At about 23:40 LT the turbomolecular pump of the WE tower went in failure and automatically shut down. Riccardo went on-site for a visual inspection, and found that the issue is likely due to the pump driver electronics.
The pump driver has been disconnected from the mains and then we switched on the spare turbo-pump. The system is now restored in normal conditions.
On Tuesday we will replace the pump electronics and investigate the cause of the failure.
During the maintenance break of this morning we replaced the electronics of the recently failed WE pump and then performed some functional tests: the malfunction appears due to the pump mechanical hardware.
The failed pump needs to be completely replaced, this is planned for the next maintenance shift. The tower remains pumped with the spare pump , under normal vacuum conditions .
Looking at the spectrograms of the microphone and velocimeter (Guralp on the tower floor), we noticed that the spare turbo pump does not produce excess acoustic and seismic noise.
Browsing backwards VIM spectrograms, it is evident that the WE turbo pump started misbehaving (extra noise) on November 28. This is very evident in the ENV_WEB_SEIS sensor (Figure 1).
On the same day, correlated effects are also seen in B8: prompt enhancement of some lines at 800Hz, and more in the 2kHz range (Figure 2 and 3). At the same time some noise started rising in a wider region 100-150Hz, but only on Nov 30 the 120Hz bump shaped clearly in B8 (and Hrec) (Figure 4).
The best monitor of that noise currently in te DMS is accelerometer ENV_WE_CT_ACC_Z. Yet, the threshold was set too high (Figure 5), and shall be adjusted. It is alsoadvisable to add the ENV_WE_ACC_Z accelerometer in the DMS monitor.
