GitLab

Higher-order modes aren't being calculated properly for signals. physics/m_fsig.kat from the test repo shows the issue:

Code

l l1 1 0 n1
gauss g1 l1 n1 10m 0 

s s1 1 n1 n2
m m1 0.9 0.1 0 n2 n3
s s2 1 n3 n4

fsig sig1 m1 1 0

xaxis m1 phi lin 0 180 100

ad ad_00_c 0 0 0 n2
ad ad_02_c 0 2 0 n2
ad ad_20_c 2 0 0 n2

ad ad_00_u 0 0 1 n2
ad ad_00_l 0 0 -1 n2
ad ad_02_u 0 2 1 n2
ad ad_02_l 0 2 -1 n2
ad ad_20_u 2 0 1 n2
ad ad_20_l 2 0 -1 n2

ad adr_00_c 0 0 0 n3
ad adr_02_c 0 2 0 n3
ad adr_20_c 2 0 0 n3

ad adr_00_u 0 0 1 n3
ad adr_00_l 0 0 -1 n3
ad adr_02_u 0 2 1 n3
ad adr_02_l 0 2 -1 n3
ad adr_20_u 2 0 1 n3
ad adr_20_l 2 0 -1 n3

maxtem 2

yaxis abs:deg

gnuterm no

Running this gives:

Output

Reference Array		Output Array
[[ 9.00000000e+01  9.00000000e+01]
 [ 9.48677857e-01  9.48677857e-01]
 [ 1.80000000e+02  1.80000000e+02]
 [ 2.27191998e-03  2.27191998e-03]
 [ 8.98059504e+01  8.98059504e+01]
 [ 2.27191998e-03  2.27191998e-03]
 [ 8.98059504e+01  8.98059504e+01]
 [ 9.48677857e-01  9.48683298e-01]
 [-9.00000000e+01 -9.00000000e+01]
 [ 9.48677857e-01  9.48683298e-01]
 [-9.00000000e+01 -9.00000000e+01]
 [ 2.27191998e-03  0.00000000e+00]
 [ 1.79805950e+02  0.00000000e+00]
 [ 2.27191998e-03  0.00000000e+00]
 [ 1.79805950e+02  0.00000000e+00]
 [ 2.27191998e-03  0.00000000e+00]
 [ 1.79805950e+02  0.00000000e+00]
 [ 2.27191998e-03  0.00000000e+00]
 [ 1.79805950e+02  0.00000000e+00]
 [ 3.16227766e-01  3.16227766e-01]
 [ 9.00000000e+01  9.00000000e+01]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]
 [ 0.00000000e+00  0.00000000e+00]]

The various detectors looking at HOM signal sidebands at n2 give zero output. This is because the mirror Knm matrices aren't being filled for the signal simulation - they're all just identity matrices. After discussing with Sam, it seems we need some way for the signal simulation to access the Knm matrices from the carrier simulation (to avoid recalculating them)