Intro

    Each transmitter has output ports where you can monitor the output of the various tubes:

• rf output ll ampl (low level broad band amplifier)
• rf output IPA anode
  
• rf output DRV AMPL anode
• xmtr  output FWD power sample (after the filter)
• (we skipped the rf output PA anode)

Setup

• front panel output
  
• 151102: monitor port-> 6dbPad -> bnc cable -> tectronix scope
• Later setups:
  
• broad band port -> 6db pad -> bnc cable -> tectronix scope
• ipa,drv,fwdPwr port-> bnc cable -> tectronix scope.
  
• Scope setup:
• acquire: sample, 100000 or 1000000 pnt buffer
• hor: 100ns/division
• chan: all with 50 ohm termination, 250Mhz bw
• 1 - 1 volt/division
• 2,3,4: 200 millivolts/division
• save: as .isf all traces (170207: .csv files, only 100000 points).
  

plots/images:

• save file
• The screen data was saved as an .isf binary file, read into idl, and then plotted:
• page 1: voltage vs time (used 100000 length buffer)
• page 2: spectral density (computed from 100,000 length buffer). Scaled to db below peak
  
• Scrn image:
• This is the screen capture from the scope
  
• trace 1:low level broad band
• trace 2 :IPA Anode
• trace 3 : driver Anode
• trace 4 : rf  fwd power output
  

Comparing the  tube stage monitor ports.

• for each tx:
• plot tx separately
  
• plot 2 usecs of voltage data for each tube output
• Each  title line includes the bias voltage that was used
• The mean,rms,and dbm are computed from the measured voltage (assuming into 50 ohms).
  
• plot the computed spectra (using all  sampled data 100k Or 1e6 sample points) for each tube stage.
• over plot all tx
  
• for each tube stage, overplot 2 usecs of data from each tx.
• overplot the relative gain of each  tube stage for all tx. the relative gain is defined as:
  
• stageN PwrOut/stage(N-1)pwrOut.
• The power  was taken from the value of the spectrum at the fundamental (so harmonic power was not included).
  
• Since the monitor outputs have different hardware scaling, the absolute gain is not known, but the gain variation tx to tx is meaningful (assuming the hardware scaling is the same between tx)
• For the broadband input, we used the dbf drive level.
• The scope setup
  

Summary

• When over plotting the tube outputs, we see a phase jump tx3,tx4 ipa output relative to the other transmitters. 
• This jump was there back on nov11. It continued through 23feb17 (07feb17 tx3 was not measured).
• According to dana, the jump could be in the tube, or  more likely in the tube to tube matching network.
  
• The monitor data is taken at the tube output before the tube to tube matching port.. So a problem in the IPA output may be coming from the matching network between the  broadband and the ipa.
• An attempt was made to compare the forward power output phase to the phase used when setting up rcp or lcp
• a sine wave was fit to the voltage output. This  gave amp,freq,phase, and offset.
• the phase of  the pol ref (tx2 for 5Mhz, tx5 for 8Mhz) was subtracted and the result inspected.
  
• The results looked promising for some of the tx, but we made not have had enough data samples for an accurate phase measurement
• 1million points at 5Ghz sampling gave 200usecs of data. 
  

• 151102:
• Notes said 6db pad on tx port input. Looks like tx5 (first tx) didn't have a pad on broadband.
  

• 170207:
  
• This data was taken in .csv mode rather than .isf (binary) mode. there were 100k points rather than 1000k points;.
• The spectra skirts are much larger than the .isf 1000K point data.
  
• Used a 6db pad on the broadband output. The processed data was scaled by 6db to correct for this.
• tx3 was not recorded
• The power ratio DriverOutput/IPA output differs from the other Tx by 3 to 4 db. this is not seen in the 8mhz data.
  
• 170223:
  
• Used a 6db pad on the broadband output. The processed data was scaled by 6db to correct for this.

home_~phil