Cyclotron trim coil has discontinuous derivative

follow up from issue #110 (closed) as discussed there.

As can be seen in the figure the derivative of the B field is discontinuous in the middle of the trim coil. This comes by construction of the calculation.

I think it would be good to:

• find out where the calculation comes from (see also comment #110 (comment 2391))
• get some measured trim coil profiles to see how important the effect could be.

Task list added 26 March (from #157 (comment 5500)):

• PSI-trim coils polynomial fit (with minimal order) (@frey_m)
• OPAL trim coil polynomial input (@snuverink_j)
• Comparison with current implementation for TC15, make polynomial input default (@frey_m, @snuverink_j)
• Documentation
• Regression/Unit test