TGBT14A_912_120 - Testing New Analog IF box from Galen (inserted into Bank D)
Overview
A new analog filter block was inserted and hooked up to Bank D. Tonight I will run some tests of it. The telescope is in overnight shutdown mode, so I will be running these tests without movement. NOTE: Bank A is hooked up to the noise source...
Tests
Start in L band to see how Hydrogen looks
- Scan 37 - frequency switching the 1420 line in banks A,C,D,E, and G. - BAD configuration
- Scan 38 - frequency switching the 1420 line 2 second integration and switching period, 10 minute scan - Cursory analysis looks Good!
- Scan 39 - as in 38, but using the hi noise diode, looks good!
- Scan 40 - tp, L-band centered at 1.5 GHz, 10 minute scan, 10s integration, hi cal
- Scan 41 - same as 40, repeated to use as an on/off pair with scan 40
- Individual scans look fine, though the scan pair does not align well, possibly an RFI issue?
- Scan 42 - tp_nocal, Lband centered at 1.5 GHz, 10 minute scan, 10s integration no cal (balance on scan 42 - diode off)
- Scan 43 - tp_nocal, Lband centered at 1.5 GHz, 10 minute scan, 10s integration high cal on (balance on scan 42 - diode off)
- all LL data seems just fine but super noisy (RFI), the RR data differs between the banks, with C, E, and G agreeing, and D showing a much lower value. The sigref of band D is more accurate, showing the appropriate negative value, nevertheless, the L band data is fairly worthless, too much RFI.
Move to X band, as the L band data is terrible for wide bandwidths (Donna is back in control room so a change is easy)
Check Linearity:
- vegas_test_xband_mode02_tp_hical_slow_bal_low
- Scan 44 - tp_nocal, Xband centered at 8.4 GHz, 10 minute scan, 10s integration no cal (balance on scan 44 - diode off)
- Scan 45 - tp_nocal, Xband centered at 8.4 GHz, 10 minute scan, 10s integration high cal on (balance on scan 44 - diode off)
- getsigref,45,44 looks great for all banks. For pol 0 (LL), all line up very well. For pol 1 (RR), banks C and E line up, as do D and G (likely minor CM issues, as is commonly seen in other sources when further calibration is not applied.
- vegas_test_xband_mode02_tp_hical_slow_bal_high
- Scan 46 - tp_nocal, Xband centered at 8.4 GHz, 10 minute scan, 10s integration no cal (balance on scan 46 - diode on) VEGAS @~-18dB
- Scan 47 - tp_nocal, Xband centered at 8.4 GHz, 10 minute scan, 10s integration high cal on (balance on scan 46 - diode on) VEGAS@~-20.5dB
- getsigref,46,47 looks good for all banks. The filter roll off does not seem as great for Bank D. For pol1 (RR), D and G are different, with a possible deviation around 8.24 GHz (see newIF_Xband_slow.pdf).
Check Stability:
- Scan 48 - vegas_test_xband_mode02_tp_hical_1sint_3min - check short time scale stability - Looks OK
- Tsys Variation Across the band:
-
- See here to see that this is OK..
- Scan 49 - vegas_test_xband_mode02_tp_hical_30sint_60min - check longer scale stability - Looks OK
- Scan 50 - vegas_test_xband_mode02_tp_local_30sint_60min - longer scale stability - Looks OK
- Scan 51