Krunal proved test result of four modules and Dinko analised them.
M1630: Grade A, VCal calibration: Slope=43.5e-/Vcal, Offset=-145.4e-
M1632: Grade A, VCal calibration: Slope=45.4e-/Vcal, Offset=-290.8e-
M1636: Grade A, VCal calibration: Slope=45.9e-/Vcal, Offset=-255.2e-
M1638: Grade A, VCal calibration: Slope=43.3e-/Vcal, Offset=-183.1e-
A few comments:
1) Rates. One should distinguish X-rays rate and the hit rate seen/measured by a ROC (as correctly Maren mentioned).
X-rays rate vs tube current has been calibrated and the histogramm titles roughly reflect the X-rays rate. One could notice that
number of hits per pixel, again roughly, scaled with the X-rays rate (histo title)
2) M1638 ROC7 and ROC10 show that we see new pixel failures that were not observed in cold box tests. In this case it's not critical, since only
65/25 pixels are not responcive already at lowest rate. But we may have cases with more not responcive pixels.
3) M1638 ROC0: number of defects in cold box test is 3 but with Xrays in the summary table it's only 1. At the same time if one looks at ROC0
summary page in all Efficiency Maps and even in Hit Maps one could see 3 not responcive pixels. We should check in MoreWeb why it's so.
4) It's not critical but it would be good to understand why "col uniformity ratio" histogramm is not filled properly. This check has been introduced
to identify cases when a column performace degrades with hit rate.
5) PROCV4 is not so noisy as PROCV2, but nevertheless I think we should introduce a proper cut on a pixel noise value and activate grading on
the total number of noisy pixels in a ROC (in MoreWeb). For a given threshold and acceptable noise rate one can calculate, pixels with noise
above which level should be counted as defective. |