[Alexa and Kiwamu]
We became unable to lock the reference cavity this morning. The autolocker wasn't capturing a fringe. After some investigation and fiddling some parameters it started locking.
There was an oscillation which seemed preventing the FSS from a stable lock. In the end the oscillation somehow went away.
-
What we did:
-
We changed the temperature scanning range and its sweep speed
-
Neither narrowing nor expanding the range helped
-
We decreased the transmitted light threshold from 1 to 0.5 with a hope that this gives an earlier trigger for a fringe
-
We increased the oscillation threshold from 0.5 to 2.0 because we saw the fast feedback hits this threshold.
-
We switched the system to a manual mode by toggling the button for the autolocker in the FSS medm screen
-
We manually set the temperature to be 0.1359 [K]. This is where we found the zero-zero mode.
-
We diabled the active temperature feedback by disabling the input of H1:PSL-FSS_TEMP_LOOP
-
Once we reached the point close to the resonance we toggled on the fast loop.
-
The FSS was found to be oscillating.
-
We were still in the manual mode at this point
-
In H1:PSL-FSS_FAST_MON_OUTPUT display we could see that the voltage fully swung +/- 10 V. We didn't check the oscillation frequency.
-
We toggled the noise eater (see Rick's alog).
-
No difference was observed.
-
We decreased the common gain by 19 dB to see if the oscillation is due to a control issue.
-
The oscillation disappeared.
-
H1:PSL-FSS_COMMON_GAIN was 8 dB
-
We set the common gain back to the nominal
-
Strangely we didn't see the oscillation any more
-
H1:PSL-FSS_COMMON_GAIN was 27 dB which is nominal
-
So the oscillation was fixed somehow
-
We switched it to the autolocker mode and restored the parameters which we changed.
-
It has been locking and fine so far.
After the oscillation went away I temporarily increased the common gain to 30 dB which is the maximum. The oscillation didn't happen. I set it back to 27 dB.