Possible stores drag inconsistencies

[Bestandskraft]

I have noticed what seem to be abnormalities in stores drag modelling for the F1.

All empirical data for the following study were obtained in an F1EE with 10% fuel remaining, unlimited fuel selected, starting at 1000 ft MSL, 0.65 M, on a standard day, no wind.

The study consisted in air starting the jet with MIL power already selected, then, upon entering the cockpit, immediately engaging AP in altitude hold mode and measuring the time until the jet reached 600 KIAS as per the F2 view.

The jet was loaded only with the store given in the diagram on station 4, 0% gun, 30 Chaff, 15 Flares.

The cx_pil values were obtained as per the following table:

Store	cx_pil	Loadout	URL
Belouga	0,000413		Link
CLB4-BLU-107	0,0006	X	Link
CLB4-Mk-82	0,00076	X	Link
CLB4-SAMP-125	0,00052	X	Link
CLB4-SAMP-250LD	0,00092	X	Link
CLB4-SAMP-400LD	0,00124	X	Link
GBU-10	0,000508		Link
GBU-12	0,000365		Link
Mk-82	0,00016		Link
Mk-83	0,000225		Link
SAMP-250LD	0,00012	X	Link
SAMP-400LD	0,0002	X	Link
Clean	0

Results

Interpretation

1. All stores with an "X" in the Loadout column above have a cx_pil defined in the /_G/launcher/ folder of the datamine. The cx_pil from other stores had to be obtained from other places (see respective link).
2. Stores with a cx_pil in the Loadout folder (blue in diagram) seem to follow a certain logic in that a higher cx_pil causes a longer time for the aircraft to reach the target speed. Notable exceptions are the SAMP-400LD and the CLB4-SAMP-125 with cause the aircraft to reach the target speed in the same time span even though the cx_pil is significantly different. The same can be said when comparing the CLB4-Mk-82 and the CLB4-SAMP-250LB.
3. Stores without a cx_pil in the Loadout folder (orange in diagram) do not seem to follow a discernible logic in that an increased cx_pil does not necessarily imply an increased time to reach the target speed. This is most notable with the GBU-12 and Belouga.
4. Empirical results not provided here, but my testing (acceleration runs with stores and with the same stores after jettison) has shown that the only pylons that seem to have a measurable drag are the CLB-4s.

Hypotheses

1. Stores drag is incorrectly modelled in the Mirage F1EE, at least for some stores and/or pylons.
2. "External" stores not specifically created for the Mirage F1 module interact with the F1 source code in a unexpected way, causing deviations from expected behaviour.
3. There is another factor in play that determines stores drag, such as the store's cross sectional area. Since the loadout's (that is, weapon plus pylon) cross sectional area is not defined in any .lua I know of, reverse-engineering and accurately cataloguing stores drag might be impossible. This is especially true for CLB-4 loadouts whose cross sectional area cannot simply be calculated using a weapon's diameter and assuming a circular shape.

Request

• I'd appreciate if anyone could shed some light on / explain the above apparent inconsistencies.

Edited July 20, 2023 by Bestandskraft

[fausete]

Hi @Bestandskraft,

So, cx_pil is only used by the AI, the drag coefficient of the stores that the flyable aircraft uses is calculated internally. The real effect of a store on drag coefficient will depend on the interference drag too, so it will vary on each aircraft and even on the specific configuration. Using our own values allows us to reflect the specific data we have and to take Mach number into account, which would not be possible just by taking the value from the lua. For example, we use values based on the performance data that we have available.

That being said, I believe you have noticed a bug in the Belouga and maybe GBU-12 bombs. Thanks for the thorough report, we will try to correct the Belouga bug for the next update.