Background:

• We needed to test the hf for the upcoming hf campaign (in jul17)

What we did:

• We measured the rf rise time through the TX as well as the relative delay between the broadband "CW" and "CYCLE" modes.
  
• The hf was run at 5.125 MHz
• The broad band amplifiers have a keyline switch that enables/disables the amplifier.
• To try and stop the  broadband amplifiers from overheating, for each broad bandamp,  Alfredo replaced the normal termination (broad band enabled) with a line running back to the hf control.
  
• Each of the 6 lines is controlled by a switch in the hf control room
• Switch up: cw mode
• The broad band amplifier is always enabled
• Switch down: "cycle mode"
• the broad band amplifier is enabled when rf is on , and disabled when rf is off.
• We did some tests to see the response times of the amps in cycle mode.
• The hf mesh was set to the 5 MHz position.
• 430 tx, dual beam
• gr -> usrp
• ch -> echotek [-5.5,5.5]
• tx 4 is off, tx 6 won't come up .. so using tx 1,2,3,5
• 09:36   start to bring up tx
• 09:54   move dome to za 1.06, mov ch to 11 (ch was overloading at za=0)
• 10:00  switched echotek [-5.5,5.5] to [5.5,9.5]
• 10:15  hf back on  with tx 1,2,3,5
• 10:24  60 KW on 4 tx
• 10:43  cable car moved to 1/2 way up the cable. has test equipment for drone.
• 11:06  bringing  up tx6
• 11:0x  set ntp on raspberry pi hfmon to ntp0
• 11:12  power to 305KW with 5 tx: 1,2,3,5,6
• 11:16 turn off broad bands to see which is arcing.. looks like tx2
• 11:22 turn off tx2 , broad band was overheating again.
• switch to another broad band that carlo had refurbished.
• 11:57  (11:27?) disconnect tx1,3  rf coupler readback from spectrum analyzer.
  
• Attach to scope so we can measure the relative delays between "cycle mode" and "cw mode" on the new switches.
• 11:38  tx2 new broad band installed
• 11:48 send tx1,3 rf coupler readback -> scope.
• rf coupler -> cable To hf control -> 10db pad -> scope (50 ohm termination)
• auto toggle 2 sec on , 4 sec off
• Auto toggle was set to run 2 sec on 4 sec off.
• using autotoggle 2sec, 4 sec
• This would occur if the 5.125 MHz was getting into the ch 400 MHz Local oscillator.

Measuring the tx rise time  and  "cycle,cw mode" relative delays

• tx1 and tx3 rf coupler readback were disconnected from the rf switch (that feeds the spectrum analyzer).
  
• the plumbing was
• rf fwd coupler output -> cable to hf room -> 10db pad -> scope input with 50 ohm termination
• Put the broadband switches to "cw" mode
• we measured a negligible delay between the 2 signals. So tx1 and tx3 have similar delays.
• Leave Tx3 broadband as "CW" and set tx1 broad band to "cycle"

The plots show the rise times and relative delays  (.ps) (.pdf)

• Page 1: The rise time for the signal coming out of TX3 ("cw" mode)
• The broad band amplifier was always enabled.
  
• the frequency was 5.125 MHz
• It took about .4 milliseconds for the signal to rise up to full strength.
• This does not show the delay through the tx, only the rise time.
• Page 2: compare "cw" and "cycle" mode delays
• Top: Start "CW" mode to start "Cycle" mode
• there is a delay of 3.9 milliseconds from the start of tx 3 (cw) to the start of tx1 (cycle)
• Bottom: extended time scale showing cycle mode rise time
• Time tx3 "CW" start to tx1 "CYCLE" full scale is 17.4 milliseconds.
• So there is a delay of 3.9 milliseconds and then about 13 milliseconds of rise time.

Summary:

• Tx 2 broad band was overheating.. it was replaced with a second
• the timing for the  "cw/cycle" mode switches for the broad band amplifiers were tested
• Normal CW (always on).
• the rise time is for the signal is about .4 milliseconds
• Cycle mode (broadband on,off with rf enable)
• there is a 3.9 millisecond delay
• the rise time is about 13 milliseconds.
• So if you are doing fast  rf enable,disable you need to be careful about using the "cycle mode".
• If the broad bands are overheating, and you have a long rf disable period:
• start in cw mode
• wait for the rf disable
• switch to cycle mode (this turns off the broad band)
• before the rf enable, flip the switch back to cw mode
• If you don't care about a fast turn on, then you could just leave it in cycle mode.

home_~phil