Reports until 13:27, Monday 24 October 2011
robert.schofield@LIGO.ORG - posted 13:27, Monday 24 October 2011 (1610)
Excess vibrations on the H2 PSL table from crystal cooling, and possible mitigation schemes
Summary. The crystal cooling circuit increases the PSL table motion an order of magnitude around 80-90 Hz and in some other bands. Mitigation schemes are discussed.

We measured vibrations on the PSL table shortly after the laser and auxiliary cooling systems were running. The accelerometer and data acquisition system were huddle tested against the LVEA seismometer and the accelerometer was mounted about a meter –X of the crystal box on the PSL table. 

Figure 1a and b show that the full cooling system increased table motion by an order of magnitude in some bands. 

We made several tests to see if a particular part of the system was the biggest problem. This turned out to be the case: when the crystal cooling circuit was bypassed or shut down it was difficult to distinguish the cooling-on and cooling-off spectra even though the blue auxiliary system and all parts of the red system, including the 30W laser system, were running except for the crystal cooling circuit. The culprit crystal cooling circuit is identified in the photo of Figure 2. Figure 3 shows that the excess noise disappears when the crystal circuit is bypassed but all other circuits are running. 

With regard to mitigation, Figure 4 shows that the vibration can be reduced by a factor of 2 by reducing cooling water circulation from 18 liters/minute to 15. However, this reduction may effect the performance of the laser. If a pressure-variation reduction system (pulse filters and hydropneumatic accumulators), such as were installed for the TCS laser cooling system, were installed near the input and output of the crystal cooling circuit, we might reduce vibrations from the purposefully generated turbulence in the crystal circuit. We might also have some success with structural damping: Figure 1a and b show that a sharp peak at about 83 Hz is evident in Y (the short axis of the table) and Z but not X. This would be expected for a rod-like resonator aligned along the X-axis. Candidates include the cable tray that supports the crystal cooling hose and the top stiffening bar of the crystal enclosure, both visible in Figure 2. 

Robert Schofield, Lutz Winkelmann

Non-image files attached to this report