craig.cahillane@LIGO.ORG - posted 17:38, Tuesday 31 January 2023 - last comment - 10:29, Wednesday 01 February 2023(67142)
awg injections are a factor of 100 lower than you might expect
Kevin, Craig when I ask foton to interpret a 'gain(1e-5)' filter, it gives me a line at 1e-5, as one might expect. if I try to inject this via DTT, awggui, or python-awg, it gives me the below result: a line at 1e-7. we checked the gain and amplitude of the injections, they were both one. there may be some amplitude scaling that I don't understand happening in the backend of awg here. I don't know when this became a problem, but I imagine it's been a long time. we probably should not repair it, otherwise all our .xml templates will be a factor of 100 stronger. just something to note when we are excitation smithing.
Images attached to this report
Comments related to this report
I would expect that you specify the rms amplitude. With 10kHz bandwidth, the spetcral density would then be 100 times lower.
Tested Daniel's comment, he's right (again). I injected 1e-5 into the PSL_SECONDLOOP_EXC (while the output was off, so no real injection into the IFO). The difference is our normal LSC EXC have a sampling rate of 16384 Hz, while SECONDLOOP EXC is 32768 Hz. Injecting 'gain(1e-5)' brought us down a factor of two. This will be important when excitation smithing for ASC, which has a sampling freq of 2048 Hz.
Images attached to this comment
