L shaped array

[luigiV]

Sorry if I propose again a post for which I have not received an answer. It seems to me important to have clear ideas on this subject.
Hi Marc,
having to process a L-shaped array (21 sensors equally spaced by 4m, 45min long) I applied the Mspac technique, but I was struck by the configuration of the co-arrays which are all concentrated to form a rectangle on the vertical side (see image). The result still seems good to me, being able to clearly (I think) define a dispersive trend. However, if I compare the ReMi dispersion curve with that obtained from the Spac plot, I observe a significant difference in terms of phase velocity. The velocities of the Spac curve are too high and poorly in agreement with the geology of the site where a drill hole did not find the bedrock within 15m of depth. The H / V curve shows a nice natural peak at 5.3Hz.
Is it possible that the maxima phase velocities do not define the true dispersion of the fundamental mode?
Some commercial softwares for surface wave analysis propose the Esac method for the L shaped array. In your opinion this kind of process could give a different result from the one obtained with Mspac?
Is there a general rule for the correct definition of the ring's radius in the case of non-circular arrays?
Thanks

Luigi