Terra and I used the PI damping infrastructure to excite the butterfly and drumhead modes on EY, and then ring them down.
We excited the butterfly mode (6053.9 Hz) during a 50 W lock. The observed ringdown time was 23.5 minutes (= 1410 s), giving a Q of 27×106.
We excited the drumhead mode (8158.0 Hz) during at 2 W lock. The observed ringdown time was 13.5 minutes (= 810 s), giving a Q of 20×106.
The templates containing the spectrum data for these ringdowns live in my directory under Public/Templates/SUS/BodyModes.
In PI model:
MODE 29 = ETMX Drumhead
MODE 30 = ETMX Butterfly
MODE 31 = ETMY Butterfly
MODE 32 = ETMY Drumhead
The attached plot shows the expected ratio of the surface strain energy (in J/m) on the test mass face to the total strain energy (in J) in the test mass for the body modes between 5 kHz and 11 kHz. This is a simple Comsol model with a perfect silica cylinder.
Evidently, the drumhead and butterfly modes have similar energy ratios, so we should not expect their Q factors to be too different. It might be good to try the 9.2 kHz modes, since their energy ratio is rather different from the drumhead and butterfly modes, and they produce test mass strain in the beamline direction (the modes at 8.25 kHz and 9.4 kHz do not).