Stability analisis

[Demetrio Cucarella]

Hi everyone, i am currently working on the analisys of stability of a modified MQ9 Reaper. I have a few problems with the results of the derivatives that the steady mode give me. 

In first place i worked with the V 3.47.0 in wich i ran a Thin-Thik analysis. I started by analisisng the longitudinal polar adding 1 by 1 the surfaces and cheking the output. Then i try to obtain the derivatives with the "Steady" stablity analisis. When i ran only canard-wing-tail on Thin surface the output in terms of magnitud and sign looks pretty well, but when i add the Thik fuselaje the CS derivative change tu positive and change the sign of all the derivatives related to taht one, this doesnt makes any sense to me, as the side force should not change of direction and maintain the absolute value only by adding the fuselage... 

Secondly, I performed a full Panel Method analysis using version 3.42.3. In this case, the derivatives for the wing–canard–tail configuration also appear reasonable, and after adding the fuselage the sign of the derivatives remains consistent, unlike in the Thin–Thick analysis.

This difference in behavior makes me suspect either:

• a problem with the Thin–Thick solver itself, or
• that the beta perturbation (Δβ) used internally by the stability analysis in version 3.47.0 is significantly smaller than in version 3.42.3, making the results more sensitive to numerical noise.

Would you say that this is a problem with the model or that is better only to use a a full VLM solver when running an stability analysis, modeling the fuselage as a cruciform body, because when i ran a beta sweep analysis in mix solver case with the body the CS slope is the opossite of the stability output, so maybe the problem is the stability module when using it with a Thin-Thik solver or a Full panel solver? In case the problem is the model, what would you suggest to modify?

Below i attached the stability outputs on the Thin-Thik and Full PM and the VSP3 file. 

Thanks for the help!!