In case of a synchronous hand-over related channel activation, 3GPP TS 48.058 Section 4.1.4 is very specific:

BTS starts transmission immediately on the main channel in the indicated mode and with encryption if so indicated. If
the MS Power and Timing Advance element are present, BTS shall start transmission also on SACCH with the timing
advance and MS power control parameters indicated. If only the MS power element is present the BTS may start
transmission also on the SACCH.
When receiving a correct access burst with the correct handover reference, BTS starts the normal reception process on
the main channel in the indicated mode, with deciphering applied if so indicated, and starts receiving (and sending if not
started earlier) on SACCH. The handover detection procedure towards BSC is also started. Alternatively, the reception
of a correctly decoded frame from the MS on the main channel, in the indicated mode and deciphering applied if so
indicated, allows the start of sending on SACCH (if not already started) and starts the handover detection procedure
towards the BSC.

However, as uncovered by an upcoming BTS_Tests.TC_sacch_chan_act_ho_sync test case, we appear to be activating the SACCH unconditionally from the first moment.

The problem here is quite obvious: Until we have received the access burst from the MS, we don't yet know the timing offset, and hence the timing advance that we should advertise in the downlink SACCH. If we start SACCH transmission too early, it means that a wrong TA is advertised, which may be picked up by the MS, which will then apply a wrong TA value -> boom.