The SNEB bench is under vacuum since about 11 days, a time long enough to draw first conclusions.
All these boxes are equipped with pressure and temperature sensors. The challenge with these measurements is the coupling between pressure and temperature, due to basic physics, but also to the probe themselves and the possible long term drift of the probes.
The first figure is showing the temperature of the bench (top plot), which evolved during these 11 days, first with a large change due to the initial pump-down, then due to the change of the thermal load of the bench (with quadrant turned on for instance), and also to temperature variation of the building with day/night effect and also some slight increase toward the end.
The second figure is showing the data for the electronic container of the bench itself. The top plot is the raw pressure, which is correlated with the temperature (second plot). The bottom plot is an attempt to remove the effect of the temperature variation by rescaling the measured pressure according to the temperature. The temperature used is not Kelvin, which would produce an overcorrection but Celsius plus an offset of 200. This offset value is empirically adjusted to minimize the short-term variations like the initial pump-down and the day/night effect. To give a sense of this sensitivity to the offset, the red and blue curves are showing the corrected data with the temperature offset changed by plus or minus 20. From these data, there is no evidence of leak. One can say that the pressure has not been reduced by more than 1 mbar over 10 days (forgetting the first day with large transient).
The third and forth figures are showing the data for the two photodiode air boxes. The temperature offset used for the model is different, because the type of probe is not the same as the one for the electronic air tank of the bench. But again, if there is a leak, its rate is below 1 mbar/10 days.
Figure 5 is showing the data for camera 1 (camera 2 had a big leak and is turned off). Since the temperature probe of this air box is not working, the model is using the temperature of a photodiode box. The result is again that the leak rate is not exceeding 1 mbar/10 days.
Finally, figure 6 and 7 are showing the data for the Nikhef quadrant air boxes. The time spam is reduced here, because the data were not collected from the beginning. The limited resolution of the temperature sensor makes also the temperature correction more difficult. QPD2 is the only case where there could be a hint of leak (not confirmed), but at a rate lower than 1 mbar/5 days.
For all the air boxes present in SNEB (except camera 2) the worse possible leak rate is 1 mbar/5 days for QPD2. This translates to 72 mbar after one year of operation, an acceptable value, meaning that we should be safe for the expected O2 data taking. More data will allow us to better quantify the leak rate in the future.
