Sorry, I don’t have an answer. What about the sd drop in the cell 2 and cell 3 ?
What are the actual values of sd drops before and after ? (rate and number of drops, please)
As a workaround, have you tried resetting the TRX and the TCU ?
My first thought is regarding the TCU and RSL capacity : perhaps there is too much signalling now, leading to drops.
You could check the LapD congestion (thanks to counter L1.18, counter type 7)
Also, to isolate the problem : try removing the extra SDCCH on cells 2 & 3. See if it reverts to the initial values.