The example attached features an fccee lattice that twisses properly from the IP (with tapering), but when cycling to the RF section for tracking, the twiss fails, likely due to the large momentum offset at that location. The option to use a 'co_guess' at the new location seems to work, however the option 'co_search_at' the IP (which should be equivalent) fails.
sample_cycle.py
sample_cycle.json