Background:

• This was the first test of the hf system following hurricane maria
  

What we did:

• The mesh was in the 5 MHz position (i assume..)
• we transmitted at 5.125 MHz rcp.
• Only 2 generators were available to provide power to the hf. the 430 tx was on commercial power.
• part of the testing was to see how much hf output we could get from 2 gen (while still leaving room for the site to switch in case of a power dip).
  
• We measured the rf rise time through the TX as well as the relative delay between the broadband "CW" and "CYCLE" modes.
• 430 tx, single beam -> gr
  
• gr -> usrp
• ch -> echotek [-5.5,5.5]
• 10:00 bring up 430tx (approx)
  
• 11:00 rf off 430 tx.. we could hear arcing upstairs.
• 10:45 start to bring up hf tx.
• 11:06  hftx 4 trips on reflected power. output power was only 15kv
• Leave tx4 off
• 11:30  leave 5 tx at 32-38kw to see if arcing subsides
• 12:40  hf off
• 13:10  hf back on
• 13:30  430 back up . heard arcing when first turned rf on, but then sound went away.
• 13:55 350KWatts cw
• 14:23:26 autotoggle 4 sec on 4 sec off
  
• 14:27:37 autotoggle off
• 14:31:26 autotoggle on
• 14:38:24 autotoggle off
• 14:51:28 autotoggle on
• 14:55:27 autotoggle off
• 14:xx:xx during autotoggle measured pulse rise times at tx 5 then in ctrl room
  
• used tx5 outputs.. along with rf pulse going to tx4
• scope inputs:
• 1 tx4 rf input
• 2 broadband output
• 3 ipa output
• 4 drv output
• .000.isf
• then looked at tx 4 rf pulse coming from phase box
• saw slow rise time. so problem is in dana's box
• .0001 isf (just chan 1)
• turn off auto toggle, turn rf on, off using rf enable of siggen.
  
• saw fast rise times.. .0002isf (just chan 1).
• So problem is with enable, disable function in dana's box. he is switching the dac  voltages for then enable,disable.
• 15:00 done

HF output power

Plots showing hf tx drive and output power (.ps) (.pdf)

• Top: hf rf coupler output vs hour of day
• Each color is a different tx.
• the low blue level is when tx4 was disabled.
• middle: output power in KW
• we got up to about 350 KW total output power (70KW*5).
• bottom: synth drive amplitude (in dbf)
• The red line with noise it the total power from the gregorian receiver measured with the interim correlator
• The large spike at 11:05 may have been from the waveguide arcing?
  
• Around 12:00 the arcing died down a bit. It back up again in the afternoon.

Generator power vs hf output.

    The data from the 2 running generators was used to measure how much generator power was needed to drive the 5 hf transmitters.

Generator output power vs hf output power (.ps) (.pdf)

• Top: generator output power vs hf output power.
• 08:30 to 10:30 - this has the cooling pumps on , and all 6 tx have the hv up (but no rf output)
• 10:30 to 12:30 - bringing up hf tx output.
  
• only tx 1,2,3,5,6 outputting
• tx 4 hv may have been left on (we used the disable switch in the control room
• 12:45: dip in power. I wonder if the hv for tx4 was turned off here...
• 13:12 to 15: 5 hf tx output. tx4 hv is off.
• 14:20  start autotoggle  4sec on , 4 sec off.
• middle: total rf output from 5 hf tx
• bottom: gen power vs hftx output power

GenPwr notes:

• We used 400 KW genpwr to run the water pumps and 6tx with HV on.
• Turning off 1 Tx HV looks to save about 70KW.
• For 350KW output (from 5 tx) we needed 1100 KW.

Measuring the rf pulse rise time:

    On 26jun17 we measured the rise time of pulses through the transmitter by looking at the rise time of the rf coupler output
on an oscilloscope. We found that:

• the "cycle mode" switch added around jun17 had a 3.8 ms delay and 13ms rise time.
• The normal "cw" mode position of the switch had a 400 usec rise time.

    On 07mar18 we measured the rise time of an rf pulse at the hftx5 monitor ports. The setup was:

• Scope chan1: rf pulse from hftx4. This was the cable coming from the ctrl room. This was used as the reference.
  
• scope chan2: broadband output port
• scope chan3: ipa output port
• scope chan4: drvr output port
• Different attenuators were placed in front of the scope channels so the levels are not the original values from the port.

The plots show the rise time of the rf pulse (.pdf)

• Top rf pulse from control room
• 2nd: output of broadband.
• 3rd: output of ipa
• bottom: output of driver tube.
• the delay is occurring before the pulse gets to the hf xmiter.

plots showing a blowup  of the sine wave after the rise time (.pdf)

• this is a blowup of the 4 previously signals, after the pulse has risen to full power.
• the data is still taken at the monitor ports of hf tx 5

The final plots show the rf pulse rise time in the ctrlroom (.pdf)

• top: pulse rise time at the output of dana's phase shifter
• you can already see the slow rise time.
• bottom:
• the pulse rise time using the signal generator rf on,off.
• Still measured at the output of the phase shifter.
• In the case, the rf enable, disable line is not used..
• there is no delay in the pulse rise time.
  

Summary:

• Tx 4 would not come up. It had lots of reflected power at low output powers
• the splice in the waveguide down by the towers showed the heliax to be full of water.
  
• We had lots of arcing.
  
• we ran 5 with 350 KW using 5 transmitters.
  
• the 2 diesel generators were output about 1.15 MWatts
• the generators can probably output 1.6 MWatts (800KW each).
• 1.15 MWatts leaves plenty of room for the site to switch if a power dip (the site is probably 400 KW on a warm day).
• We found that the 300usec rise time of the rf pulse is coming from the dana's phase shifter
• Autotoggle uses the rf enable,disable line to run the rf on, off.
• To disable rf, the voltage to the dacs is removed.
• To enable rf, the voltage to the dacs is applied.
• This must be causing the slow rise times.
• When using the signal generators rf enable,disable we see instantaneous rise times.
• We still need to check the output of the rfi coupler.
  

home_~phil