There are different ways to specify the harmonics number for PSS simulation, 2 of them are "number of harmonics" and "Array of coefficients".
My question is, if this choice may effect the results accuracy. Or it will only effect on post analysis (Fourier transfer)

(test case) If I know a circuit will only have strong output at odd harmonics, and 9 harmonic is good enough for the simulation. What is the difference for me to set the "number of harmonics" to be 9 (harms=9), compared with "Array of coefficients"  to be [1 3 5 7 9] (harmsvec=[1 3 5 7 9])
(1) because the fundamental frequency and the highest harmonic is the same, the "period" and "maxstep" for 2 simulation is the same.
(2) If the difference of this 2 selection will only effect on post analysis, the results value at harmonic 1,3,5,7,9 should be the same
(3) chose  harms=9, the results may have small value at harmonic 2,4,6,8 from the "software noise", but if we chose harmsvec=[1 3 5 7 9] the results will not have even harmonic because the post analysis don't do Fourier transfer at those tones.

If that true?

 Simon