FSXA (DC-9) Effect of chord and span decrease on lift and drag

[Heretic]

I'm working on a FDE for a payware DC-9 at the moment.

My reference data only covers the series 30 model and a bit of the series 10 model. As a starting point, I'm working on tackled lift, drag and engine tables, along with aircraft.cfg tweaks.

Table 401 was reset to 1.
Table 404 was constructed from series 30 flight test data (C_L max) and some reference images* (corresponding AoA).
Tables 15xx will be recycled JT8D tables made for the 727.
Table 154a was constructed from a drag polar for the series 30.

For verfication, I'll have to use the cruise charts from another payware DC-9 since the original flight manual only features an EPR table without any N1/2 indication whatsoever.

So while I can cook up a believable FDE for the series 30, 40 and 50 the wing changes between the series 10 and series 30 and the series 30 and series 20 have me a bit stumped.



Problem 1 - Series 30 to series 10:

Essentially, the difference between the series 10's and series 30's wings is composed of a 6% increase in chord (at the root) and a 4 ft increase in span.
See here**:

In Schaufele's/Shevell's paper on the design features of the DC-9***, the difference and effect of the modification is described:

Almost all airplanes have follow-on versions, but very few have them under construction before the original airplane enters service. The DC-9 Series 30 has a fuselage longer than that of the series 10 by 179 in., wing span greater by 4 ft, a wing-chord increase of 6%, and leading-edge slats.
...
Since the Series 30 is equipped with leading-edge slats, and the maximum lift coefficient is much less dependent on the shape of the airfoil leading edge when slats are used, the airfoil ahead of the front spar is entirely different on the Series 30 airplane. The chord is extended by 6% and the contour redesigned to minimize the gradual drag rise before the drag-divergence Mach number, at the expense of basic-airfoil maximum lift coefficient. With slats extended, the maximum lift coefficient is very little different from what it would have been if the original airfoil had been retained. The result is an improvement in specific range in the cruise region and a very large increase in the maximum lift coefficient in the takeoff and landing regimes with slats extended. Use of slats required extensive wind-tunnel optimization of slat angle and position to optimize maximum lift coefficient, slat drag in ground acceleration, and slat drag in the takeoff-climb configuration.

I am unsure how to translate this into the FDE. For me, the critical items in the above text are:

1) Drag-rise before drag divergence
Since this is less for the series 30 than for the series 10, I interpret this as having to slightly increase the slope of the drag curve at higher mach numbers in table 154a.
But the exact amount is a bit of a mystery, so I'd probably use the ratio between aspect ratio of the series 10 and series 30 as a scalar.
Or can I get away with simply increasing the induced or parasite drag scalars in the aircraft.cfg?

2) Basic airfoil maximum lift coefficient
This is lower for the series 30 than for the series 10 without slats, but nearly equal when the slats are extended.
I have a maximum lift coefficent diagram for the series 30 that plots C_L_maxA versus flap deflection angle (see attachment), so I'd just plug the C_L values for slats extended into table 404 and keep the lower AoA values from the series 30.



Problem 2 - Series 30 to series 20(/40/50):

The problem is based on my lack of knowledge about what the actual wing incidence of a DC-9-10/30/40/50 is. The diagrams from that website* imply that zero-lift incidence is around 2° (tail-off), and Erick's estimations have it at 3° for the wing****.
For the series 20, the only information available is conflicting. The Airliner Café guide says that the series 20, 40 and 50 got an increase in incidence by 1.25° while a comment on the guide said that the increase was 2.5° and only applied to the series 20.
I have no idea what to make of this, but assuming that I implement a change in wing incidence, table 404 would be the only one I'd have to modify, right?
Would I have to shift the entire relevant part of the table (zero lift AoA to AoA max) or is a mod near the zero lift line all it takes?



Lots of text, but I'd appreciate some input on this.



* https://web.archive.org/web/20161206135705/http://adg.stanford.edu/aa241/highlift/highliftintro.html
** http://www.airlinercafe.com/page.php?id=396
*** https://arc.aiaa.org/doi/abs/10.2514/3.43770
**** https://www.fsdeveloper.com/forum/t...rom-an-idiot-savant.438902/page-7#post-788354

 

[Roy Holmes]

1) Drag-rise before drag divergence
Since this is less for the series 30 than for the series 10, I interpret this as having to slightly increase the slope of the drag curve at higher mach numbers in table 154a.
But the exact amount is a bit of a mystery, so I'd probably use the ratio between aspect ratio of the series 10 and series 30 as a scalar.
Or can I get away with simply increasing the induced or parasite drag scalars in the aircraft.cfg?

2) Basic airfoil maximum lift coefficient
This is lower for the series 30 than for the series 10 without slats, but nearly equal when the slats are extended.
I have a maximum lift coefficent diagram for the series 30 that plots C_L_maxA versus flap deflection angle (see attachment), so I'd just plug the C_L values for slats extended into table 404 and keep the lower AoA values from the series 30.

1. I think what he is saying in that document and your quote from it is that the drag rise is reduced in the region slower than where it rises rapidly due to divergence (caused by local exceeding of mach 1). This could be reflected in lower drag coefficient drag values on the left side of 154a. At any given Mach number the total wing drag coefficient is the sum of CD0 (fixed value) and CD delta Mach (what you have in 154a) You should not be increasing any scalars in the aircraft.cfg. You can do all that in 154a since the X scale increments are what you make of them, unlike its predecessor 403 which had 0.2 Mach increments. As for the size of the value changes I would just say that you would have to make a SWAG unless you have cruise Altitude/thrust data or altitude/Mach/fuel flow. Since fuel flow and TSFC give you thrust and thrust in the cruise equals drag there is a path to follow. Aspect ratio affects induced drag, but that should be at a minimum at cruise speeds.

2. Your 19.jpg shows total lift coefficient versus flap extension at some minimum speed. Different flap extensions give you different minimum speeds with different AOA since the lift coefficients and wing areas are different. About all you can take from that chart is that you have a gain in maximum lift coefficient with slats compared to flaps only. But bear in mind that at any given AOA slats do not increase CL. Slats allow higher AOA to be used before airflow diverges. So you would have a big decrease in minimum speed (good) but it would be at a higher AOA (bad). But without having a substitute 404 table using something like XMLTOOLS that came into play when slats are down and which gave higher stall AOA than the one which it replaces, all slats do in the sim is add to the total CL Flap.

As regards wing incidence angle for zero lift it depends on the airfoil shape. A symmetrical airfoil generates no lift at zero AOA if its chord line is parallel to the aircraft fuselage. A cambered airrfoil generates zero lift if its chord line is negative (nose down) relative to the fuselage. So if you want the airfoil to be at 2 degrees AOA when the fuselage is horizontal and in level flight, you put the symmetrical one at 2 degrees positive chord line and maybe have the cambered one parallel to the fuselage.
This means in sim terms which reference the fuselage AOA, the zero lift point for symmetrical airfoils would be at zero AOA and the cambered one at -2 AOA. If you wanted to modify a 404 to reflect a change from symmetrical to cambered, all other things being equal, you would subtract 2 degrees or whatever from the CL reading locations in the central curve and leave the rest unchanged.

The sim only considers an average CL for the whole wing. You can do separate calculations for the root and tip chords which as far as I can see go from near symmetrical at the root in the DC-9 to quite cambered at the tip. This gives effective washout of incidence for the tips. There is a slot for wing twist in the aircraft.cfg but it is eye candy only, does nothing. So if you did separate calculations what you can do is estimate the amount of the wing area which is affected by the root and a smaller area which is affected by the tip. You would then estimate an overall lift coefficient versus fuselage AOA curve and use that.

To sum up, the key factors are the wing area and the average spanwise CL for the flight conditions.
In the sim you can only have one wing area in the aircraft.cfg. so when flaps increase the effective wing area you have to increase the flap CL accordingly.
Hope this helps and PM me if you want anything more definitive.
Roy

 

[Heretic]

1. I think what he is saying in that document and your quote from it is that the drag rise is reduced in the region slower than where it rises rapidly due to divergence (caused by local exceeding of mach 1). This could be reflected in lower drag coefficient drag values on the left side of 154a. At any given Mach number the total wing drag coefficient is the sum of CD0 (fixed value) and CD delta Mach (what you have in 154a) You should not be increasing any scalars in the aircraft.cfg. You can do all that in 154a since the X scale increments are what you make of them, unlike its predecessor 403 which had 0.2 Mach increments. As for the size of the value changes I would just say that you would have to make a SWAG unless you have cruise Altitude/thrust data or altitude/Mach/fuel flow. Since fuel flow and TSFC give you thrust and thrust in the cruise equals drag there is a path to follow. Aspect ratio affects induced drag, but that should be at a minimum at cruise speeds.

So for the series 10, I'd have to increase drag rise.
I don't have a cruise chart to verify any mods though (haven't tried to find any yet), so a SWAG (had to look that up) it is.

2. Your 19.jpg shows total lift coefficient versus flap extension at some minimum speed. Different flap extensions give you different minimum speeds with different AOA since the lift coefficients and wing areas are different. About all you can take from that chart is that you have a gain in maximum lift coefficient with slats compared to flaps only. But bear in mind that at any given AOA slats do not increase CL. Slats allow higher AOA to be used before airflow diverges. So you would have a big decrease in minimum speed (good) but it would be at a higher AOA (bad). But without having a substitute 404 table using something like XMLTOOLS that came into play when slats are down and which gave higher stall AOA than the one which it replaces, all slats do in the sim is add to the total CL Flap.

Forgot about the AoA limitation and XMLTools won't be of much use for an entire replacement table.
The only thing I could do is setting up 404 for C_Lmax/AoA_max for slats extended and then use a gauge that triggers a stick push at the clean stall AoA when the slats are not extended.

As regards wing incidence angle for zero lift it depends on the airfoil shape. A symmetrical airfoil generates no lift at zero AOA if its chord line is parallel to the aircraft fuselage. A cambered airrfoil generates zero lift if its chord line is negative (nose down) relative to the fuselage. So if you want the airfoil to be at 2 degrees AOA when the fuselage is horizontal and in level flight, you put the symmetrical one at 2 degrees positive chord line and maybe have the cambered one parallel to the fuselage.
This means in sim terms which reference the fuselage AOA, the zero lift point for symmetrical airfoils would be at zero AOA and the cambered one at -2 AOA. If you wanted to modify a 404 to reflect a change from symmetrical to cambered, all other things being equal, you would subtract 2 degrees or whatever from the CL reading locations in the central curve and leave the rest unchanged.

The sim only considers an average CL for the whole wing. You can do separate calculations for the root and tip chords which as far as I can see go from near symmetrical at the root in the DC-9 to quite cambered at the tip. This gives effective washout of incidence for the tips. There is a slot for wing twist in the aircraft.cfg but it is eye candy only, does nothing. So if you did separate calculations what you can do is estimate the amount of the wing area which is affected by the root and a smaller area which is affected by the tip. You would then estimate an overall lift coefficient versus fuselage AOA curve and use that.

To sum up, the key factors are the wing area and the average spanwise CL for the flight conditions.
In the sim you can only have one wing area in the aircraft.cfg. so when flaps increase the effective wing area you have to increase the flap CL accordingly.
Hope this helps and PM me if you want anything more definitive.

I'll need a bit to process that and will get back to you.


Anyway, thanks for the extensive reply!

 

[jx_]

always set wing geometry first to the real numbers! Wing incidence is set in the cfg

then CL vs AoA should be set so that the CL is referenced to the fuselage (body angle negative to reach wing angle zero). Your Douglas charts will already be this way, copy the clean wing data into the table as is. Do not make any changes!

Zero out and ground effect above 100 feet
Zero out CL/AoA vs Mach for now (unless you have the real charts...then fill em in!)

IMPORTANT: CL vs AoA chart is clean (no flaps or slats). Douglas will mark it as Vsr or Vf0 That means V reference speed: Slat Retract or Flap Zero. Do not use Vse or V## data

Plot the DIFFERENCE between flap50/slat extended lift and drag vs clean lift and drag for the real airplane and add to the cfg file flap sections. (the Douglas flap/drag charts are usually mach 0.20). To Roy's point, real slats don't really add lift so you'll probably have a really small number like 0.05 for the slat lift scalar. FS doesn't model this correctly. Don't sweat it, a pilot won't notice that he is stalling at the wrong AoA nor is it useful for a pilot to fly a DC9 beyond 16 degrees AoA. Just make sure the drag and pitching moment are correct. A DC9 pitches up slightly with slats out as you can see in the document (if you need that document let me know, I have a free copy.)

IMPORTANT: FS uses a generic formula to find lift and drag. Make one set of flaps that has 0 lift then set drag and one with 0 drag then set lift; when setting lift and drag you need to verify the actual lift additive at full flaps, divide full flap angle/57.2957, take that number "N" and divide your Douglas chart full flap CL/"N" to get the theoretical CL for 1.0 radian flaps. Remember the lift and drag are set at 1 radian in the air file. Lift and drag will be slightly different, hence the two entries. Lift will not be linear, therefore, you will want to center the accurate lift number on the normal landing flap angle. If you want hyper accurate modeling, I can explain how to do that as well but it get's complicated using the in-sim dynamics engine.

Set your center of lift correctly in the air file! (this affects trim and pitch moments)

verify the reference position in your CFG is calibrated to the reference position in your data and verify the wing position, chord, geometry, etc

That wing has an Oswald e of 72.


Once that's done, your questions are simple:


1) Don't touch induced or parasite, those are not relevant here. Your changes will be in CDm Mach drag vs Mach. The newer wing is more efficient in compression. You need a drag polar at higher mach or you'll need to reverse engineer fuel burns to find the exact number. I've attached a table below.


2) use the Stanford table at zero flaps slat retracted....the one at the bottom. Put that in table 404. The effect of slats on CL are shown on that table as well. There's no way to model that in FS (and it's rarely encountered.)


Problem 2) No! Do not change your CL table for incidence, the CL table is written AT the body angle of attack with incidence built in already. So for each model that has a different incidence you will need a different CL table. If you don't have one that's okay they are easy to make if you have any kind of flight data with aircraft pitch, temp/pressure alt, and speed. The FS entry of incidence is only used to adjust the 3D model and HSI pitch, so if you have the correct CL and the airplane is at the wrong pitch, offset it to correct it (assuming all your other entries are correct, the airplane will fly at the right lift and drag even if the incidence is set wrong, but the model and gauges will be off.)



And here are the mach drag curves for the DC9... (the bottom is the -30, in the first two lines it curves up to exceed the -10 above .75 mach)

 

[Heretic]

Thanks Joe. A bit late, but better than never.

I'll get back to it if I ever pick up enough motivation again.