TGBT14A_912_83 Bank to Bank problems with real hardware, real software

Goals

  • Total power and spectrum ratios ADC / filtersnap / HPC spectra
  • Identical ratios for varying input power
  • Do this with noise source on inputs to ADCs, inputs to VEGAS IF Modules, and with real (converter rack) signals, with these Teed off to DCR.
  • Using the new noise source, 300MHz highpass filter, amplifier, attenuator and 8-way splitter

Details

  • Set up at 8:00am, Jason connects noise source to ADCs
  • Set up lbw_gain. From Randy, empirically the digital filter requires a gain of 5.664 to get -20dBFS into each of real and imag separately, for a -20dBFS input.
  • So, set lbw_gain to 2**16 (unity in BOF) * 2**8 (Ray correction for old config_tool)*5.664
  • Output power for filtersnap will be 3dB above input power, since real and imag powers are summed together.

Mode 4 - noise source to ADC input

  • bank_gain_mode4_cband_nocal
  • archivist
  • setValues with 5.664 * 2**24
  • Attenuator setting of 0dB corresponds to ~ -15.8dBFS. filtersnap goes to around -12.5 dBFS - close enough
  • Change attenuator by 8dB seems to change ADC value by 8dB also - is this what I saw before? Start from attenuation of 0dB, and set in powers of 3dB.
  • Run 60 second tracks
  • Scans 1-6 = 0,3,6,9,12,15 dB attenuation
  • quick exploration with GBTIDL; everything seems to be working perfectly!

Mode 1 - noise source to ADC input

  • bank_gain_mode1_cband_nocal
  • archivist
  • SetValues is irrelevant
  • Scans 7-12 = 0,3,6,9,12,15 dB attenuation
  • Looks sensible, but slightly off, e.g. ratio between 7 and 9 looks like 3.5 (should be 3.98) = 5.44dB iso 6dB

Mode 4 - noise source to VEGAS Analog IF input, and VEGAS IF monitor output to DCR

  • Need to remove the amplifier from the noise source, as the VEGAS IF module has gain. Nominal -20dBFS corresponds to 18dB on the attenuator
  • bank_gain_mode4_cband_nocal
  • archivist
  • setValues with 5.664 * 2**24
  • Manually set up the Analog Filter Rack and the DCR to send the VEGAS IF module monitor output to the DCR
  • On Analog Filter Rack, set 1.6 GHz samplers to have SGfilters set to "external". This is where Galen will hook in.
  • Can monitor at All Samplers / SG Power
  • Outputs go to DCR A9 - A16
  • Do a test scan to see whether DCR samplers will start toggling. Add DCR to Scan Coordinator
  • Note, Bank H pol 2 (J16) seems to have a dodgy connector out of VEGAS IF module to ADCs; value was jumping around until Galen poked at it.
  • Scan 13 60 sec track. No sign of DCR values updating..
  • DCR FITS file has no rows...
  • Try a quick pointing scan - scan 14
  • Turns out it seems like DCR has to be the switching signal master, even for no switching
  • Do all of the above setup. Set DCR to 1 sec integratgions, DCR as master, no switching. Include in scan coordinator
  • attenuator at 18 dB gives ~ -20dBFS. Ratio of ADC to filtersnap is now around 6dB (filternsap = -14dB) - presumably because of different passband
or something? I don't really understand this...
  • Scan 15 12 dB attenuation. DCR is now around 22 counts. DCR FITS file is around 4.39e+06, VEGAS is around 1.5e+11
  • Scan 16 15dB attenuation. DCR is now around 12 counts. DCR FITS file is around, VEGAS is around 2.17e+06 0.75e+11
  • Scan 17,18,19,20 = 18,21,24,27dB attenuation
  • switch to mode 1

Mode 1 - noise source to VEGAS Analog IF input, and VEGAS IF monitor output to DCR

  • bank_gain_mode1_cband_nocal
  • archivist
  • SetValues is irrelevant
  • re-include DCR
  • Scans 21-26, attenuations 12 - 27dB

Mode 1 - C-band - receiver IF to VEGAS with VEGAS IF module monitor output to DCR

  • bank_gain_mode1_cband_nocal
  • archivist
  • balance - to get IF input levels around -20 dBFS
  • SetValues is now irrelevant
  • Not getting the values I expect in the IF Manager - for example, IF Manager is now showing 800MHz filter.... resend and everything is ok.
  • Suspect CLEO IF Manager may not be updating somehow?
  • Balance to -15dB
  • DCR now showing ~ 26V
  • Scan 27 -15dB into VEGAS ADCs. But DCR didn't read any data... Somehow SSS got switched back to VEGAS!
  • Scan 28 - 15dB
  • Scan 29 - 18dB - approx factor of two difference from 28, not exact as converter rack attenuation change is not exact
  • Scans 30-32, -21,-24,-27 dB
  • Scan 32, only 4 of 8 channels seemed to balance properly... Balance manually
  • Scan 33 -27dB

Mode 4 - C-band receiver IF to VEGAS with VEGAS IF module monitor output to DCR

  • Config tool puts in a 320MHz filter, so system is being overdriven to get ADC to -20dBFS. Need a smaller lbw_gain.
  • By experiment, need a value of 2 * sqrt(2) smaller - don't really understand this value.
  • Scan 34 -15dB into VEGAS ADCs.
  • VEGAS Bank H aborted due to invalid data..., twice...Turn Bank H off and on again. Doesn't cure it. Call Ray. Now it is worse... then cures itself
  • Scan 39 -15dB - DCR got no data, not SS Master
  • Scan 40 -15dB Now we are in business.
  • Scans 41 - 44 -18, -21, -24, -27 dB. Ratio of scans is huge! What on earth is going on here....?

Mode 1 - X-band receiver IF to VEGAS with VEGAS IF module monitor output to DCR

  • bank_gain_mode1_xband_nocal
  • Archivist
  • can't balance to -15dB again in one input, but proceed anyway.
  • SetValues is irrelevant
  • Scans 45-49: -15 - -27 dB

Mode 4 - X-band receiver IF to VEGAS with VEGAS IF module monitor output to DCR

  • bank_gain_mode4_xband_nocal
  • archivist
  • balance to -15dB fails, -18dB is close; start from here.
  • need same value of lbw_gain as for C-band (not too surprising)
  • Scan 50-53, -18 - -27dB

Mode 1 - X-band, tp with cal with DCR running also.

  • band_gain_mode1_xband - high Tcal
  • archivist
  • balance -20dB
  • tell VEGAS that the DCR is the switching signal master
  • DCR is not happy. Configure for DCR, then add VEGAS. See if that works? Not sure how to set it up...
  • Scan 55 configure with bank_gain_mode4_xband_dcr. Then add VEGAS back in. Need to check IF filters. Runs at least, takes data
  • Scan 56 low tcal

Mode 4 - X-band, tp with cal with DCR running also.

  • Configure VEGAS with bank_gain_mode4_xband, then dcr with above
  • Scan 57 - high Tcal - fails
  • Scan 58 - high Tcal
  • Scan 59 - low Tcal
  • 4:30pm - get kicked off telescope!
  • Didn't have VEGAS switching! Argghh!!

-- RichardPrestage - 2015-03-13
Topic revision: r6 - 2015-03-17, 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