x2 does...not the only way, but you never know what people are going to try.

 

Tom K

Laminar / IXEG

Hey TK,

 

Thanks for helping out!

 

You kind of lost me on the "it is perfectly feasible to have two rudders" idea.  You said that most likely those definitions are not being used - do we know for certain?  I'm also wondering if I can simply delete or zero out Rudder 2, just like Aileron 2 is zero'd out.

 

You also mention that the author might have made something of a small data entry error.  Ok.  But, this is not a Third-Party Cessna 172SP.  I'm actually trying to use the X-Plane 10 Global Default C172SP.  So (correct me if I'm wrong), that would have come from Laminar, no?

 

Lastly, the idea of using a "2nd Rudder" to get "closer to the real thing" is a viable option.  Can you explain how that works, please.  I know you said that X-Plane is a mathematical model, but I don't understand having multiple flight controls with dissimilar moments and different aerodynamic forces (one having a larger deflection angle than the other - almost twice as much).  That part confuses me just a bit.  Like I said, I could certainly see this being plausible if it were a Skymaster, or something with a dual vertical stabilizer by design where the designer made a data entry error with either chord ratio or deflection angle.  Then, based on the deflection numbers in the pic I posted, it would simply be a matter of tuning those numbers to force symmetry in the control surfaces - if that was desired.

 

But, with only one (1) mathematical Rudder in the flight dynamics equation, all of the moment, aerodynamic force and control surface effectiveness is derived from a single flight control.  If there are two separate sets of values for the same input into the stability control algorithm, then I'm not quite sure how X-Plane differentiates in real-time which one to use.  If you can help with that issue, I'd appreciate it.

 

I'll also take a look at some other aircraft flight control values, to see if they are in the same ballpark as their respective FAA TCDS.

 

Thanks again!

Link to comment
Share on other sites

I believe the promo video used some add-on aircraft, but don't recall on scenery.  It certainly can look as is seen so long as your specs are bumped up! It's also important that you keep updated with X-Plane updates. It's come a very long way since version 10.0.0. 

 

Hey Cameron,

 

I running a decent rig - not the best but it should be good for both simulator platforms.  I get good graphics with FSX for example and only a slight amount of stutter that I have yet optimized out, but I'll get around to it shortly (I'm working with X-Plane right now, so I don't have the time to optimize FSX for frame rates).  I have two problem with X-Plane 10 Global:

 

1) I get precious little ground based topological accuracy.

 

The bridges don't look like they do in the X-Plane 10 video.  The streets, roads and highways don't have nearly as much detail as the video.  There are no recognizable cities anywhere!  I depart from Oakland International on runway 27L and you can't even distinguish San Francisco - there is no downtown, no financial district, no Fisherman's Warf that you can readily identify and the Golden Gate Bridge is almost non-descript.  The entire downtown area of the city of Oakland, is nowhere to be found and major landmarks within the city are just not there, or are extremely hard to find from any altitude.

 

2) Airport topological accuracy is low.

 

The airport lighting is either extremely low, or basically non-existent at night, or dusk/dawn.  I tried to do a short VFR flight around the bay area today in a military jet.  But, because the jet had so much power and speed, I ended up departing Oakland International and by the time I leveled off, I was lost - literally.  So, I picked a mountain range and followed the ridge back to the water. 

 

Once I got over the water, I headed east because I knew that would take me back toward the east bay area.  I was basically flying directly over my own back yard and never even knew it, because the land topology was so inaccurate and/or so incomplete.  I dropped the jet to 150kts at 1,000agl and began a circular search for identifiable landmasses, major highways that I drive on every day and major cities and down that I know exist.  I found none that I could recognize.  So, I decided to fly the bay as far south as I could, knowing that would put me south of runway 29.  I then made a 180-degree heading change and back tracked until I came to the Oakland International airport.  My airspeeds were between 150 - 200 kts during this dead-reckoning airport search.

 

So, I have another question, if you don't mind:

 

- What exactly was supposed to ship with X-Plane 10 Global, in terms of Landclasses, Scenery, Airports and Topological refinement?  Because, when I watch Austin's video from Feldberg, he is very clear that everything must be three dimensional and built-up from the smallest piece.  He talks about being able to see "headlights" and "tail lights" of vehicles, the generation lf three dimensional clouds with thousands of variations and all the little details down below in the cities from altitude.  Yet, I'm not getting a lot of that detail down below on the ground.

 

The more I use X-Plane 10, the more I like it better than FSX.  But, the more I don't get the same quantity of autogen that Austin talks about in his presentation and in the official video (which causes you to believe that is what you are buying), the more frustrated I become - because I do feel myself starting to lean heavily towards X-Plane, as my sim of choice.  The default aircraft that I have flown thus far in X-Plane 10, simply blow the FSX default aircraft out of the water in terms of simulating actual flight characteristics on my computer and I'm not even using a physical yoke yet!  So, I do like X-Plane 10, more than FSX now.  However, I would like to have the exact same out of the box experience that Austin talks about having and presents to people in public!

 

Am I missing a setting, a configuration, a check-box, a toggle switch, something, anything in the X-Plane shell, that would give me that same out of the box experience that Austin is showing here:

 

 

I love the way X-Plane 10 flies.  I would like to appreciate the way it looks, too.  Thanks!  :)

Link to comment
Share on other sites

You kind of lost me on the "it is perfectly feasible to have two rudders" idea.  You said that most likely those definitions are not being used - do we know for certain?  I'm also wondering if I can simply delete or zero out Rudder 2, just like Aileron 2 is zero'd out.

Ah, the default 172.  Yes, the author is using one rudder for the top half of the vert stab and the 2nd rudder for the lower half.  he does this because he wants to encompass the wetted area that a real 172 has for the "fin" between the vert stab that runs up along the spine of the fuselage.  This would allow for x-plane to simulate the crosswind effect on this area.  This is not the way I would do it myself.  As you can see from the picture,  x-plane will not only calculate side forces on the fuselage, but also on the "lower rudder" as well, resulting in a lot of side force on all that surface area.  X-Plane has no concept of "wind shadow"...that 2nd rudder is contributing to the behavior in yaw.  Being there is no real surface "inside the fuselage", this  could yield some inaccruate results in some instances.  It could also be "approximated by inceasing the vertical depth of the fuse a bit to get some of that side area from the "dorsal fin" and leaving out rudder 2 entirely. 

 

post-5-0-49351600-1359423703_thumb.jpg

 

 

 

You said that most likely those definitions are not being used - do we know for certain?

The way you tell  is to go to the "wings" page and click on the vertical stabilizer tab.  There are two rudders and if either one of them has a semi-length value greater than zero...then the values for the control deflections and chords are being used.  In the case of the 172, they are definitely getting used.

 

 

You also mention that the author might have made something of a small data entry error.  Ok.  But, this is not a Third-Party Cessna 172SP.  I'm actually trying to use the X-Plane 10 Global Default C172SP.  So (correct me if I'm wrong), that would have come from Laminar, no?

It would, but several of these aircraft are a bit older and contracted out to individuals...so its possible to encounter anything.  We have done a measure of quality control though...but plan more.  We are as prone to making errors as anybody else.

 

 

Lastly, the idea of using a "2nd Rudder" to get "closer to the real thing" is a viable option.  Can you explain how that works, please.  I know you said that X-Plane is a mathematical model, but I don't understand having multiple flight controls with dissimilar moments and different aerodynamic forces (one having a larger deflection angle than the other - almost twice as much).  That part confuses me just a bit.  Like I said, I could certainly see this being plausible if it were a Skymaster, or something with a dual vertical stabilizer by design where the designer made a data entry error with either chord ratio or deflection angle.  Then, based on the deflection numbers in the pic I posted, it would simply be a matter of tuning those numbers to force symmetry in the control surfaces - if that was desired.

 

But, with only one (1) mathematical Rudder in the flight dynamics equation, all of the moment, aerodynamic force and control surface effectiveness is derived from a single flight control.  If there are two separate sets of values for the same input into the stability control algorithm, then I'm not quite sure how X-Plane differentiates in real-time which one to use.  If you can help with that issue, I'd appreciate it.

 

This is definitely a heavier subject.   I'm a mechanical engineer by education, with focus on finite elements and numerical simulation...so I see things a bit differently than a lot of users.   If you have 1 rudder with surface area X and deflection of angle X.  THEN you place a 2nd rudder with surface area Y and deflection value X into the model.  That's the same as a single rudder with surface area X + Y to x-plane.    When it comes to forces on  a body, you can sum up all forces and moments and replace them with a single point on a body applied at the center of mass.  If you have two rudders, then x-plane simply calculates  the moments from both rudders , superimposes the forces and the net effect on the body is a single moment.   If I have two guys stand on your chest, each 100 lbs...thats no different than 1 200 lb guy standing on your chest...you don't need to figure out which 100lb guy to use, they both affect you.  x-plane doesn't figure out which to use, it simply calculates all forces from all force generating bodies and computes a net force on the aircraft.   It would be no different if you stuck your hand out the window.....you introduce a new surface and a new force...x-plane would figure the force, and add it to all the others.

 

Tom K

Laminar / IXEG

Edited by tkyler
Link to comment
Share on other sites

Just want to say WELCOME to the forum, I read a thread of yours on the Org and look forward to watching you pick X-Plane apart with your critical eye. ;)

Sent from my iPhone.

 

I'm actually flying X-Plane more than FSX at this point because it gives me a better overall experience - especially in the flight characteristics compartment.  So, I'm much more likely to pick apart the loss of frame rate I get with FSX, which I do not get to anywhere near the same degree with X-Plane 10 Global running on my box.

 

FSX, is killing it with Scenery and Add-Ons. There is a ton of stuff out there to try.  But, I'm looking hard for X-Plane 10 Scenery and Airport models that can give me the coverage I need in the United States.  There are tons of U.K. scenery and airport models, but I have not seen that same level of effort on the U.S. side of the scenery/airport equation.

 

I took two (2) X-Plane 10 Global flights today and zero (0) FSX flights.  Just in case you were keeping score!

 

I really want to settle down with this thing, but I need more surface topology coverage, more airport coverage and more aircraft coverage - namely the Phenom 300 with the Prodigy Flight Deck 300 (G1000).  If could get some good topology/scenery, enough airport models, a good C172 or PA28 with G1000, a King Air C90 or Super 2000 or Beechcraft Baron or Beechcraft Royal Duke and a Phenom 300 with Prodigy (G1000), then I'd be set with X-Plane.

 

I'm trying to get all of those components lined-up and tuned.  Then I can stop researching and start flying and familiarizing myself with these aircraft appropriately on cross-country flights with either PilotEdge, Vatsim or IVAO.  Since this will be my real world training track at some point later this year, I want to get a mental head start on these aircraft now.

 

This is not a game for me.  I plan to use the sim to front-run my actual training, so that I am not shock-cooled by a fast flight training program/regime.

 

How do I put that package together?  I'm leaning heavily towards XP at this time.

Edited by Wetted Area
Link to comment
Share on other sites

1) I get precious little ground based topological accuracy.

X-Plane is relatively good in its resolution of mesh, but not quite the good you and I would like. Long story short, the way mesh is created in X-Plane it would be cost and space prohibitive to ship hi-res mesh. It would also likely not perform very well.

 

 

The bridges don't look like they do in the X-Plane 10 video.  The streets, roads and highways don't have nearly as much detail as the video.  There are no recognizable cities anywhere!  I depart from Oakland International on runway 27L and you can't even distinguish San Francisco - there is no downtown, no financial district, no Fisherman's Warf that you can readily identify and the Golden Gate Bridge is almost non-descript.  The entire downtown area of the city of Oakland, is nowhere to be found and major landmarks within the city are just not there, or are extremely hard to find from any altitude.

So I have had a look at the video you speak of. Everything in there is without a doubt default and what ships with X-Plane. Most of the video is taken around the KSEA region. X-Plane scenery is only as good as the data it is built with. The streets, mesh, etc come from free sources. The mesh has an algorithm which determines the best plausible texture set to apply to a region given the topology. Sometimes this is accurate, sometimes it's not. For Austin, plausible means just that, but not LITERAL...not exact. If I told you about a city you had never been to and you try to imagine it in your head, that's essentially what Austin would mean. You know what cars look like, you know what buildings look like. X-Plane does NOT accurately place autogen buildings (hence the reason you see no landmark buildings or wharf), but it does accurately place roads, power lines, railroads, etc. That said, this data comes from OpenStreetMap, and if not all roads are in there then so too will the roads not be there in X-Plane. Laminar is also aware of some problem areas in the default scenery and, as far as I am aware, does have plans to re-cut some scenery to look better.

 

 

 

2) Airport topological accuracy is low.

 

The airport lighting is either extremely low, or basically non-existent at night, or dusk/dawn.

Perhaps I don't understand you on this one. In my experience, Laminar has done quite well on the airport lighting. It's important to remember that unless you are at just the right angle (even in real life) you will not see airport lights. Usually it is just a black patch in a city. That said, to really get the best experience that one sees in the videos from Laminar, you need to enable HDR mode at night.

 

 

   

- What exactly was supposed to ship with X-Plane 10 Global, in terms of Landclasses, Scenery, Airports and Topological refinement?  Because, when I watch Austin's video from Feldberg, he is very clear that everything must be three dimensional and built-up from the smallest piece.  He talks about being able to see "headlights" and "tail lights" of vehicles, the generation lf three dimensional clouds with thousands of variations and all the little details down below in the cities from altitude.  Yet, I'm not getting a lot of that detail down below on the ground.

Everything I have mentioned above. From Austin's video I believe he has pretty much delivered on what he stated would be there. Cars, tail lights, night lights, etc...it's all there. HDR is key in this though! These objects use "real lights" and are really only rendered in HDR mode. Without it on...say hello to darkness.

Link to comment
Share on other sites

X-Plane is relatively good in its resolution of mesh, but not quite the good you and I would like. Long story short, the way mesh is created in X-Plane it would be cost and space prohibitive to ship hi-res mesh. It would also likely not perform very well.

If you - well I don't mean Camreon, but "Wetted Area" - are interested in higher res version of the Global Scenery, I have cut a few areas of the planet in "HD" quality (approximately doubling the mesh resolution and also inducing better representation of the raw landclass data), like Alaska, USA West, some parts of Europe. This way you can check out what the extra res can (or can't) give you (its free):

 

http://www.alpilotx.net/downloads/x-plane-10-hd-scenery-mesh/

 

Or there is an even more "special" scenery (with non OSM based data, and also very high res mesh) for New Zealand (its free too):

 

http://www.alpilotx.net/downloads/x-plane-10-new-zealand-pro/

 

Both of them are meant to showcase, what X-Plane can do (but at the cost of higher space usage, more strain on graphics hardware ... and in the case of NZ Pro, the need for outstanding geodata and a lot of extra processing needed).

 

What exactly was supposed to ship with X-Plane 10 Global, in terms of Landclasses, Scenery, Airports and Topological refinement?

Well, I do point to this article every now or then, but I think it wont hurt to do it again (as not everyone can know about this old piece :) ). In it, I gave a more or less detailed description about how the base mesh etc. (of the Global Scenery) works, how it is done, what data it uses. Its a lengthy read, but has quite some info in it:

 

http://xsimreviews.com/2011/12/10/developer-interview-andras-fabian-mr-x-terrain/

Link to comment
Share on other sites

Ah, the default 172.  Yes, the author is using one rudder for the top half of the vert stab and the 2nd rudder for the lower half.  he does this because he wants to encompass the wetted area that a real 172 has for the "fin" between the vert stab that runs up along the spine of the fuselage.  This would allow for x-plane to simulate the crosswind effect on this area.

 

168iz3b.jpg

 

That explains the range of values between both rudder chord ratios having union at 0.38 and extremes at 0.16 and 0.40 respectively in the Plane-Maker Control Geometry page.  What I don't understand is the more than 100% increase in the rudder 2 effectiveness?  Doing that, essentially warps the the deflection of the composite rudder (1 and 2) such that the lower section (rudder 2) deflects more than double the amount of the upper section (rudder 1).

 

As far as mathematically tweaking the increase in wetted area by causing the dorsal portion of the vertical stabilizer that runs on top of the aft section of the fuselage to be "involved" in the overall net effect of the elevator deflection is concerned, I've never heard or seen such a thing before.  Quite frankly, I don't know how that works out mathematically, or if it is really just a "cheat" for the designer to math-wise optimize that which would be physically impossible to do in a real Cessna 172SP with the same empennage nomenclature.

 

It is like throwing a formula or algorithm at the "wind" and getting an actual response to it in terms of aircraft behavior.  I'd have to think about that one for a while before I actually understood it beyond what I just wrote.

 

But, sure enough - you are correct.  You can see the actual cut-off point in both rudder chord ratios with one ending at .38 and the other starting at the same exact number.  LOL!  Like the designer welded them together that way.

 

Good eye.

 

This is not the way I would do it myself.  As you can see from the picture,  x-plane will not only calculate side forces on the fuselage, but also on the "lower rudder" as well, resulting in a lot of side force on all that surface area.

 

That's what I'm talking about.  And, when you include the doubling of the rudder effectiveness, that would only serve to augment the already side force, causing an even greater yawing moment - I would think.  The question is - does that increase in yaw moment around the Z-axis exceed the design intend for yaw of the OEM?  That's the kind of stuff that need to be paid attention to when making these types of modification.

 

Who was the designer - do you know?  I'd love to talk with them.

 

 

X-Plane has no concept of "wind shadow"...that 2nd rudder is contributing to the behavior in yaw.  Being there is no real surface "inside the fuselage", this  could yield some inaccruate results in some instances.  It could also be "approximated by inceasing the vertical depth of the fuse a bit to get some of that side area from the "dorsal fin" and leaving out rudder 2 entirely.

 

Which would basically turn the empanage section into a remake of the Glasair III tail design - quite nearly.  Take a look:

 

348146c.jpg

 

I really like the Glasair III.

 

 

The way you tell  is to go to the "wings" page and click on the vertical stabilizer tab.  There are two rudders and if either one of them has a semi-length value greater than zero...then the values for the control deflections and chords are being used.  In the case of the 172, they are definitely getting used.

 

Vert Stab 1:

10ihr41.jpg

 

Vert Stab 2:

mu686.jpg

 

(VS1+VS2) = 11.65ft total length.  But, once again, I don't understand why VS1 has an Incidence of -0.3 for all of its six (6) components, while VS2 has an Incidence of 0.0 for its four (4) components:

 

Vert Stab 1:

34e6pvs.jpg

 

Vert Stab 2:

2cfbmtd.jpg

 

This being the vertical stabilizer of the C172SP, the resultant combined net force(s) acting upon its surface in straight and level flight should be symmetrical/neutral.  However, to counter single engine P-factor, this -0.3 angle of incidence could be the Vertical Stabilizer "Twist."  But, then that begs the question: Why twist only the Vert Stab 1 section?  I don't know the answer to that.  Maybe the designer found that it was just enough to achieve a proper counter to P-factor.

 

 

 

If you have 1 rudder with surface area X and deflection of angle X.  THEN you place a 2nd rudder with surface area Y and deflection value X into the model.

 

Which is precisely what the designer did not do in this case.  They increased the Rudder 2 by 100% over Rudder 1's effectiveness.  Not the size, but the mathematical (theoretical) net effect of the rudder deflection.  How that affects combined rudder effectiveness is beyond me, until I either get more information - or better yet, until I can talk with the designer.  But, the chances of that happening are slim to none and slim just left town.

 

Hey, thanks for the input!  It was nice talking with a fellow engineer.  ;)

Link to comment
Share on other sites

X-Plane is relatively good in its resolution of mesh, but not quite the good you and I would like. Long story short, the way mesh is created in X-Plane it would be cost and space prohibitive to ship hi-res mesh. It would also likely not perform very well.

 

Hey, Cameron.

 

This brings up two questions:

 

1) How does FSX get it done where X-Plane 10 Global cannot out of the box?

 

Essentially, what I'm seeing in FSX are all the football stadiums, basketball arenas, lakes, bridges, cities, towns, roads, highways, streets and major landmarks, straight out of the box.  It is not perfect topological replication, but to my surprise, FSX had a couple of sports complexes located pretty much where they should be as well as to the correct scale, based on other ground based object in the same vicinity.  So, FSX and X-Plane, should have the same cost per data element factors to overcome.  Yet, when I look at X-Plane, these things are clearly missing. (this might also explain the difference in the size of the total install of each application as well).

 

2) What happens in real-time on screen when you have a scenery package installed that covers X square miles and then you fly beyond the boundary of that coverage area?

 

Does FSX simply start to render its own native scenery instantly upon crossing the coverage boarder.  As you are approaching the coverage boarder, would you be able to look down and see over into the native FSX scenery, while still flying through the "package" scenery and clearly notice the difference between the two.

 

 

...but it does accurately place roads, power lines, railroads, etc. That said, this data comes from OpenStreetMap, and if not all roads are in there then so too will the roads not be there in X-Plane. Laminar is also aware of some problem areas in the default scenery and, as far as I am aware, does have plans to re-cut some scenery to look better.

 

My X-Plane X Global places massively tall power lines on top of the east bay hills directly east of runway 27L\27R.  Those hills do exist and they are very prominent in the area, but there are no power lines anywhere on the ridge adjacent to KOAK.  Here's a pic of OpenStreetMap's depiction and coverage of the immediately Oakland Bay Area:

 

2mzlgr9.jpg

 

I've circled just three (3) items:  KOAK airport.  Oracle Arena (Oakland Coliseum).  Downtown Oakland.  Out of these three landmarks, only KOAK shows up in X-Plane 10 Global out of the box.  All three show up prominently in FSX out of the box.  We can see that this is a very prominent part of the State of California and OpenStreetMap clearly shows full land development.  But, what X-Plane 10 Global shows in between KOAK in the south and Downtown Oakland in the North, is a vast area of mostly what looks like agricultural land, with just a handful of buildings in between.  As you get closer to the Downtown area, you do start to see more AutoGen buildings in what should be the downtown area, but the density is is far from even 10% of the actual area - the rest is green/brown looking agricultural type landclass in X-Plane 10 Global.

 

I'm not wearing this out on purpose.  It is just that Austin, did indicate that if the data exists in OpenStreetMap, then accurate AutoGen of that area will be produced i X-Plane 10 Global.  That includes streets as well.  Streets can be seen all over OpenStreetMap in this pic, but literally less than 5% of the street grid shows up in X-Plane 10 Global covering this same exact area.  In X-Plane 10 Global, this same densely populated area you see in OpenStreetMap, looks like some of our areas old historic photographs of the Bay Area, back in the late 19th century.

 

 

 That said, to really get the best experience that one sees in the videos from Laminar, you need to enable HDR mode at night. Cars, tail lights, night lights, etc...it's all there. HDR is key in this though! These objects use "real lights" and are really only rendered in HDR mode. Without it on...say hello to darkness.

 

Interesting.  I did not know that you had to enable a function to be able to see certain autogen elements.  I'll look up HDR in the manual and figure out what it means, how it works and how to enable it.

 

I'm pretty much over the issue of X-Plane 10 Global Scenery and the official video showing a nicely rendered Seattle, Washington, area.  I'll just focus on what I can change at this point, knowing that things are the way they are.

 

Thanks for the tip and I appreciate the help on my question above, if you don't mind.  :)

Link to comment
Share on other sites

(VS1+VS2) = 11.65ft total length.  But, once again, I don't understand why VS1 has an Incidence of -0.3 for all of its six (6) components, while VS2 has an Incidence of 0.0 for its four (4) components:

 I'm not going to address this question literally, but rather figuratively from a "design theory" point of view.   Way back when I got in a discussion with a designer that absolutely insisted on using real numbers everywhere in the flight model.  I couldn't really argue this approach, but at the same time, it puts the complete onus on Austin for simulating reality in code.   Having done a bit of numerical simulation myself, I know that the effort involved to get realism in all possible dynamic inputs to aircraft performance is just not realistic.  

 

Now having had some classes in numerical methods, finite elements and mechanical thingys,  I don't see the x-plane model as a "build it like the real thing and it will work".  I see forces, moments, translations and rotations....and rates thereof.    If one were to input numbers exactly and then "seat of the pants" common sense or flight testing says x-plane just isn't getting it quite right, then it's necessary to depart from norms to do it.  A lot of time, we'll just "feel" our way there.  Say the rudder was ineffective with actual chord an deflection values...perhaps austin's propwash model or wheel friction model isn't quite right...whatever the reason, we need more rudder control.  Well you either deflect the rudder a bit more or increase it's area via chords or semi-length...or even incidence....only partial in this case as incidence on both might have been too much effect.  You never really know what course a designer may take as is there is always multiple options.   The thing i have found in my non-stop discussions over the flight model is that many time, people tend to discuss in "one regime".....usually cruise, whereas from a design perspective, I at least like to consider all phases.  Sometimes I just have to compromise...reminding myself that this is a 70.00 sim that ain't too bad.   Anyhow, that nice rudder incidence you put in to counter for P-factor at low speeds / takeoff will create unrealistic yaw at cruise speeds.   

 

The designer of the 172 is a gentlemen named "max".  He sells products under the "DMax" label.   Many of the default x-planes have been "built" for x-plane by "serious hobbyists"....generally pilots or 3D artists with an interest in aviation.   The quality of the flight models is totally unknown as I have no idea the depth of their knowledge applied to their models....so just because a flight model comes from the "Laminar camp" is no guarantee as to its accuracy.   I think as engineers, we will always have a more critical eye to these things since that's the way our brains work and why we got into the discipline in the first place....but still no guarantee.  I've been guilty of working on a model and saying...."I just don't feel like spending x hours getting this right for 1 in a 1000 customers to notice".

 

Tom K

Laminar / IXEG 

 

 

 

 

 

 

1) How does FSX get it done where X-Plane 10 Global cannot out of the box?

 

 

10 year head start....monopoly on the market and all the revenue thereof to support an army of artists to build all that stuff.   x-Plane, while making great money for Austin initially...and now enough to support a small team is still only perhaps 15-20% of the flight sim market.  Artists need to get paid and not enough folks are pouring into x-plane, funding such an effort.  They don't come because x-plane lacks these things so it's a catch-22 of which the only solution I think is time.   Being FSX is no longer developed and folks seem to like a lot of things about x-plane, they don't walk away, but neither do they flock.   What I foresee is developers coming over slowly...like dripping water in a bucket.  It grows slow, but eventually tips.   I see no way around this currently.   Could Austin set aside 100,000.00 and say, "this is for artists to make nothing but landmarks"...yea, that could happen but only Austin can ascertain if he would go there.  Being that x-plane is a business and supports his other pursuits well, his incentive is going to be less than ours who are without these features.    Austin maintains two other business and the mobile platform, which is a great revenue stream...better than the desktop possibly.  This leaves x-plane desktop...our choice as enthusiasts as the "last in line" sometimes.   We ARE however, discussing the issue and strategizing ways to deal with it, so it's not completely hopeless.

Edited by tkyler
  • Upvote 1
Link to comment
Share on other sites

1) How does FSX get it done where X-Plane 10 Global cannot out of the box?

 

Essentially, what I'm seeing in FSX are all the football stadiums, basketball arenas, lakes, bridges, cities, towns, roads, highways, streets and major landmarks, straight out of the box.  It is not perfect topological replication, but to my surprise, FSX had a couple of sports complexes located pretty much where they should be as well as to the correct scale, based on other ground based object in the same vicinity.  So, FSX and X-Plane, should have the same cost per data element factors to overcome.  Yet, when I look at X-Plane, these things are clearly missing. (this might also explain the difference in the size of the total install of each application as well).

Microsoft/ACES spent considerable time with a team to develop bigger cities with some accurately placed CUSTOM scenery, not autogen. This would be why you would see more accurate depiction of such elements. This comes down to time, size of team, and money, all of which Microsoft has (had) a ton more of. That said, X-Plane 10 is a complete re-write (or near it) of the scenery engine compared to version 9, and is leaps and bounds better. As time moves forward I think it would be safe to assume the scenery engine will make great forward improvements as well.

Until X-Plane 10, Laminar's focus has never really been eye candy. It's always been about flight dynamics, but Laminar is now starting to cater to both sides with the latest changes. For someone coming from X-Plane 9 it's a massive improvement eye candy wise, for someone coming from MSFS it is a different story.

 

 

2) What happens in real-time on screen when you have a scenery package installed that covers X square miles and then you fly beyond the boundary of that coverage area?

 

Does FSX simply start to render its own native scenery instantly upon crossing the coverage boarder.  As you are approaching the coverage boarder, would you be able to look down and see over into the native FSX scenery, while still flying through the "package" scenery and clearly notice the difference between the two.

I don't understand this question. You are asking about FS X? Perhaps that was a typo? Either way, both sims are not repetitive and there are technically no "boundaries". For X-Plane there are tiles that just fit together like square pieces in a puzzle. Each piece is intended and unique for its section of the world. You don't even know you reach the seam/boundary in X-Plane (or FS X for that matter).

 

 

  

I'm not wearing this out on purpose.  It is just that Austin, did indicate that if the data exists in OpenStreetMap, then accurate AutoGen of that area will be produced i X-Plane 10 Global.

I don't believe Austin made such a statement. The way the system works is it takes the power lines, railroad, roads, etc. data and fills in autogen in between it to the best it *thinks* it should plausibly do. Nothing is perfect here, but it's also not horrible. 

 

That includes streets as well.  Streets can be seen all over OpenStreetMap in this pic, but literally less than 5% of the street grid shows up in X-Plane 10 Global covering this same exact area.  In X-Plane 10 Global, this same densely populated area you see in OpenStreetMap, looks like some of our areas old historic photographs of the Bay Area, back in the late 19th century.

You would need to pump up your street settings to see them all. Be aware that doing so does and will have an impact on your frame rates. This is why the default setting does not set it to show ALL streets in the scenery even though they are there.

 

 

Interesting.  I did not know that you had to enable a function to be able to see certain autogen elements.  I'll look up HDR in the manual and figure out what it means, how it works and how to enable it.

There's not much to read up on here. Open X-Plane and the Rendering Settings menu and enable HDR!

 

 

I'm pretty much over the issue of X-Plane 10 Global Scenery and the official video showing a nicely rendered Seattle, Washington, area.  I'll just focus on what I can change at this point, knowing that things are the way they are.

I actually don't think you fully understand still. For the most part, all scenery locations are pretty well consistent so long as you turn up the rendering settings. It is not exclusive to Seattle.

  • Upvote 1
Link to comment
Share on other sites

Having done a bit of numerical simulation myself, I know that the effort involved to get realism in all possible dynamic inputs to aircraft performance is just not realistic.

 

What area of math sim have you done btw?

 

Caveat, indeed on the phrases: "the effort involved," and "all possible dynamic inputs."  Those two extractions pretty much says it all, when you consider the price tag of both FSX, P3D and X-Plane 10 combined.  Clearly, only a volume sales economic model would make that worthwhile.  And, I agree with you - doing it right down to the finite numbers is indeed a daunting task - especially when developing an aircraft model from scratch.  But, this is part of what I'm trying to get at.  The more I look at this, the more I realize that none of them have mostly likely worked any of the aircraft flight models from anywhere near scratch.

 

What they most probably did was use a 3D modeling application to put together a plausible visual representation of the real thing on screen, laid in some textures, cut an instrument panel using the same 3D modeler, dropped in some standard graphics file formats for the gauges, some kind of pixel or vector graphic elements for needles, some more 3D modeling for the flight control (yoke/stick) and a few in-cockpit controls (dials, toggle switches, buttons, etc.) and then animated the whole thing -  complete with dithered spinning circles made to look like propellers and turbine engine fan blades. 

 

The rest was probably a matter of increasing the shear number of aircraft parts that animate real motion, to give the aircraft a sense of complexity in design.  Ok, so that's the Aircraft base Model and you can literally make that look anyway you want.  It is not really interacting with the air in a 3-dimensional fluid environment while running inside your computer.  Any AC3D or SketchUp type modeler will take a two dimensional drawing and allow you to extend it into the third dimension.  Ok, I pretty much get that part.

 

What I'm most interested in figuring out is how they created the other model.  The Flight Dynamics Model.

 

When I look at the file system, the aircraft files themselves seem really small in size.  So, I'm wondering of the flight simulator itself is encoded to be a Flight Dynamics Engine of sorts, where it then uses its own native files (probably through a series of .dll file types), in conjunction with the aircraft.acf and all of the .afl (Air Foil) files, to produce the simulation result based on its mathematical interpretation.  This is the part I'm trying to understand with respect to the application architecture.

 

What is clear is the fact that for the default C-172SP, there are a specific set of NACA files - three (3) to be exact. When I look up those files (NACA  0006, 0009 and 2412), two of them are listed as symmetrical (which I confirmed with an outside source) and one is significantly larger and not symmetrical - it also has a significantly higher Max_CL at 1.204 and a Max_L/D of 50.702.  Apparently, 2412 is the primary wing and 0006 and 0009 must be the horizontal and vertical stabilizers.  So, clearly the NACA profile is being read or called by a .dll implicitly for use by what has to be some kind of flight dynamics engine somewhere in the base code of X-Plane.

 

I'm trying to understand how the performance is being modeled mathematically.  I've already taken a look at Plane-Maker and I've gone through its file menu and looked at each window.  Clearly, there is a lot that can be modified in both Plane-Maker and Airfoil maker, if you know what you are doing.  The question is why - why would we even bother going into this level of detail to change these airfoils and thus, the behavior of the flight dynamics.

 

I guess the answer is that we wouldn't - we would not bother changing these configurations unless the X-Plane itself is predicated upon a proprietary pseudo flight dynamics model, that is not geared to produce the mathematically predictive behavior of the aircraft in an open flow 3-dimensional air filled environment, but instead, one that is designed to simply appear to be in-line with the OEM performance requirements of the POH.  And, for that, you would not need a real "flight dynamics engine" at all.

 

So, for me - I guess I am in search of the truth behind how both FSX and X-Plane 10, go about conducting flight dynamic calculations.  The reason that is important to me, is because it will tell me which flight simulator most closely approximates actual aircraft performance, which means a lot more than just how well the aircraft handles.  It gets into very important issues like take-off, climb, cruise, descent (from an aerodynamics standpoint) and the landing performance of the aircraft.  It will also tell me which flight simulator environment will provide the best predictions on endurance and range at various differentials in density altitude, winds aloft, weight of the aircraft, etc.  When flying instrument approaches, aircraft performance becomes critically important in the descent profile, as the rate of descent being registered by the VSI is another predictive element that tells the pilot where the aircraft should be with respect to altitude, after descending for N-number of minutes while on glide-slope.

 

So, there are a number of reasons why I think these things are important and I will continue to evaluate each platform based on these kinds of assessments until I come to a conclusion as to which platform best approximates a real "simulated" environment.

 

On the issue of Austin's commitment to advancing X-Plane (I keep hearing a lot of people talk about how the platform is nowhere near maxing out its true potential), I think you are correct.  Only he can decide for himself what he wants to do.  However, I would think that in an economy of scale, one would look at the potential to grow a market as well as maintaining the one already in place.  When you consider what Lockheed Martin has done with Flight Simulator code by reworking the source and trying to bring it up to current standard of presentation and performance, I quite frankly can't see how Austin, fails to boost the efficiency and presentation of X-Plane, and make a bold step with X-Plane 11, that once and for all deals with the matter of graphical under achievement.

 

I appreciate your input - thanks!

Link to comment
Share on other sites

I don't understand this question. You are asking about FS X? Perhaps that was a typo?

 

The question was directed at both, really.  It applied to both platforms - though I obviously mistyped FSX - this is an X-Plane related forum.  Second, the question wasn't how the scenery was tied together. The question was what happens when you buy/install third-party scenery that covers X-square miles and you end up flying out beyond that boundary of the custom scenery.

 

After having spend more time in general with both simulators and downloading one airport scenery add-on for FSX yesterday and seeing how it works, I believe I can figure what happens in X-Plane under the above scenario.

 

 

I don't believe Austin made such a statement.

 

I didn't believe it either - until I actually heard him make the statement in a video.  I use the video as my source where Austin, himself, is speaking.  When quoting someone, I don't want to post hearsay.  Not admissible.

There's not much to read up on here. Open X-Plane and the Rendering Settings menu and enable HDR!

 

I actually don't think you fully understand still. For the most part, all scenery locations are pretty well consistent so long as you turn up the rendering settings. It is not exclusive to Seattle.

 

I'll check Rendering Settings and enable HDR.

 

Thanks!

Link to comment
Share on other sites

What area of math sim have you done btw?

Mostly "low level" numerical methods and dynamics modeling.  By dynamics modeling, I mean deriving governing equations of motion based on mass, motion, forces, free body diagrams to build a "model of the system" and then evaluating step response, natural frequencies, vibration, accelerations.....all that stuff.  I can't say I'm still proficient at it though having not practiced it for some time.  The "low level" numerical methods is just that.....developing computer algorithms for solving large scale linear algebra issues, numerical integration techniques, root-finding techniques etc, but again, haven't done it in 15 years.  It does still remain in my head though ready for use when needed as well as the challenges / limitations thereof.

 

 

I'm trying to understand how the performance is being modeled mathematically

That's a tough one in that you are really asking for details from years of programming that the programmers just aren't going to give out and the best we can do is infer from what we do have....which fortunately, is enough I think, at least for me it is.   There are several ways to model a phenomena of course.   One can use empirical methods as FSX does....using  point data, curve-fitting and statistical equations....but this requires obtained data from measurements of some type....OR, like X-Plane, you can use governing equations of state in a predictive manner.     The predictive manner, using equations, generally does not account for variances in it's base form but you do not need obtained data either and can thus model aircraft that don't yet exist.   The effort expended in either approach determines the accuracy of the outcome...both are simply different methods.  Being that 99 out of a 100 times, you want to simulate a plane that already exists though, then the empirical approach will give the most accurate results, but only if you provide a sufficent number of data points.   The greater the diversity of phenomena though, the more data you have to have and the data can get large and unweildy very quickly.   If you say, "hey, lets model aircraft performance from still air to hurricane force, from -40c to 100c, from -15 AoA to +15 AoA, from min weight to max weight, from no power to full power...you could probably end up with a few billion data points in no time.    With predictive methods, governing equations of state and a sufficient discretized model, you can calculate as many points as the computer can handle.  The faster the computer, the more accurate the results (in general).   You lose variance with this method though and tend to lose the "nuances" of modeling associated with the cumulation of minor inputs....such as cable stretch, skin flex, response time, wind gusts, etc. ...but the effects of these "peripheral phenomena" can be approximated in a limited capacity, an example of which is x-plane's "turbulence" and "prop vortex" vector fields.

 

So each has a limit.  X-Plane's approach will result in a smoother performance envelope....more accurate?  not necessarily, but smoothness is really important to me as far as feel goes so therefore I choose x-plane as my platform in which to try and improve the accuracy.  I think accuracy can be improved in both sims, but I think it's easier in x-plane and I think the final result is smoother in x-plane.  Just looking at something as simple as "turning up the cockpit lighting" shows the difference to me.

 

Tom K

Edited by tkyler
Link to comment
Share on other sites

I didn't believe it either - until I actually heard him make the statement in a video.  I use the video as my source where Austin, himself, is speaking.  When quoting someone, I don't want to post hearsay.  Not admissible.

 

I watched the video. He did not say what you stated he did. What he did say:

 

1. OpenStreetMap was utilized to produce the roads and they are not custom, but auto generated based on the roads that people have contributed into the OSM database (almost all). This is a true statement, and Austin left little for interpretation here. He even went to the extent to state that meant overpasses, etc...not buildings. Since the roads are accurately placed in the sim and you can see them all if you pump up your roads settings, I believe the entirety of Austin's claims to be true.

 

2. He stated that the X-Plane world is plausibly 3D. He further clarified this by saying that the clouds, the aircraft, the buildings, the cars, and whatever else were not superimposed or billboarded to a 2D image, be it flat or vertical. They are truly 3D objects. This statement is also true.

Link to comment
Share on other sites

Mostly "low level" numerical methods and dynamics modeling.  By dynamics modeling, I mean deriving governing equations of motion based on mass, motion, forces, free body diagrams to build a "model of the system" and then evaluating step response, natural frequencies, vibration, accelerations.....all that stuff.  I can't say I'm still proficient at it though having not practiced it for some time.  The "low level" numerical methods is just that.....developing computer algorithms for solving large scale linear algebra issues, numerical integration techniques, root-finding techniques etc, but again, haven't done it in 15 years.  It does still remain in my head though ready for use when needed as well as the challenges / limitations thereof.

 

Then it sounds like you really know some things about flat and spherical earth point mass modeling with respect to kinematic equations involving local horizontal/vertical vectors, velocity vectors in dynamic equations, force balance equations and the like.

 

Maybe you can help me out with something. One of the problems I used to have many years ago, was in working out some of the composite relationships in my dynamic equations.  For example, take: mV[dot] = T cos[alpha] - D - mg sin y.  In trying to resolve (as just one example) the force components to the velocity vector: mVy[dot] = T sin [alpha] + L, I have a hard time with figuring out how to position -mg cos y.  Bot the lift vector and the gravity component in that arrangement was always a problem for me.  I had no problems visualizing the vectors on a graph, but when it came to solving for the derivatives, I tended to have problems with the integration steps.  Probably not too appropriate for this forum, but I don't often chat with people online in the sim world about stuff like this.

 

Anyway, it is nice to another "low level" type on this forum.

 

That's a tough one in that you are really asking for details from years of programming that the programmers just aren't going to give out and the best we can do is infer from what we do have....which fortunately, is enough I think, at least for me it is.   There are several ways to model a phenomena of course.   One can use empirical methods as FSX does....using  point data, curve-fitting and statistical equations....but this requires obtained data from measurements of some type....OR, like X-Plane, you can use governing equations of state in a predictive manner.     The predictive manner, using equations, generally does not account for variances in it's base form but you do not need obtained data either and can thus model aircraft that don't yet exist.

 

That's the crux of the entire matter that I'm going after.  I'm glad I ran into you - you understand where I'm coming from.

 

If that's what X-Plane does, then it is more of a theoretical presentation of actual flight dynamics than FSX - but also more rigid.  Not rigid in terms of its inability to be more "fluid" in its predictive capability, but more rigid in the strict sense that (if what you say is true) it does not adhere to specific OEM data points.  Therefore, we can't trust that the resultant dynamics derived from X-Plane are "by the numbers" relative to any given OEM derived aircraft.  In other words, the structural replication of the aircraft (mode shapes and geometry) becomes much more critical, if their goal was to be as precise as possible in "simulating" flight dynamics.  Of course, the reason being the absolute dependence on getting the planform data as accurate as possible, because those inputs to the theoretical flight dynamics model (in the aggregate) are causal for actual aircraft performance.

 

It would seem to be the contrary with FSX.  Again, if what you say is true about them - they have apparently obtained "real data" from "real production aircraft" (or, as you say - "some source") in their default models and can therefore, rely upon the empirical, data driven formulations to recreate the range of actual aircraft performance, without using the more "organic" theoretical methods for obtaining their flight dynamics results.  It is interesting how both took two entirely different paths in their attempt to achieve seemingly the same things.  I would guess then, based on what you have said here, that if X-Plane sticks to actual measurements derived from OEM aircraft, then I would expect its results to be more true to form.  This might explain the "smoother," more non-linear transitions through the XYZ planes in the flight behavior that I'm experiencing with X-Plane default aircraft, as opposed to the more jittery flight behavior with a slightly detectable "stepwise" type motion that I seem to get from the default FSX aircraft.  It is not obtusely obvious, but I do detect traces of it in the initial reaction to moments generated by larger aileron deflections, as well as almost any attempt to apply opposing aileron to halt rotational momentum around the X-axis in the FSX default aircraft that I've looked at thus far.

 

 

If you say, "hey, lets model aircraft performance from still air to hurricane force, from -40c to 100c, from -15 AoA to +15 AoA, from min weight to max weight, from no power to full power...you could probably end up with a few billion data points in no time.

 

Or, variable CG changes caused by a number of design factors like changes in the fuel load mid-flight, or variable CL caused by pilot induced changes in the wing configuration during certain phases of flight, or one of the biggest factors of all, variable changed in density altitude.  The reality of the limitations of desktop flight simulation does begin to set-in at some point when you start to look at things this way.  But, this is part of what I wanted to know, so that I could understand which components of the simulation were being impacted the most - whereupon I could make a judgment about the usefulness of such a tool for my purposes.

 

So each has a limit.  X-Plane's approach will result in a smoother performance envelope....more accurate?  not necessarily, but smoothness is really important to me as far as feel goes so therefore I choose x-plane as my platform in which to try and improve the accuracy.  I think accuracy can be improved in both sims, but I think it's easier in x-plane and I think the final result is smoother in x-plane.  Just looking at something as simple as "turning up the cockpit lighting" shows the difference to me.

 

For me, I'd like to see these flight simulators at least provide a 'decent approximation' of point mass, six DOF, etc., to set up a good inertial frame (force equations) within which the aircraft functions.  After that, I'd like to see the actual aircraft perform decently in areas such as rate of climb, V speeds, polar curve related L/D ratios, FC/SFC, HDA performance, etc., so that I can make some 'reasonable' (not absolutely spot on perfect) assumptions about flight planing, emergency procedures and the like.

 

If it is just a game to someone then hey - no big deal - things like specific fuel consumption derived as a function of accurately predicting things like a headwind component (as just one example) on a long cross country, which itself is tied to the aerodynamic coefficients throughout the entire planform of the aircraft (getting straight back to the issue of theoretical -vs- emperical modeling approaches), will be no real issue for the end user.  But, if someone wants to use this stuff as a tool for enhancing there knowledge about the net effects of what they put into their flight planning in a given aircraft, well then - that's a horse of another color, as getting the aircraft performance data as tight as possible in a desktop world, all of a sudden becomes far more important (actual flight dynamics aside).  There are a lot of people talking about flight dynamics in relationship to "aircraft handling."  However, the issue of flight dynamics in relationship to the "practical flight regimes," is just as important (if not more) for the one seeking a platform for working on procedures and detailed flight planning.

 

I think we are speaking the same language and I appreciate your relevant input to this thread.  I also appreciate the recognition that the "low level" stuff does indeed matter, depending on the needs/requirements of the end user.

 

Thanks.

Link to comment
Share on other sites

I watched the video. He did not say what you stated he did. What he did say:

 

1. OpenStreetMap was utilized to produce the roads and they are not custom, but auto generated based on the roads that people have contributed into the OSM database (almost all). This is a true statement, and Austin left little for interpretation here. He even went to the extent to state that meant overpasses, etc...not buildings. Since the roads are accurately placed in the sim and you can see them all if you pump up your roads settings, I believe the entirety of Austin's claims to be true.

 

2. He stated that the X-Plane world is plausibly 3D. He further clarified this by saying that the clouds, the aircraft, the buildings, the cars, and whatever else were not superimposed or billboarded to a 2D image, be it flat or vertical. They are truly 3D objects. This statement is also true.

 

What Austin said is here:

 

 

And, here:

 

Edited by Wetted Area
Link to comment
Share on other sites

Maybe you can help me out with something. One of the problems ....

Man,  I am rusty on that stuff, I haven't done DiffEQ in some time...thats one of those subjects where I did well at the time but when done said, "Ok noted...lets get back to algebra"

 

 

It is interesting how both took two entirely different paths in their attempt to achieve seemingly the same things.

MSFS goes back to 1980 and desktop computers then were really in their infancy and you just didn't have the processing power you have now to run calculations as fast as you can now.   By the time Austin got into it (he's an aeronautical engineer by the way and is why X-Plane is structured the way it is),  computers were fast enough where he took a more modern approach.  It's not unlike CFD or FEA....stress analysis years ago was done with simplified equations, but now is generally FEA...the reason designers took different approaches is becasue different tools were available to them at the time.  Because MSFS was the first though, their method is what "took hold" and dominates.

 

 

This might explain the "smoother," more non-linear transitions through the XYZ planes in the flight behavior that I'm experiencing with X-Plane default aircraft, as opposed to the more jittery flight behavior with a slightly detectable "stepwise" type motion that I seem to get from the default FSX aircraft

That is right.  FSX will never get smoother as computers get faster.  X-Plane will run with higher framerates and get smoother as 'h' gets smaller (for the old timer engineers out there)

 

 

For me, I'd like to see these flight simulators at least provide a 'decent approximation' of point mass, six DOF, etc., to set up a good inertial frame (force equations) within which the aircraft functions

This is the way x-plane works now....its "decency" is debatable of course.   There is no frequency component of course, only forces.   Austin uses a series of user definable point masses for Cg, fuel cell locations and weapons locations.  Beyond that, he "empericizes" things like engine mass, radius of gyration, passenger location.    These values can be dynamically modified through custom programming to get more realistic.     Basically the way it works is Austin takes an element of a wing surface, it has an airfoil specification and that airfoil has a lift/drag curve associated with it (X-Plane's foil-maker app).  Based on the  AoA and airflow vectors, austin can figure the force and drag on that wing element from the airfoil curve.  The more wing elements you spec, the more accurate you can get....just like FEA.  That holds true for all the lifting surfaces.  He estimates drag from gear (but the user can add more)....lift and drag from flaps is set via coefficients....the fuselage body is used for side area and side loads...but Austin also implements drag with rapid changes in cross-section of the fuselage along the air stream.  Without sitting down with Austin and going over equations and code, it's hard to say how in-depth his state equations are...what I purport here is what I've gathered over the years of discussing certain aspects of the flight model with him and my own experimentation.  

 

The end result of all Austin's force calculations in a "calculation loop" are 6 instantaneous DoFs....for position, velocity and accelerations...these values can be examined in sim.

 

TomK

Edited by tkyler
Link to comment
Share on other sites

Man,  I am rusty on that stuff, I haven't done DiffEQ in some time...thats one of those subjects where I did well at the time but when done said, "Ok noted...lets get back to algebra"

 

Yeah, no kidding.  The "use it or lose it" axiom pretty much covers almost anything in aeronautical and aerospace engineering.  I ended up working in enterprise software after leaving school, so I never actually applied my formal education in aerodynamics.  However, I did end up going back to work on projects with all the major aerospace and aircraft manufacturing companies in the U.S., but as an enterprise software systems engineer through the software companies that I worked with over the years. 

 

So, I was able to stay somewhat "in touch" with aerospace, but not at ground zero on anything.  I filled a niche role for the software companies, having some knowledge of both aerospace and enterprise software technology, I was able to work between both aeronautical integrations engineers and product drivers on one hand, and then go back to the office and articulate problems/solutions to the software engineers. Back when I was coming out of school, there were not a lot of jobs for aerospace science engineers.  Most of the tech jobs were all about electrical engineering or double-e computer science (I did not see that freight train coming).

 

Bottom line - multi-dimensional mathematics a language, and if you don't use it every day of your life, you will lose it over time.  That sure did happen to me.  But, I can still do a small amount of Rock'n Roll sometimes - typically in the area of differential equations, or non-linear extractions/transformation type stuff.  But, I don't have super chops anymore, that's for sure.  The software business has been good to me and I cannot complain about the career shift away from pure aerospace.  However, with that said - when I look at what Elon Musk, is doing right now, it makes me drool.  Not in envy, but in awe.  If I were able to build a company today focused on aerospace, he's doing exactly what I would want to do with SpaceX.  He's also doing some amazing stuff with Tesla Motors, but I digress quite far.

 

 

MSFS goes back to 1980 and desktop computers then were really in their infancy and you just didn't have the processing power you have now to run calculations as fast as you can now.   By the time Austin got into it (he's an aeronautical engineer by the way and is why X-Plane is structured the way it is),

 

Ah, yes.  Now, it starts to make better sense.  I did not know he was also indoctrinated in the field.  So, we are pretty much aligned fairly well and fairly consistently in terms of educational background and interests.  Do you know if he worked in the software industry before getting started with X-Plane?  I got my initial start in the business with Oracle, but I've worked with a number of enterprise tech companies in SV.

 

Small world.

 

 

computers were fast enough where he took a more modern approach.  It's not unlike CFD or FEA....stress analysis years ago was done with simplified equations, but now is generally FEA...the reason designers took different approaches is becasue different tools were available to them at the time.  Because MSFS was the first though, their method is what "took hold" and dominates.

 

If you have not seen it already, you'll most likely get a chuckle out of this they way I did when I first watched it:

 

 

 

This is the way x-plane works now....its "decency" is debatable of course.   There is no frequency component of course, only forces.

 

Adding frequency is just probably not doable at these retail box levels.  The transformations (FT and even FFT) on those calculations alone would shut down most the desktop machines currently running X-Plane right now.  Of course, there are some variant transforms techniques that might be worth exploring (maybe, lol!).

 

 

Austin uses a series of user definable point masses for Cg, fuel cell locations and weapons locations.  Beyond that, he "empericizes" things like engine mass, radius of gyration, passenger location.    These values can be dynamically modified through custom programming to get more realistic.

 

I assume he's documented a client side API for that?

 

Basically the way it works is Austin takes an element of a wing surface, it has an airfoil specification and that airfoil has a lift/drag curve associated with it (X-Plane's foil-maker app).  Based on the  AoA and airflow vectors, austin can figure the force and drag on that wing element from the airfoil curve.  The more wing elements you spec, the more accurate you can get....just like FEA.

 

True.  And, I think it is important to also realize that FEA can and should be bolstered with other methodologies such as lumped parameter analysis, finite differentials, finite volume studies, etc.  Another potential issue is that fact that more independent corollary proofing methods need to be integrated into not just vehicle flight dynamics, but vehicle structural performance analyses as well.  There also needs to be a better bridging of the gap between the human interface components of aircraft concept performance analysis and the actual flight testing environment, to provide a tighter (more integrated) level of fluidity between both disciplines.  This is one of those areas where I used to function a technical liaison type role between aircraft developers and software & system component developers.

 

It is also one of the reasons why I want to approach Garmin.  I have an idea that I think can revolutionize the way pilots interact with Garmin Avionics and the entire cockpit and as far as FAA certification is concerned, they would only be adding one additional STC to their growing list of type certs.  I think that if the industry (General Aviation) is going to "say" that it now features "Technically Advanced Aircraft," then it really ought to have genuine advancements that drive that point home and that make flying safer at the same time.

 

We always talk about a "fully integrated avionics."  But, we never talk about a "fully integrated pilot," and that is what I visualize on the horizon. I think Garmin might be in one of the best positions to make that happen - with the right product vision.  If not Garmin, then Avidyne.

 

Now, I really digress! (sorry)  I was sort of thinking aloud on that last paragraph.

 

...the fuselage body is used for side area and side loads...but Austin also implements drag with rapid changes in cross-section of the fuselage along the air stream.

 

I would assume that he's not actually changing the area rule geometry of the fuselage?

 

 

Without sitting down with Austin and going over equations and code, it's hard to say how in-depth his state equations are...what I purport here is what I've gathered over the years of discussing certain aspects of the flight model with him and my own experimentation. 

 

The end result of all Austin's force calculations in a "calculation loop" are 6 instantaneous DoFs....for position, velocity and accelerations...these values can be examined in sim.

 

TomK

 

After discussion some of this with you, it sounds like he knows what he's doing to a large degree and knows what he's trying to achieve and that was one of my main questions about Austin, and about X-Plane.  With the shear processing power of today's desktop computer, you can make almost any "animation" look so-called "real."  FSX, is a great example of that.  What Lockheed Martin has done and will continue to do with P3D, is another story entirely. 

 

However, my early suspicion about four weeks ago, was that there was something different about X-Plane.  My analytical nature caused me to investigate further and to go ahead and try both applications side-by-side and on alternating days.  My hope (as I have mentioned before on this forum) was that I could use one of these solutions as a tool for memory work and procedures skills enhancement which does include long-range flight profile modeling in specific aircraft.  My fear was that the flight dynamics might be ok, but the modal aircraft performance (in those dimensions mentioned earlier in this thread) might not be ok, making any attempt to simulate actual "flight planning" a moot point.  Apparently, I can get a good dose of both with X-Plane, without learning bad habits.

 

Now, that speaks for just the Flight Simulator itself.  However, there is potential pitfall I need help with.

 

I you recall, I need four (4) specific aircraft types:

 

1) SEL trainer with steam gauges (152, 172, 182RG , PA28R, AA-5x [either])

2) ME prop with G1000 (Duke, Duchess, Seminole, Baron, Comanche [either])

3) ME turbo-prop with G1000 (King Air C90, King Air Super 200, Cessna Conquest 441 [either])

4) Phenom 300 with either G1000 or G3000 (exclusive)

 

Each aircraft increasing in weight, performance and complexity, finally arriving at 17,000 - 18,000 lbs. MTOW.

 

Who has the absolute best X-Plane based aircraft models for these four areas?  I'm trying to narrow my search for only the best in class that fit these requirements.  Or, will I end up having to make these aircraft models myself - something I am trying to avoid because just getting the geometry models together, looks like it would take considerable time:

 

 

Regards.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...