At some point we got an errata report for Exercise 5.6 (
exercises\dynamics\powergrid.tex) saying that "we need to shift $\varphi_0$ in part (c)" (which defines a curve for the saddle). We can't find the original errata submission, but we should make sure that the equation is correct. The solutions show that the equations of motion are satisfied by given curve, but don't verify that the curve actually goes through the saddle point.
At a minimum, we should update the solutions to verify that the curve goes through the saddle point.