Setup:

• transmitter 3b (#4) was connected to the dummy load.
• A monitor cable was run from the forward power directional coupler to the hf control room to monitor the power
  
• (Note: the power monitor cables were hooked up in correctly.  We repeated this test on 150506)
• The signal generator was used to inject the RF signal at 5.125 MHz.
• into dana's box  with no attenuators..
• The anritsu spectrum analyzer was used to record the power.
• To look at the higher harmonics, a 75 MHz hipass filter was used (to block the fundamental).
• We ran 75 to 78 KW into the dummy load
• Reading from the transmitter front panel meter
• It would trip at 80KW.
  

The data:

    The plots  show the dummy load results (.ps) (.pdf):

• Page 1: the fundamental and 1st two harmonics
  
• the spectrum analyzer was set to :0 to 100 MHz,  rbw:1MHz, average 10 sweeps, preamp off,atten=40db
• If we had made the rbw narrower, we might have seen the 4th harmonic.
  
• The dotted red lines are multiples of the 5.125 transmitter frequency.
• The measured levels are:

• freq	harmonic	amp(dbm)	delta (db)
  5.125	fundamental	17.0
  10.25	2nd	-20.2	37
  15.375	3rd	-37.0	17
  20.5	4th	NoiseFlr:-49 expectedVal:-54	not seen

• Page 2: 0 to 200 MHz with the 75 MHz hipass filter installed.
• The spectrum analyzer was set 0 to 200 MHz, preamp on, 1MHz rbw,atten=0
• The green dotted lines are spaced every 5.125 MHz.
• When comparing these plots with the transmitted power.
• Page one had fundamental at 17 dbm
• On these plots, the same scale would be a few db less (from the filter loss).
  
• Top: first attempt shows intermittent spikes
• Red trace is peak hold, black trace is single sweep
• There were intermittent burst that filled in the red envelope
• the spacing between the nulls was about 13.6 MHz
  
• bottom: 2nd attempt: no intermittent spikes.
• The black trace is now an average of 10 sweeps
• The intermittent bursts were not seen
• There is power starting around 100 MHz spaced by the tx frequency.
• The largest peak was at 128.125 MHz (25 times the fundamental).
• Page 3: 0 to 500 MHz, 75 MHz HiPass filter. Rf on the off.
• Top: 0 to 500 MHz, RF is on
• You can see the harmonics around 120 MHz
• Bottom: turn off the RF power
• The harmonics around 120 MHz went away, so they are coming from the transmitter.

Summary:

• The first two harmonics were seen on dummy load. down 37 db and 54 db from the fundamental.
• We saw intermittent bursts with an envelope that was spaced by 13.5 MHz at the nulls.
• This did not repeat on the second measurement.
• Stronger harmonics were seen at 128.125 (25*F0) about 85 db below the the fundamental.
• these extended from 100 MHz to around 160 MHz.
• When banging on the 300 watt solid state amplifier for 3B, we would see power jumps in its output.
• We need to verify that we were correctly connected to the forward power coupler correctly.
• I need to get the coupler and the hipass filter frequency response,.
  

150506: tx4 dummy load, harmonics, no bursts     (top)

Setup:

• tx4 (3b) was  hooked to dummy
• a single monitor cable was run from the control room to the power monitor connector above the tx door.
• The anritsu spectrum was used (peak detection). 
• Looked at harmonics and wideband glitches.
  

 dummy load results:

    The plots  show tx4  dummy load results (.ps) (.pdf):

• The RF drive was set to 3.05 dbm and the signal generator.
  
• page 1: look at fundamental and 1st 5  harmonics
• Top plot: 10 khz rbw
• bottom plot: 1 khz rbw
• page 2: 75 MHz hipass filter installed
• 0 to 500 Mhz.
• green lines are multiples of the fundamental
• black line: RF drive on
• red line RF drive off.. the harmonics went away.
• page 3: blowup hipass filter in
• top: 0 to 200 Mhz.
• The strongest harmonic: 128.125 . 5.125*25.
• bottom: 20 Mhz around 430.
• the strongest harmonic is at 425.x
  

Summary

• tx4. 3.05 dbm siggen drive

• freq	pwr @ analyzer dbm	pwr delta (db) FN - F(N+1)
  5.125=F0	17.4	-
  2*F0	-23.7	41.1
  3*F0	-34.7	11.0
  4*F0	-48.6	13.9
  5*F0	-59.9	11.3
  6*F0	-68.5	8.6

• After the 1st harmonic, they fall off at about 11db.
• The harmonics at higher frequency:
• I need to get the directional coupler frequency response.
• the insertion loss of the 75 Mhz filter is probably a db or so
• the 128.125 (25*F0) harmonic is at -66dbm .. It is higher than the 6th harmonic (without the filter)
• the 425.375  (83*F0) is at -110 dbm.
• We didn't see any broad band glitches... but we were not running with peak hold on..

Calorimeter measurements using dummy load. (top)

    The dummy load was used to do a calorimetric calibration of the output power.

procedure:

• connect tx to dummy load
• set waterflow to about 26.5 gpm
• use anritsu spectrum analyzer to monitor the forward power
• set the siggen RF drive, record the water flow, input, output temps,  compute power
• pwr=.264 * gpm* deltaT   where deltaT is  (averageOuputTemp - inputTemp) degC.
• record the anritsu fundamental and 1st harmonic.
• increase the RF drive. and repeat.

07,08may15: tx4 calorimeter

• page 1: Calorimeter power vs 60db RF coupler output.
• Top
  
• green: 07may15 data,black 08may15 data
• blue *: 07may15 input water temp, red *, input water temp 08may15
• Bottom: there are 2 output temp meters. compute the power using meter1,meter 2, and the avg
• also over plot the avg pwr for 07may15
• Page 2:  Dummy load water temperature
• Black is the input water temperature, red,green are the output water temps for meter 1,2
• The x axis is the order we did the measurements.
  
• top : 07may15
• bottom: 08may15
• At the beginning of the test, the two output temp meters varied by 2deg C.
  
• later in the day they came closer together
• Page 3: Gain variation of the transmitter
• Top: RF coupler power/ rfDrivePwr (linear)
• red line is input water temperature
• Bottom: value of RF output output in dbm.d
  

Results:

• the curves do not remain linear over the entire measurement range
• The input water temperature dropped by 25 C during the tests.
• The transmitter gain (or the RF coupler output) varies .. probably with the temperature
• Things to do to improve the measurements:
• Try to stabilize the water temperature
• Check the water meters before anything is turned on.
• measure any offsets between the input and output meters
• Measure the RF coupler output before and after measuring the dummy load temps
• It was varying..
• wait a little longer between raising the RF drive and doing the measurements to give things a chance to stabilize.
• when using low rf drive, we should use a narrow rbw so we don't add in too much of the noise baseline.
  

12may15: tx 4 calorimeter  measurement and fit

    We replace one of the output temperature sensors on dummy load 4. It differed from the other 2 temperature sensors with no power input.
The input water temperature remained stable for the entire measurement (29 to 31 degC)

    The plots show the data and a linear fit to the data (.ps) (.pdf)

• Top: output power vs rf coupler for 12may15 data
• black: measured points
• green line: linear fit
• red points: excluded from fit.
• the linear fit was
  
•  pwrKw=-.121221 + 1.34914*(10.^couplerdbm*.1)
• Bottom: over plot data from all tx4 dummy load measurements
• the different days do not overlap
• changing the temperature sensor, and stabilizing the input water temperature probably helped.

12may15: tx 3 calorimeter measurement

    tx3 dummy load calibration. we got measurements till 50 KW then the tx would faults with ac drv Amp fault. The water temp was stable for the entire run. (input 29 to 31 C).
    The plots show  Calorimeter power vs coupler power and the linear fit (.ps) (.pdf)

• Top: data  (black) and fit (green). The red * were excluded from fit.
• The linear fit was:
• Pwr=-7.04039 + 1.27153* 10^(clprDbm*.1)
• the y intercept should be 0. Non zero causes the bending of the line on the log,log plot
• Bottom: same points, force fit yintercept to be
  

We need more high power points to get a better fit.

processing: x101/150512/dltx3_calib.pro

22may15: tx6 calorimeter measurement

23jun15: tx 4 on dummy load, calorimeter

Note:  At end of run  we found that connectors in the back of the broad band amp were loose. This was probably causing the changes in the drive input vs coupler output. This data wasn't used for the calorimetric fits (although it might be ok).

Dana first measured S11 on coax from tx4 to dipole 3b (with dipole end shorted). This is the dipole that had the arc. Saw about a 40 ns wide reflection at the far end.

Then ran tx 4 on dummy load the rest of the afternoon.
What we did:

• recorded drive, power levels and water temp as we increased the power.
• it eventually tripped  with PA Amp dc  fault at 93KW (from coupler value).
  
• as we were raising the power, the drive level was 3 or 4db higher than what was needed normally for tx4.
• The PA cathode current was also high.
• After the trip, the drive level needed to get to  a particular power was 3 to 4db  less than before the trip.
• We looked at the lowlevel broadband amp to see if it was varying. banging on the front caused the power to drop out.
• We found that the output of the broadbad (that goes into a coupler and then on to the next stage was loose.
•  Looking at the other tx's some of them also had loose connections on the output of the broadband amp.
  

• We had 26.5 gallons/minuter for the entire time).
• the reflected power always read less than 1 kw.
  

• The last two columns show the delta db in drive level and coupler reading. At the higher levels we were compressing
• after the trip, the drive power need to get to a particular output power dropped by about 3.5 db.
  

150629: tx4 on dummy load

    We started with tx2, but kept getting trips with drv AC A  .. with less than 10Kw of output power. So  we switched to tx4.
    Tx4:

• we adjusted the bias's first. the data
  
• took data till 5.8 drive. The spectrum analyzer readout wondered a bit
• This data was then added to the previous tx4 calibration data  and fit for coupler value vs calorimetric power (see the plots)
  

150811- tx2 on dummy load (tx5 host burst)

• pa screen was 400 ma with 0 drive
• at 100KW tripped with drv A amp after about 5 minutes of running.
• At 80KW the PA bias was about 180 instead of 380 (at 80 kw)
• this data was fit coupler value vs calorimetric power (see the plots)

home_~phil