Board index FlightGear Development Aircraft

JSBSim Piper J-3 Cub  Topic is solved

Questions and discussion about creating aircraft. Flight dynamics, 3d models, cockpits, systems, animation, textures.

Re: JSBSim Piper J-3 Cub

Postby Bomber » Fri Aug 04, 2017 9:04 pm

Tune up ?
"If anyone ever tells you anything about an aeroplane which is so bloody complicated you can't understand it, take it from me - it's all balls" - R J Mitchel
Bomber
 
Posts: 1934
Joined: Fri Dec 14, 2007 7:06 pm
OS: Windows XP and 10

Re: JSBSim Piper J-3 Cub

Postby dany93 » Fri Aug 04, 2017 10:08 pm

erik wrote in Fri Aug 04, 2017 5:23 pm:I'm not a big fan of messing with the numbers just because they feel right.
Any change that the center of gravity is too far to the tail for instance?

Don't you feel that "Messing" is rather strong?
The numbers I propose are neither invented nor "taken out from the hat".

I've already written several times that I do not like cheating with the numbers either.
But in this case the FDM has a problem and I'm looking to find where it can be.

The factor for "Pitch moment due to alpha" in your FDM is in the -2 to -1.3 range
Code: Select all
    <function name="aero/moment/Pitch_alpha">
       <description>Pitch moment due to alpha</description>
       <product>
          [...]
                <independentVar lookup="row">aero/Re</independentVar>
                <tableData>     <!-- Du Y -->
                    1668183  -2.0327
                    3707224  -1.3432

I proposed this change, not randomly but, because Aeromatic (v. 0.96) for the J3Cub and Megginson's tables for the Cessna182 give a factor from -0.5 to -0.65 (see Cmalpha on Megginson's tables). And even Aeromatic v 0.96 is not so bad IMO.
After that, as it can depend on the aircraft, we can consider changing it a bit (hence my factor to make tests easier) but, in my spirit, around these values. Unless we have a more reliable source, or a good reason to presume that Megginson's values are wrong.

Unfortunately I do not know how to calculate it. I would be a bit embarrassed on which area to take (horizontal stabilizer alone? with the elevator?) for the alpha variation. The elevator participates to this effect but its effect for each alpha depends on its deflection. And the issue here is that the horizontal stabilizer counteracts the elevator moment. Possible (by CFD?) but difficult to translate in linear effect or by tables. And will it be noticeably better than a more simple solution?
How did you have these numbers? Yongliang Du Thesis? Do you know how to calculate them?

About the center of gravity, its location is rather relative to the AERORP. At first view, I would not change this.
Also, I have observed that the CG location for different loads remains inside the books limits. Which makes me think that the empty CG location is right.
Or not far. (unless I've done this checking too quickly)
I guess that this is your opinion too, otherwise you would have proposed to change it.
dany93
 
Posts: 783
Joined: Mon Sep 07, 2009 3:43 pm
Location: France (Paris region)
Version: 2020.2.0
OS: Linux Mint 18 (64 b)

Re: JSBSim Piper J-3 Cub

Postby Alant » Fri Aug 04, 2017 11:07 pm

The Yongliang Du Thesis (https://dspace.lib.cranfield.ac.uk/bits ... s_2011.pdf) uses Datcom and AVL to estimate aerodynamics. Either of these is an advance on Aeromatic, which gives a generic model.

Perhaps there are wind tunnel or flight test data publically available. These should give better results.

Anyone with a private pilots license and undergraduate aerodynamics training should be able to instrument an aircraft and conduct a simple flight test to get this parameter. Doing this was part of my degree course in the 1960´s.

I would stick with the Datcom numbers. They should be quite accurate for this class of aircraft.

Alan

Alan
Alant
 
Posts: 956
Joined: Wed Jun 23, 2010 5:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: JSBSim Piper J-3 Cub

Postby wlbragg » Sat Aug 05, 2017 3:56 am

@dany93

Thank you for clarifying the numbers you propose.

just inventing numbers is never good;

I totally get that and I agree, obviously so does dany93. But at the same time, there are a couple issues here that need correcting. I think everyone wants this aircraft to perform as it does in real life. Much time has been spent implementing the various configurations to make this a really versatile aircraft simulation and I really want it to be the best it can.

"Politically" correct or not (so to speak), the recent changes to the FDM have improved the versatility drastically.

You can now land on all terrain without endowing.
You can stall the aircraft.
You can perform a 3 point landing.
The flaps in the SuperCub actually work and serve a purpose.
You can touch down on water in a realistically short distance.

None of these things were possible or at best were flawed before these changes.

All I am after is the simulation of reality.

I don't want the FDM compromised in a less than desirable fashion, but I desperately want it to work and feel right.

So, where does this leave things and what to do?
Kansas(2-27-15)/Ohio/Midwest scenery development.
KEQA (2-27-15), 3AU, KRCP Airport Layout
Intel i5 3570K AMDRX480
User avatar
wlbragg
 
Posts: 5450
Joined: Sat Aug 25, 2012 11:31 pm
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/AMDRX480

Re: JSBSim Piper J-3 Cub

Postby erik » Sat Aug 05, 2017 8:08 am

dany93 wrote in Fri Aug 04, 2017 10:08 pm:Don't you feel that "Messing" is rather strong?

Maybe we have a different feeling with that word, it was not meant to be strong.
The numbers I propose are neither invented nor "taken out from the hat".

Maybe, but neither is the one I put in the configuration. In fact, unless some mistake was made somewhere either by me or the author of the paper, it should match every other flight dynamic number in the config file.
It is good to search for solution for a problem, but I'm not really sure we already know what the problem exactly is. For instance, tonight I remembered that the pilot is in the backseat for the J3 Cub. This will hurt elevator responsiveness. Maybe it's a tad to far too the back and moving it slightly too the CG will solve the problem.
I've already written several times that I do not like cheating with the numbers either.
But in this case the FDM has a problem and I'm looking to find where it can be.

This is great, it gives an indication whit might be wrong. But after that it is always good to see if you fixed the problem or just covered it up.
How did you have these numbers? Yongliang Du Thesis? Do you know how to calculate them?

Yes and yes. I did write AeromatiC++ (which by the way gives yet another number for Pitch_alpha somewhere between your proposal and the one from the thesis).
About the center of gravity, its location is rather relative to the AERORP. At first view, I would not change this.
Also, I have observed that the CG location for different loads remains inside the books limits. Which makes me think that the empty CG location is right.

About the locations: those did not come from the thesis but were generated by AeromatiC++ (unfortunately I don't remember if I double checked them by comparing them to three views). If anything I have less faith in them than in the aerodynamic numbers from the thesis.

Erik
erik
 
Posts: 1588
Joined: Thu Nov 01, 2007 1:41 pm

Re: JSBSim Piper J-3 Cub

Postby dany93 » Sat Aug 05, 2017 9:20 am

erik wrote in Sat Aug 05, 2017 8:08 am:......but I'm not really sure we already know what the problem exactly is.
.....But after that it is always good to see if you fixed the problem or just covered it up.

I agree. As I globally agree with, or accept, what you wrote (except maybe for the CG which would not be my first search).

I'm not at all an aeronautics engineer. When I proposed changes, I expected someone's reaction (yours in priority) if there was some doubt or error. I am right or not, that's just proposals for wlbragg after he reads our discussions.
This thread does the job of exchanging ideas and opinions.
But until this day, neither other solution nor better understanding is brought to address the issues. The first thing that we have to be certain is that they are really issues, not in our heads. If the real Cub behaves like this for part of our issues in reality, we have to accept it.
Otherwise, we cannot let it like this.

Not easy, as long as part of our opinion of what should be a "good" behavior has no other base than some subjectivity. However, a good feeling is not a negligible part...
Last edited by dany93 on Sat Aug 05, 2017 4:35 pm, edited 1 time in total.
dany93
 
Posts: 783
Joined: Mon Sep 07, 2009 3:43 pm
Location: France (Paris region)
Version: 2020.2.0
OS: Linux Mint 18 (64 b)

Re: JSBSim Piper J-3 Cub

Postby erik » Sat Aug 05, 2017 11:49 am

First of all, I don t have experience flying the Cub and nobody I've asked ever responded back to me.

But what I think is causing the confusion is that Pitch_alpha is increasing at low speed (aero/Re is velocity and air pressure related).

You could make both values the same for Pitch_alpha and see if that meets your expectations. If it does then I don't think there is a problem after all since a higher velocity means more lift (end hence a lower velocity means less lift for the elevator).

Erik
erik
 
Posts: 1588
Joined: Thu Nov 01, 2007 1:41 pm

Re: JSBSim Piper J-3 Cub

Postby dany93 » Sat Aug 05, 2017 12:51 pm

(I guess you are meaning "aero/Re is velocity and air density related". Velocity here)
A lower value for the Pitch_alpha factor is not incompatible with a Re number dependency.

I am looking for if the disagreement between Megginson's tables and your values may be explained.

For your Pitch_alpha (= Cmalpha) calculations in AeromatiC++, which area did you take? The [horizontal stabilizer] alone? Or the [horizontal stabilizer + elevator] considered as a solid part?
And for Yongliang Du Thesis?
In the FDM, I see <htailarea unit="FT2"> 24.50 </htailarea>, which is the [horizontal stabilizer + elevator] area.

- If the aircraft is with more or less high AoA, but with the elevator held in neutral position (alpha not due to the elevator moment) the area which counteracts alpha to realign the aircraft with the relative wind is [horizontal stabilizer + elevator]. This is a transitory situation.

- If the elevator is deflected (for our case, rather upwards), the elevator does not counteract alpha, it is the cause (as long as alpha is not too high). Here, the area which counteracts alpha is the horizontal stabilizer alone.

This gives about a factor 2 between both.
(By only taking these area to make reasoning more simple)
Last edited by dany93 on Sat Aug 05, 2017 1:35 pm, edited 1 time in total.
dany93
 
Posts: 783
Joined: Mon Sep 07, 2009 3:43 pm
Location: France (Paris region)
Version: 2020.2.0
OS: Linux Mint 18 (64 b)

Re: JSBSim Piper J-3 Cub

Postby erik » Sat Aug 05, 2017 1:29 pm

The tail data was taken from this site: http://www.eaa62.org/technotes/tail.htm

Erik
erik
 
Posts: 1588
Joined: Thu Nov 01, 2007 1:41 pm

Re: JSBSim Piper J-3 Cub

Postby dany93 » Sat Aug 05, 2017 1:50 pm

That is SH = 24.5 Sq. Ft. [horizontal stabilizer + elevator] area.
I just discovered it in the FDM when you were responding.
I also found these areas in the J3Cub Owner Manual, Piper Cub J3C-65, p.12:
Stabilizer: 15.1 Sq. Ft,
Elevator: 11.7 Sq. Ft.

In this case, see my previous message, this area is excessive when the pilot pulls the stick to pitch up. To calculate the moment in this case, the area to take into account, which counteracts alpha, is the horizontal stabilizer alone. Otherwise, the Cmalpha is too high when the pilot pulls up to reach a stall. It is correct only when the elevator is held in neutral position.
Taking into account the horizontal stabilizer alone, the arm for the moment calculation is also a bit smaller.

In fact it is much more complicated. The resulting moment is a combination of alpha and elevator deflection. But as the [high AoA + (close to) neutral elevator] situation is much less durable and important than the pulled up elevator, I would give priority to the second one. Unless willing to build a hard to do table.
dany93
 
Posts: 783
Joined: Mon Sep 07, 2009 3:43 pm
Location: France (Paris region)
Version: 2020.2.0
OS: Linux Mint 18 (64 b)

Re: JSBSim Piper J-3 Cub

Postby Bomber » Sat Aug 05, 2017 3:07 pm

So.......

If I've read it right there's now some thought that having both a h-stab and elevator moment calculation running together could be as I've suggested producing too much moment.

I use javafoil to give me the coefficient for the h-stab.... but I also use it to give me the coefficients for the h-stab when the elevator is fully deflected in both directions.

And I use a 3d table with these results to give me the moment.

If you're interested in going this route I'm quite prepared to knock a table up and talk you through its creation.

Simon.
"If anyone ever tells you anything about an aeroplane which is so bloody complicated you can't understand it, take it from me - it's all balls" - R J Mitchel
Bomber
 
Posts: 1934
Joined: Fri Dec 14, 2007 7:06 pm
OS: Windows XP and 10

Re: JSBSim Piper J-3 Cub

Postby Thorsten » Sat Aug 05, 2017 3:14 pm

The resulting moment is a combination of alpha and elevator deflection


Yes, the pitching moment due to elevator deflection should be a function of alpha as well as deflection angle (I suspect I said this earlier).
Thorsten
 
Posts: 11580
Joined: Mon Nov 02, 2009 8:33 am

Re: JSBSim Piper J-3 Cub

Postby Bomber » Sat Aug 05, 2017 3:21 pm

So are you now agreeing that the way it's been done here at FG for the past 15 years could have been done better ?

And that we could maybe think about doing it a different way ?
"If anyone ever tells you anything about an aeroplane which is so bloody complicated you can't understand it, take it from me - it's all balls" - R J Mitchel
Bomber
 
Posts: 1934
Joined: Fri Dec 14, 2007 7:06 pm
OS: Windows XP and 10

Re: JSBSim Piper J-3 Cub

Postby Thorsten » Sat Aug 05, 2017 4:10 pm

I have no idea what 'the way it's been done here at FG for the past 15 years' is supposed to be - different planes are clearly different and unless you come with a specific example, the statement has no meaning.

Data (wind tunnel, CFD) based aircraft will have elevator moments alpha (and possibly Mach) dependent and as far as I know they've done this for ages (certainly longer than I am around). So have I pretty much from the beginning of the Shuttle FDM and so have I said early in this thread.

Perhaps you have been habitually looking at bad FDMs and ignored the rest?

Anyway, there's no fresh perspective needed, this is the way it's been done for ages and (read back) I've said the same thing all along.

Edit: The upshot is: There's 400+ planes in the repository. The fact that you can pick a number of them and show that they're wrongly modeled neither proves that 'FG does anything wrong' nor that you do things right. It just proves that you can show a red herring.
Thorsten
 
Posts: 11580
Joined: Mon Nov 02, 2009 8:33 am

Re: JSBSim Piper J-3 Cub

Postby Bomber » Sat Aug 05, 2017 5:01 pm

Oh good....post a link to your shuttle FDM and I'll have a look at how YOU do it, as you're not posting an example of 'good practice'...
"If anyone ever tells you anything about an aeroplane which is so bloody complicated you can't understand it, take it from me - it's all balls" - R J Mitchel
Bomber
 
Posts: 1934
Joined: Fri Dec 14, 2007 7:06 pm
OS: Windows XP and 10

PreviousNext

Return to Aircraft

Who is online

Users browsing this forum: No registered users and 5 guests