TGBT14A_912_76 - Continued bank discrepancy tests

NB: Pol 1 becomes unstable half-way through this session

NB: Scan 38 and onwards were all done with VEGAS switching between two buffers, even though some were supposed to be total power / nocal. But the cal diode was turned on and off as described below.

Goals

  • stick with mode 4, BOF 123
  • Run Mode 1 and Mode 4 next to each other for comparison purposes
  • since we believe we have established linearity of the FPGA, I don’t think I need to run through a range of coverter rack attenuator settings. I’ll just use the nominal -20dBm input power. But, I will set every bank / polarization to have the same lbw_gain setting.
  • probably switch to another receiver (probably L-band), which has a lower system temperature, and a high cal diode setting.
  • for nominal inputs, run with the blinking cal, and do a few scans allowing GBTIDL to calculate the system temperature.
  • for nominal inputs, do separate “cal on”, and “cal off” measurements, with the data all going to a single buffer. Calculate Tsys manually.
  • for nominal inputs, drive the cal with DCR while running VEGAS in tp_nocal mode
  • repeat all of the above with as high as possible safe input power, to try and minimize effect of spikes
  • repeat all of the above, with MMCM and OGP calibration turned off, so the center spike should be intentionally exaggerated.

Details

  • 8:40 start setting up
  • 8:45 get Ray to switch to 123 BOF
  • Various Turtle-related faults, finally get started 9:10pm
  • Telescope is at snow dump
  • Weather is absolutely lousy; pouring with rain
  • Above two mean Tsys's are very high.

Mode 1 - normal

  • Run Archivist
  • Run bank_gain_mode1_lband - switch mode = tp
  • Run balance() - target level -20dBm
  • lbw_gain is not relevant in this mode
  • IF Rack output power is too high. Set attenuators by hand, then balance, all seems ok now. ??
  • Cal diode switching is really obvious in Vegasdm
  • Scans 1-3 one minute Tp (with cal) scans. Some strong RFI visible in Vegasdm. Tsys 46K, Tcal 20K
  • results look reasonably good in pol 1, bur obvious RFI present. try switching to X-band
  • X-band - cal switching again clearly visible, no RFI now.
  • Scans 4-6 one minute Tp (with cal) scans. Tsys 135K, Tcal 17K
  • Try C-band
  • Scans 7-9 one minute Tp (with cal) scans. Tsys 86, Tcal 56, no RFI
  • Tsys's look reasonably consistent, although with at least one outlier (scan 7, plnum=0,ifnum=3)
  • Tsys's repeatable from bank to bank; go with this.
  • Switch to Tp nocal, toggle cal diode manually.
  • Scan 10, cal off
  • Scan 11, cal on
  • Scan 12, cal off
  • Scan 13, cal on
  • Scan 14, cal off

Mode 4 - normal

  • Run bank_gain_mode4_cband (with cal)
  • Run balance
  • Scans 15-17 one min Tp (with cal) scans, Tsys 86K, Tcal 57K, no RFI
  • didn't set lbw_gains to be the same between banks.
  • values 104256245, 51954335, 49468797, 99432075 - 2**25.6 - 2**26.6
  • set lbw_gain to 2**26 = 67108864
  • Scans 18-20 one min Tp (with cal) scans, lbw_gain fixed.
  • Scan 18, pol 0, Tsys = 80 - 90K, pol 1, all 85K !!!
  • Scans 21-25 Cal off,on,off,on,off

Mode 1 - High input power

  • bank_gain_mode1_cband (with cal)
  • balance to -12dB - right on the edge of saturating. Bank F pol 0 is at -14dB.
  • Scans 26-28 one min Tp (with cal) scans.
  • Scans 29-33 Cal off,on,off,on,off. Sky is varying a lot during scan 29

Mode 4 - High input power

  • pol 1 total power becomes quite unstable during this set, for reasons I cannot explain
  • bank_gain_mode4_cband
  • balance to -12dB - can't quite get there (range from -20 to -12dB)
  • Scans 34 - forgot lbw_gain, set this now
  • Scans 35 - 37 one min Tp (with cal) scans
  • Scans 38- cal off. Suddenly lost power in pol 1, for no obvious reason. pol 1 is now drifting around...
  • I do not understand what has happened here. I need to remove 10dB of attenuation from IF Rack...
  • This seems to have sorted everything out.
  • Run SetValues to set lbw_gain
  • pol 1 is still drifting around at this point...
  • Scans 39-43 Cal off,on,off,on,off
  • pol 0 was quite stable, pol 1 drifting around. But all four banks should still give the same Tsys I think (different in each pol).

Mode 1 - ADC calibration turned off.

  • edit vegas.conf to set load_mmcms and load_ogps_inls to false.
  • turn managers off, and switch modes, to make sure this takes effect.
  • I hope this has worked - how can I tell?
  • bank_gain_mode1_cband
  • balance to -20dBm
  • pol 1 totally unstable at this point..., input powers -35dBm pol 0 looks ok.
  • Scans 44-46 one min Tp (with cal) scans
  • Scans 47-51 Cal off, on, off, on, off

Mode 4 - ADC calibration turned off.

  • bank_gain_mode4_cband
  • balance -20dBm
  • set lbw_gain
  • Scans 52-54 one min Tp (with cal) scans
  • Scans 55-59 cal off,on,off,on,off

Clean up

  • Edit vegas.conf to turn ADC calibration back on, turn managers off and on, and reconfigure banks
  • 12:10am - finito

Preliminary Analysis with GBTIDL

TP with Cal scans

IDL gets the following values for these scans. NB the filler has interpolated over the center spike
   S  P M   Tcal       E        F         G        H        Mean      Std 
   7  0  1 56.018   86.5205   87.5613   87.1866   84.3013   86.3924   1.4590
   7  1  1 57.565   85.6456   84.3732   84.7545   85.4466   85.0550   0.5937
   8  0  1 56.018   86.3808   87.4134   87.0519   84.1836   86.2574   1.4472
   8  1  1 57.565   85.8220   84.5460   84.9306   85.6254   85.2310   0.5956
   9  0  1 56.018   89.1589   90.3008   89.8667   86.7209   89.0118   1.5982
   9  1  1 57.565   88.0779   86.7147   87.1876   87.8790   87.4648   0.6290
  18  0  4 57.729   85.0691   79.9959   85.7285   90.7111   85.3762   4.3828
  18  1  4 60.443   85.1313   85.8918   85.7464   85.2133   85.4957   0.3796
  19  0  4 57.729   91.6595   85.9521   92.3377   97.6913   91.9102   4.8016
  19  1  4 60.443   91.1000   91.9543   91.8252   91.2071   91.5217   0.4305
  20  0  4 57.729   92.1261   86.3353   92.7774   98.0154   92.3135   4.7784
  20  1  4 60.443   91.0381   91.8859   91.7561   91.1364   91.4541   0.4288
  26  0  1 56.018   90.9275   91.6890   90.8776   87.0634   90.1394   2.0840
  26  1  1 57.565   89.0507   88.7834   89.0541   90.3633   89.3129   0.7117
  27  0  1 56.018   87.7094   88.3397   87.7104   84.2344   86.9985   1.8665
  27  1  1 57.565   85.8594   85.6090   85.8438   86.9902   86.0756   0.6204
  28  0  1 56.018   87.1070   87.7328   87.1209   83.6872   86.4119   1.8398
  28  1  1 57.565   85.5913   85.3448   85.5769   86.7158   85.8072   0.6162
  35  0  4 57.729   81.5255   78.2200   83.7297   85.9117   82.3467   3.2826
  35  1  4 60.443   82.7898   82.1361   85.2347   83.9910   83.5379   1.3674
  36  0  4 57.729   80.8354   77.2655   82.1949   85.2287   81.3811   3.3016
  36  1  4 60.443   82.2370   81.5544   84.6105   83.3938   82.9489   1.3429
  37  0  4 57.729   78.7397   75.2369   79.8745   82.8400   79.1728   3.1422
  37  1  4 60.443   79.9816   79.3208   82.0924   81.0178   80.6031   1.2139
  44  0  1 56.018  595.1307  361.3974  506.5335  439.4442  475.6265  99.3209
  44  1  1 57.565   78.7641   78.1431   78.6468   78.6067   78.5401   0.2730
  45  0  1 56.018 1922.5547  592.1356 1790.0970  630.0169 1233.7012 721.1418
  45  1  1 57.565   78.5727   77.9551   78.4545   78.4141   78.3491   0.2711
  46  0  1 56.018 2418.8728  586.3088 4819.8232  620.2552 2111.31491998.2919
  46  1  1 57.565   78.3475   77.7351   78.2277   78.1891   78.1248   0.2684
  52  0  4 57.729  210.7865   77.5585  182.0076  143.3659  153.4296  57.6317
  52  1  4 60.443   82.3454   81.9368   82.3393   82.3054   82.2317   0.1974
  53  0  4 57.729  199.7456   76.3148  171.7908  145.2282  148.2699  52.8828
  53  1  4 60.443   81.5209   81.1188   81.5098   81.4832   81.4082   0.1936
  54  0  4 57.729  183.8923   74.9462  156.8034  141.0646  139.1766  46.3296
  54  1  4 60.443   81.1743   80.7761   81.1638   81.1393   81.0634   0.1921

-- RichardPrestage - 2015-03-04
Topic revision: r4 - 2015-03-05, RichardPrestage
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback