I unearthed an old Levell oscillator in my kit, rigged it up through an amp at low output level as suggested and ran the tests with results listed below. I judged the speaker SPLs by ear, so some margin for error - to do a more rigorous test would need to get a dB level meter etc.
So.......first, confirming the input layout for these tests:
A B C
White Black Red
1. Using B3 as negative, +ve inputs to either A3 or C3 (in turn) operates the bass units only, at similar sound levels and frequency response, from low Hz to roll-off starting at approx. 3.3 kHz.
2. Using B2 as negative, +ve input to A2 operates the MR only, with roll-in to good level at 350 Hz and roll-off from 7 kHz.
3. Using B1 as negative, +ve input to A1 operates the HF only with roll-in to good level at 1 kHz and roll-off >10 KHz.
4. Continuing with B2 as negative, +ve input to C2 operates the MR only, now with roll-in to comparable level at more like 500 Hz and roll-off from 5 kHz, and
5. Using column B1 as negative, +ve input to C1 operates the HF but now with roll-in to comparable level at 2.6 KHz and roll-off from > 10 KHz.
To operate all speakers together requires a bridge to all column B inputs (ie. B1-B2-B3) as the common negative input, and bridge either column A (A1-A2-A3) as +ve, or
the alternate column C (C1-C2-C3) as +ve.
And from that I guess one could Bi-amp the bass (Row 3 only) and the (MR HF as a pair) with just the 1 and 2 rows bridged for the latter and providing a choice of crossover as column A or C alternate.
If correct, only a lonely Swiss could devise all that! Be interested in your thoughts.