2014 November 7 VEGAS testing
Testing the new OGP calibration method to compare against the current (production) method
Same method as in the previous tests by Adam, Paul and Jason
. Using VEGAS mode 2.
Data in TGBT14A
_912_45, using Astrid script TGBT14A
- Scan 2 was a quick 60 second test to check we were switching the noise source successfully
- Scan 3 - 600 second scan, with no OGP calibration
- Scan 5 - 600 second scan, with production OGP
- Scan 6 - 600 seconds, new OGP, OG values derived from external noise source
- Scan 7 - 600 seconds, new OGP, OG values derived from IF noise source
- Scan 8 - 600 seconds, random OG number used
- Scan 9 - 600 seconds, zero OG numbers used
Data in scans 6 and 7 were compared directly by dividing through as follows in
GBTIDL -> gettp,6
GBTIDL -> copy,0,6
GBTIDL -> gettp,7
GBTIDL -> copy,0,7
GBTIDL -> divide,6,7,8 # saves the output from divide in buffer 8
GBTIDL -> show,8
All total power plots use same y axis range for direct comparison.
Total power for scan 3 (no OGP applied):
Total power for scan 5 (production OGP applied):
Total power for scan 6 (OGP applied, OG from external noise source):
Total power for scan 6 (OGP applied, OG from IF noise source):
Total power from scan 6 / total power from scan 7 for direct comparison:
Using random OG values in the OGP calibration (scan 8):
Using no OG values in the OGP calibration (i.e. phase only) (scan 9):
By comparing scans 6 or 7 to scan 5, it seems that the new OGs as calculated by the new method do not seem to produce better results then our current production method.
The comparison of scans 6 and 7 show that there does not seem to be a major difference between using the new method with an external noise source or the one in the IF.
Scan 8 shows that both methods for determining the OGs are better then generating random numbers for the offsets and gains.
Comparing Scan 3 to Scan 9 shows the affect of just
applying phase corrections (the offsets and gains are all zero in both scans).