Background:

• Problems during previous test  07Feb17:
• tx 3 and tx 6 had circuit breaker faults.
  
• on 13Feb 17 rf switches were installed at the broadband input to switch between the control room synth value and a value input locally at the transmitter.

What we did:

• mesh was at 5MHz position. used 5 MHz rcp , freq=5.125 MHz.
  
• 430 transmitter was used. data recorded with echotek and usrp.
• The interim correlator monitored the arcing using the ch receiver. It also had the 430 tx sample in the spectra.
  
• 09:50 start to bring up tx at 5 MHz.
• 10:15 50KW each tx. stay here for about .5 hours
• sampled tx power much noisier than before . Jumps around by > 1db.
• 10:54  80KW each tx. measured power still noisy .
  
• 11:12 tx 4 dropped out, then brought it back up.
• 11:31 found out about the rf switches being recently installed.
• brought down tx2 and removed the rf switch.
  
• tx 2 power measurement now stable.
• 11:48 bring down tx1 . remove rf switch
  
• 11:45 started to bring up transmitters
• 11:55  bring down power all tx. remove the rest of the rf switches.
• 12:07 all 6 tx back up to 80KW.
  
• leave the tx drive power fixed for 1.5 hours, see how the output drifts.
• 12:24  tx2 jumps up by about .5 db.
• 12:48  some arcing seen on the ch monitor.
• turned off 430 transmitter. went under the bowl to check for arcing.. nothing seen.
• 13:20 back from bowl. 430 tx turned back on.
  
• 14:03  tx 2 jumps down by 2+ db. then comes back a few minutes later
• 14:30  tx 2 jumps down by 3db then comes back up a few minutes later.
• saw broadband go from 110 watts -> 85 watts -> 110 watts without changing the input drive levels.
• 14:40  adjust power levels in preparation for toggling... about 75KW each tx.
• 14:47 toggling. 2 sec on, 4 sec off.
• 14:58  tx1 tripped with reverse power. reset.
  
• 15:07  back to cw. left power levels untouched to see what the power levels were.
• they were a little higher then when we started..probably tx cooler since 30% duty cycle.
• 15:10 tx2 dropped 80 to 45 KW. broadband went 110 to 80W.  a few minutes later came back to 110 w
• 15:15 stopped.

Plots of output and drive power:

The plots show the data we took while the hf was on (.ps) (.pdf)

• The page has 3 frames:
• each color is a different transmitter (1 through 6).
  
• top  frame: rf power (in dbm) measured at the output of the rf coupler.
• middle frame: the same output power, but displayed in KWatts
• bottom frame:
• The synthesizer drive amplitude used
• The dashed green line shows when the rf was on
• the red line at the bottom is the total power from the carriage house 430 receiver (polA 425-435 MHz). When we have arcing, there are jumps in this power (rfi can also make jumps).
• 10.1 to 12.0 - you can see the jumping around of the output power. a
• The rf switches were removed at 12.1.. things were a lot more stable after that.
• 12.1 to 13.5 output drive power left constant.
• tx 2 red. is wandering around.
• 14.0,14.4 tx2 red drops output power level by up to 3db.
  
• problem was the broadband output was changing.
• 14.8  start auto toggle 2 on 4 off.
• 15.1  tx2 red drops output power level by 3db again.. comes back by itself.
• problem again with the broadband.
• 15.25  done

clp movie during hf test

    The 430 clp data from the carriage was processed and  a movies were made. They showed how often arcing appeared.

links to the movies:

Setup:

•  ch at za=0, looking through the heated volume
• 10ms ipp, 440 usec rf pulse data taken for entire run.
  
• echotek:
• ch, 3 to 10 MHz were covered (upshifted)
• No data taken 12:47 to 13:23 (we were under the bowl looking for arcs, so the 430 was turned off).
  
• usrp:
• ch, 25 mhz about 430 mhz.
• When processing the data
• the rfi removal algorithm was not used (didn't want to clip the enhanced plasma line).
• Each frame of the movie was a 10 second integration.

• We saw lots of arcing in both the usrp and echotek
• 10 - 10:30 arcing when running at 50KW each tx
• no changes in drive level were made during this time.
  
• The morning also had the rf switches installed. After removing the switches, the arcing continued.. so they probably did not contribute to the arcing.
• 12 to 12:45 arcing got worse.. again no changes in drive level
• we then shut off the 430 tx to go under the dish and check for arcing . nothing was found
• 13:24 430 back on. arcing continued.. but not quite as bad as 12:45
• Images of the plasma line
• in both the usrp and the echotek data there is an image of the plasma line shifted by +/- 5.125 MHz,. It is only there when the heater is on.
• Since it is both usrp and echotek, we can blame it on a particular backend. The harmonic/intermod must be occurring upstairs (since we split the usrp,echotek at the crr2 output.

Summary:

• tx4 tripped once, then came back up and ran reliably.
• tx2 output power jumps down by about 3db for a few minutes and then comes back to previous value.
• saw the output of the broadband change each time.
• This happened 3 times during the test.
• saw some arcing on the ch monitor. On the scope these looked like shorter pulses. So the interim correlator 1 sec averages didn't see them.
• We see arcing in the clp movies
• There is a harmonic/intermod of the plasma line offset by  +/- 5.125 MHz. This is seen in the usrp and the echotek.

home_~phil