GBT14A _912_32


  • ADC tests for CASPER Group


In a previous test, we tested looking for spikes in Q-band while connected to a noise source and looking at blank sky. Spikes were not seen in the power spectrum from the raw ADC output. This was different than the results obtained by David and Dan at Berkeley who found the spikes in the raw ADC output when connected to a noise source (adc_snap_spec_2pol.png).

Setup and Analysis

To further test this comparison, we set up 4 Banks of VEGAS, terminating one polarization of bank A, disconnecting one polarization of bank B and running banks C and D as normal. A noise 500MHz wide noise source was inserted at the beginning of the IF chain to mimic the setup used by Dan and David. In bank A, LO2 was manually adjusted to offset the noise source to the lower side of the bandpass (in bin-space), as shown in the plots in the next section. The offset was not applied to the data from the other banks.

Separate 20 minutes scans were taken in L-band with all 4 banks in modes 1, 2, and finally mode 3 (scans 214, 215, and 218). We then summed the FFT of all 1158 snapshots before creating the power spectrums.

Spikes were clearly visible without a high pass filter in the .fits files recorded by VEGAS (as shown here). We then looked at the data sniffed from before the ADCs, and found the spikes there as well, as shown in Bank A pol 1 mode 1, Bank A pol 2 mode 1, Bank B pol 1 mode 1, and Bank B pol 2 mode 1. The spikes showed up in all banks, in all modes. As can be seen comparing Bank A and Bank B, the spikes appear to be larger in the disconnected bank (Bank B). The spikes from the ADCs appear to be in the same location as the spikes in VEGAS, as shown in the comparison plot (similar in all data).

Due to this discrepancy, we reanalyzed the data from earlier, and still found no notable spikes.


  • Bank A is connected / terminated
  • Bank B is connected / disconnected
  • Bank C is connected / connected
  • Bank D is connected / connected
  • Scan 205 - 1 minute test scan - TP NOCAL - Modes 1 through 4. Make them all Mode 1 initially
  • Scan 206 - all mode 1, test scan
  • Adam confirms that he can read the data using IDL FITS routines
  • Scan 207 - reconfigured IF/LO, but did it for L-band, not Q-band. Reconfigure for Q-band.
  • Scan 208 - 213 - tests of different passbands, etc.
  • Scan 214 - 20 min scan - mode 1
  • IF Rack = Noise Source (1-8 GHz), filter 2360-2640, Attn 20dB, Converter Rack LO2 13310MHz.
  • Scan 215 - mode 2, with modified IF system setup (20 min)
  • Scan 216 - 217 are test runs in mode 3 (not useful)
  • Scan 218 - mode 3, with modified IF system setup (20 min)
Topic attachments
I Attachment Action Size Date Who Comment
BankApol1mode1.pngpng BankApol1mode1.png manage 33 K 2014-09-12 - 18:34 AdamKobelski  
BankBpol1mode1.pngpng BankBpol1mode1.png manage 34 K 2014-09-12 - 18:34 AdamKobelski  
BankBpol2mode1.pngpng BankBpol2mode1.png manage 31 K 2014-09-12 - 18:34 AdamKobelski  
adc_snap_spec_2pol.pngpng adc_snap_spec_2pol.png manage 326 K 2014-09-12 - 18:34 AdamKobelski  
bankApol2mode1.pngpng bankApol2mode1.png manage 31 K 2014-09-12 - 18:35 AdamKobelski  
cmp_bankApol1mode2.pngpng cmp_bankApol1mode2.png manage 38 K 2014-09-15 - 16:04 AdamKobelski comparison of ADC and VEGAS spectra
Topic revision: r8 - 2015-01-27, 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