JSBSim Piper J-3 Cub

[dany93]

I do not think that is a good idea to arbitrarily change these coefficients just to make the aircraft feel better. I leave that approach to Yasim aficionados.

I do agree. But my proposed changes did not do this. They are objectively justified and / or based on publications. That's not fiddling.

[EDIT] Reminder: CoG location: see this message.

Your weight and balance for the J3 points to your own disk.

[wlbragg]

If one had access to a parked J3 in a garage with the wings off, would one be able to "easily" confirm any needed data?

[dany93]

One major problem that I have is the inconsistency of wing root incidence in the various Cub original drawings. In many of them it appears that the lower wing surface, which is almost flat, is set at 0 degrees incidence. The incidence datum is (according to the aircraft rigging instructions) the top of the cockpit door)

On these two, wing incidence 2 deg is specified.
https://scindeks-clanci.ceon.rs/data/pd ... 01031Q.pdf
https://archive.org/details/piperj3_Bin ... g_diamgram
On this one below, nothing is specified, the lower wing surface is parallel to the top of the cockpit door, but the reference axis (mixed dotted line) is something from the propeller axis to the the horizontal stab. Which clearly gives an angle (close to 3.8 deg). The reference axis is questionable. This one is probably not an aerodynamic axis.
https://archive.org/details/piperj3_Bin ... e_Airplane

[Alant]

A bit more progress.
I have modified my original Datcom model. The main changes are to better model the tailplane and the wing tips. Datcom only supports simple wing shapes, so I have done the best that I can achieve whilst maintaining the area and position of the shapes.
Here are lift, drag and pitch moment curves showing the coefficient build-up of the whole aircraft from tailplane, fuselage and wing components.
Lift, drag and pitch moment coefficients are shown for the wing, horizontal tailplane and the fuselage body, as well as for the wing+ body and wing+body+tail which is the whole aircraft.
It is noticeable that the pitch moment (CM) of the whole aircraft does not match the sum of the pitch moments of the various components!
This because when added together the wing, tailplane and body are no longer isolated, but interact with each other. In particular, the tailplane is now in the downwash of the wing, which reduces its effect dramatically.
These results are in good agreement with report https://scindeks-clanci.ceon.rs/data/pdf/1820-0206/2015/1820-02061501031Q.pdf
My next steps are to repeat this coefficient build-up using OpenVSP and recheck my old weight and balance calculations.
Alan

[dany93]

Interesting. Thank you, Alan.

[Ysop]

Looks like really good progress. Thanks and also thanks for keeping us up to date.

I was working on the aeroelastic part, but with a lot of unknowns it is more or less guesswork. Or re-engineering. Could take some time due to other commitments.

[Alant]

Ysop - that sounds interesting and potentially complicated.

Some papers have been published showing how OpenVSP could be used for this. e.g. https://vtechworks.lib.vt.edu/bitstream/handle/10919/109536/Sarode_VS_T_2022.pdf
As OpenVSP calculates spanwise lift distribution, it could be a useful tool as part of a finite element analysis (FEA) study.

This takes me right back to the 1960´s, when as part of my training I spent several months working on FEA flutter analysis for the wings (with ailerons) of the BAC 3-11 aircraft. I would code up data for overnight runs on the company mainframe, which was an ICL 1900 series machine occupying several large air-conditioned rooms. More often than not, when the results were returned to me the next morning, the run had failed due to a misplaced comma or similar error in the middle of the paper tape.

I wish you well with this.

[Ysop]

Thanks Alan for the kind words.

Yes, such a kind of approach is planned. But currently only very basic structural models are made to get an understanding, what is happening under static conditions. Control reversal for aileron might be interesting to investigate. There is not much wing structure restricting torsion - but the two struts really set the boundary conditions.
If a usable structural model has been set up, it would of course be very interesting to have a look at it under divergence or flutter effects. As always there is so much to learn in FGFS, in this case brushing up FEA skills.

As material thickness are largely unknown: Do you already have a reasonable wing weight? This would give something to crosscheck. If not I can of course provide my estimations.

[Alant]

No, I don´t have any weight estimates.
There are a large number of Cub original drawings to be found on the Internet. Sadly, most are badly scanned and the text details are hard to read. If you can find one that has readable wall thickness for the steel tube that was used for the Cub framework that would be a good start. The wing ribs were wood or aluminium at various times. Again finding a good drawing would be a start. I would assume that if a particular diameter steel tube were used in the wing, tube of the the same thickness for that diameter would be used in the fuselage and tail etc.
Piper almost certainly used stock steel tubes, so you could look at the sizes that were commonly available.

Another approach.
When I made my TSR2 I could not find moments of inertia. I made 3d drawings using Solidworks. From the published empty weight I subtracted engine, crew, payload, fuel and estimated undercarriage weights. The remaining weight I assumed to be uniformly distributed on the outer shell of the structure. From this, Solidworks gave me moments of inertia of the empty airfame. I then added the U/C, engines,payload, crew and fuel weights using JSBSim pointmass.
A year or so later I got access to the original factory archives, and found that I was within 5% of their original inertia estimates.
I think that Blender offers a similar inertia calculation facility to that of Solidworks.
Perhaps this type approach could give an estimate of the Cub framework weights.

Once again, best of luck with this. Getting the data is going to be one of your main problems.
Alan

[erik]

No, I don´t have any weight estimates.
There are a large number of Cub original drawings to be found on the Internet. Sadly, most are badly scanned and the text details are hard to read. If you can find one that has readable wall thickness for the steel tube that was used for the Cub framework that would be a good start. The wing ribs were wood or aluminium at various times. Again finding a good drawing would be a start. I would assume that if a particular diameter steel tube were used in the wing, tube of the the same thickness for that diameter would be used in the fuselage and tail etc.
Piper almost certainly used stock steel tubes, so you could look at the sizes that were commonly available.

https://www.univair.com/content/partcat ... age73.html

Erik

[Alant]

Erik - that is a good find. Using that data the weight and stiffness of the Cub´s tube framework can be estimated.
There is a lot more info available on the Univair website https://resources.univair.com/docs/general-information

[Ysop]

Thanks Erik!

For the fuselage there is a good drawing in the archive.org package.

Wood spars are 1'' sika spruce. In aluminium certainly something of same stiffness/strength.

On the part list Friese ailerons are mentioned. Maybe both Friese and plain versions were existing?

Plenty of stuff to go through! Will take time with summer activities still going on.

[wlbragg]

@david.megginson

Just in case you're interested I pushed the last version of the J3Cub that contained the YASim FDM.

You could get to it in either the GitHub history or the SourceForge svn history, but this should make it easier and I also had to fix a bug with the prop spin.

[wlbragg]

Putting the PA-18 to the test. We definitely hit the ceiling.

[david.megginson]

@david.megginson

Just in case you're interested I pushed the last version of the J3Cub that contained the YASim FDM.

You could get to it in either the GitHub history or the SourceForge svn history, but this should make it easier and I also had to fix a bug with the prop spin.

Thank you!