HF tests on antenna
05may15
150504: Tx1,2 low power. monitor with 327
rcvr.
Setup:
- transmitters 1,2 were hooked to the dipole antennas
- rf drive (5.125 MHz) was injected (-15dbm). The forward power
meter on the tx's showed 4KW on tx1 and no reading on tx2
- The 327 receiver and the mock spectrometer was used to monitor
the signal into the dome.
- 302 to 352 MHz, 8192 channels, 1 second dumps were taken.
- The dome was at az=0, za=15 degrees.
- The if/lo power levels were adjusted before the hf rf
was turned on, and then left fixed.
- 300 seconds of data was taken with the rf off (the high
voltage was on for part of this time)
- The rf was on for about 95 seconds, then dana lowered the rf
drive level in 10db steps until around 140 secs it was
off.
The data:
The plots shows the
average spectra with the rf drive off and then on (.ps) (.pdf)
- Black:rf drive off. average 300 seconds. (the high voltage was
on for part of this time).
- red: rf drive on. average 90 seconds. (before dana started
lowering the drive power).
- scaling: the median value from the rf drive off spectra was
used to normalize both spectra (so they should be on the same
scale).
- The yscale in logarithmic in units of Tsys.
- Green dashed lines are multiples of 5.125 fundamental
- Results:
- the 5.125 MHz comb is seen with values up to 5 times Tsys
(about 120 K) in the
- we need to check if the 307 MHz birdie is connected with the
hv (i think not).
- the baseline for the rf on spectra is a bit lower than the
rf drive off.
- there may be some compression. We need to fire the
cal next time.
- ripples are seen in the rf drive on spectra. This are
probably reflections of the harmonics between the platform
,dipoles, and dish.
The dynamic spectra shows the rf drive
on then off (.gif):
- the vertical scale is seconds.
- dana started to turn off the rf drive around 95 seconds (in
10db steps).
- 92 it was at -15dbm (regular tx value)
- 96 : -25dbm
- 107: -35dbm
- 120: -45dbm
- 135: -55dbm
- 150: rf off.
- You can see the 5.125 MHz comb until around 107secs (when it
went to -35 dbm.. 20db below the original -15dbm).
Summary:
- the 5.125 MHz comb was seen in the 327 rcvr
- about 5*Tsys
- the comb was not resolved in the 6.8 Khz channels.
- The tx power was about 4KW, the rf drive was -15dbm
- We saw the comb until -35dbm rf drive level.
- There may have been a slight compression in the 327 noise
floor. we need to repeat this and fire the cal.
- ripples were seen in the rf drive on bandpass. these are
probably standing waves.. platform to dish.
- The tiedown load cells operated normally during the test (not
jumps).
processing: x101/150504/hftest.pro
150506: 327 results with Tx1,Tx2 on antenna and tx4 on
load (top)
setup:
- The 327 receiver and the mock spectrometer was used to monitor
the signal into the dome.
- 302 to 352 MHz, 8192 channels, 1 second dumps were taken and
then 1millisecond dumps (with 1024 channels)
- We took data with tx4 on dummy load
- look at the hvon, rf on. to see if we could see anything in
the 327 rcvr.
- The az,za was at 7,15, then 7,6.5 (to let people into the
dome).
- We then switched to tx1,tx2 on antenna, with az,za at 0, 15
degrees.
- data was taken with 1 sec dumps, and then with 1 millisecond
spectral dumps
327 spectra with tx4 on dummy
load
- Telescope at az,za 7,15 then 7,6.5 deg
- We took data for 2647 seconds at 1sec dumps, while tx4
was on dummy load
- The 327 rcvr sees the hv turn on when the transmitters are on
dummy load.
- the 327 rcvr does not see the 5.125 harmonics when tx is on
dummy load.
The 1st dynamic spectra shows the 2647
seconds of data with tx4 on dummy load (.gif)
The 2nd dynamic spectra is a blowup
showing the first 800 seconds (.gif)
- The vertical axis is seconds
- the first image is all 2647 seconds, the second image shows
the first 800 seconds.
- seconds when things occurred:
- 106 hv on for tx4
- 276: start moving dome za 15 -> 6.5 so people could go
into the dome
- this caused the level change in the image.
- 362: solid state linear amplifier (300w) turned on.
- 462: started to turn on rf drive power (started at a low
level).
- 2198: turned off rf power for a little while.. then back on.
327 receiver with tx1,2 on
antenna
- The tx monitoring showed no power coming out of tx 2.
- we see harmonics start around -20dbm rf drive, broad band
glitches at -15dbm
- The spectra level (in band) is about 10times for the 327
rcvr (may be large out of band birdies).
- the receiver recovers rapidly after a glitch.
- 327 sees no power output from tx2
- the harmonics, glitches do no change when dana's rf driver is
bypassed.
the first dynamic spectra (file
300) has 1229 seconds of data while the rf drive was turned
on tx1,tx2 (.gif)
- The two polarizations have been summed (.gif)
- Seconds when things occurred:
- start - hv was already on
- 70-75 secs linear amp on 1a,1b. rf drive is still off
- 129 : rf drive on , at -65 dbm
- 260 : rf drive -50dbm
- 310 : rf drive -40dbm
- 790 : rf drive -30 dbm
- 815: rfdrive -20 dbm
- 5.125 MHz comb appears in 327 receiver
- 1060: rfdrive -15dbm
- Start to get broadband noise in 327 rcvr (or is somthing
out of band causing this?).
The plot shows the
spectra around second 1060 (when the rf drive went to -15dbm)
(.ps) (.pdf):
The 2nd set of 327 dynamic spectra
shows the rf drive being turned on,off for tx1,2 (.gif)
- seconds when things occured:
- start -15dbm rf drive into tx1,2
- broad band glitches and 5.125 harmonics present
- 115 secs: turn off rf drive
- glitches, harmonics went away
- 167 secs: rf back on.
- 202 secs: turn off rf drive for tx2:
- no change, glitches, harmonics remain. --> 327 rcvr
sees nothing being output from tx2.
- 233-240 secs: rf drive tx2 in , rf drive tx1 turned off
- harmonics, broadband glitches are gone
- all of the junk that 327 rcvr sees is coming from tx1.
- 390 : rf drive tx 1 back in
- 448 : rf drive off
- 575: rf drive for 1a is coming straight from siggen
(-22dbm level).
- bypassing dana's box.. harmonics, glitches still present.
High time resolution (1 millisecond) spectral
data was taken with the mock spectrometer and the 327 rcvr
while broadband glitches were occurring using tx1 on antenna.
setup:
- mock spectrometer: 56 MHz bw, 1024 chan (55 khz resolution),
138 seconds of 1 millisecond spectra, pola,b averaged.
The dynamic spectra shows the
occurrence of the glitches (.gif)
- (the display has been averaged to .155 seconds).
The 1st plots show the total power vs
time at 1 millliseconds resolution (.ps) (.pdf):
- Page 1: the total power vs time for the 138 seconds.
- at 1 millisecond resolution the power is about 100 times
tsys.
- Page 2: spectra and total power
- top: overplot the 1st 20 spectra
- middle: over plot the 1st 50 spectra
- bottom. plot the total power for the first 400 spectra (.4
seconds)
- Red line. 0 to here is the total power for the top frame
- green line: 0 to here is the total power for the middle
frame
- The receiver recovers by the next 1 millisecond sample.
The 2nd set of plots shows the spectra (abs(fft)) of the total power
(.ps) (.pdf)
- top: full 500 hz bandwidth
- bottom: 0 to 10 hz blowup
- No 60 hz is seen in the total power spectra.
Summary 150506 327 rcvr:
- On dummy load
- the 327 rcvr sees the hv on
- it does not see the 5.125 MHz comb
- tx2 on antenna
- tx1 on antenna:
- harmonics are seen at -20 dbm rf drive
- broadband glitches occur at -15dbm rf drive.
- since they are time variable.. if we waited long enough we
might have seen them at a lower rf drive power
- With 1 second resolution
- the the glitches are 10 time Tsys
- With 1 millisecond resolution,
- the glitches are 100 times Tsys.
- The 327 receiver goes back to normal on the next
millisecond after the end of a glitch.
- glitches occurred that were not resolved in 1 millisecond.
- multiples of 60 Hz was not seen in the spectra of the
total power.
processing: x101/150506/hf327.pro
150514: tx6 on antenna, 430 gr taking
data (top)
On 150514 we put tx6 on antenna with the 430
dome taking data..
Setup:
- az,za : 0,15 deg
- mock spectrometer : bandwidth 53MHz,cfr:430, 8192 channels, 1
second dumps
- transmit at 5.125 MHz.
The dynamic spectra shows the last 300
seconds of data taken (.gif):
- the image was scaled so the dynamic range was 15*Tsys
- You start to see the 5.125 MHz harmonics around second 50.
- We arrived at +2dbm siggen output (34kw) around second
282.
The plots show spectral and harmonic
strength (.ps) (.pdf)
- Top: over plot 20 spectra between seconds 270 and 290
- The larger 1 MHz ripple has a spacing of about 1 MHz (stw?)
- bottom: the strength of the 4 harmonics of 5.125MHz that fall
in the 430 dome band
- We reached +2db (34Kw) around 282 seconds.
- The harmonics don't appear to get stronger as we increase
the power.
- Not sure why they increased so much at 120 seconds:
Summary:
- we transmitted on antenna with tx6.
- siggen -20 to +2db ( 0 to 34 KW)
- occasional spectra had 1 MHz ripples up to 6* tsys
- These were averaged to 1 second. I don't know if the data
was a stronger, shorter pulse
- The 1 MHz is close to the standing wave period of platform
to ground
- The harmonics were seen in the 430 gr receiver
- they did not seem to get stronger as the power increased.
processing: x101/150514/hftest.pro
150526: tx1 on antenna, take data with
alfa (top)
- az,za: 360, 15
- beam 0 alfa, mock spectrometer, .1 sec dumps, 172 MHz/band,
8192 channels
- go from -20dbm siggen to 0dbm (20KW)
- saw no harmonics. Saw 1 MHz ripple spectra
the dynamic spectra shows that
last 600 seconds of the tx6 xmit (.gif)
- horizontal frequency scale is reversed
- The dashes at the right is punta salinas radar
- the white stripes are from the transmitter
the plots show the strength of the
total power and the spectra (.ps) (.pdf)
- page 1: total power vs time
- Top: 1390 to 1520 MHz was used.
- the yaxis is Tsys.
- the vertical dashes show when we increase the rf drive to
tx
- bottom: blowup of total power 541 to 545 seconds
- The spectra on page 2 are from this subset of the data
- page 2: overplot spectra for time: 541 to 545 seconds
- 50 .1 second spectra
- each spectra is separated by an offset of 1 (for display)
- The largest spectral peak is about 100 Tsys
- page 3: individual spectra, acf
- Top: spectra with the largest spectral peaks
- bottom: acf: abs(fft(spc)) .
- Shows the delay for the standing waves
- The peak around 1 usec is round trip dish to platform
- 2 useconds is a double bounce.
Summary:
- Did not see any obvious harmonics of 5.125 MHz
- we saw intermittent bursts. The total power of these was about
20*tsys
- spectral peaks got to 100 Tsys
- The spectral oscillations show 1 and 2 usecond delays..
- So this power is bouncing between dish and platform.
- The stw periods show that this data is coming in through the
horn. It is not the 5.125MHz modulating the bias values
processing: x101/150526/tx6ant.pro
150527 tx6 xmit 100KW on ant, record data
with alfa bm 0 (top)
We took data with alfa while transmitting on
antenna with tx 6. The setup was:
- alfa bm0. (the other alfa amps were turned off).
- mock spectrometer recording 172 MHz bw, 8192 channels, .1
second dumps, centered at 1450 MHz
- We started at az=360,za=15.
- data was taken for about 1 hour as we varied the transmitted
power
- We then moved from az=360 to az=240 deg. This pointed the dome
at tower 8 . Tx6 is in this direction.
- data was taken during the move and continued after we
arrived at az=240.
- Lightning began during this 2nd set of data. We stopped
after a power dip from lighting.
The first set of plots shows alfa
data with tx6 transmitting and the azimuth at 360 degrees (.ps)
(.pdf)
- page 1: total pwr vs time for the 1
hour we sat at az=360,za=15
- The power was computed from 1380 to 1520 MHz.
- top: 1 hour of data. Black is polA, red is PolB. the
vertical scale is normalized to Tsys.
- The green lines show the transmitted power (we actually
reached 106 Kw around 16.2 hours)
- The blue dashed lines delimit the data used for the time
blowup in the bottom frame.
- bottom: blowup in time for strongest power deviations.
- the * show the .1 second samples.
- Most of the bursts last for less than .1 seconds.
- Page 2: spectra around the largest
burst at 16.055 hrs
- 10 spectra are overplotted.
- Top: polA
- bottom: polB
- the deep nulls are from standing waves.. dish to platform.
- Page 3: spectra of largest burst,
and stw
- top: spectra of largest burst. polA
- middle: standing wave amplitude vs delay in usecs. The time
resolution 7 ns..
- bottom: smooth the stw vs delay to 64 ns
The second set of data shows
when we moved from az 360 to 240, and then sat at 240 (.ps)
(.pdf)
- page 1: total power vs time
- top: all the data
- black polA, red polB, green is transmitted power.
- we started moving 360 to 240 at 16.47 (start of plot)
- we arrived at az=240 at 16.52 (dashed light blue line) ..
the move was at slew rate: .4deg/sec
- We only transmitted 3 kw while we were moving.
- the jumps around 16.54 to 16.545 occurred between az 258
and 250.
- on 26nov15 was also saw bursts with only a few kw
transmitted.
- bottom: blowup of strongest burst
- many of these lasted .5 to 1 second
- the total power is approaching 1000 * tsys.
- This is probably being limited by the dynamic range of the
if/lo..
- Page 2: plot spectra around the
strongest burst
- top full frequency range
- bottom:show 40 MHz bw.
- Page 3: strongest spectra, and stw
vs delay
- top: spectra with most power
- middle: standing wave vs delay. delay resolution 7ns
- bottom: smooth delay resolution to 64 ns.
- You don't see the peaks at 1 and 2 usec that are the single
and double bounce times between the dish and the
platform.
- this may be lightning
the final plots show the
total power,spectra, and stw when lightning caused the power
failure (.ps) (.pdf)
- page 1: end of datataking when lightning/power dip occurred
- the power dip occurred at 16.776
- top: total power vs time
- The green line is lightning that occurred about 5 seconds
before the power dip.
- It caused the power dip
- it lasted for about .8 seconds
- the red line is another peak that occurred prior to this.
it laste for about .2 seconds
- bottom: the spectra for the red and green flagged line.
- the spectra overlay each other. so they both come from the
same source.
- page 2: the standing wave vs delay
- top: 7 ns resolution
- bottom: smooth to 64 ns delay resolution
- this looks like the other stw power vs delay that we saw
with telescope at az=240.
Summary:
- We transmitted 106 KW on tx6 and took data with alfa bm0
- tx6 is at az=240.
- When at az=360
- the strongest burst was about 80 times Tsys (in 140 MHz bw)
- the bursts were not resolved in .1 sec samples
- we saw the 1,2 usec peaks in the stw vs delay.
- So i think these were from the transmitter
- when moving az 360 to 240 and then sitting;
- we saw bursts at az 258 to 250 with only 3kw of xmittted
power
- There were large bursts of 700 times Tsys
- they lasted .1 to .9 secs
- the signal was probably saturated in the if/lo
- The 1 MHz nulls, 1,2 usecs stw delays were not as obvious as
the first data set
- A power dip at the end was caused by lightning
- the spectra and stw vs delay for this bursts, looks
identical to the other bursts at az=240
- The large 500 to 800 times tsys bursts are probably from
lightning.
processing:
x101/150527/hftest_1.pro, hftest_2.pro
150602: tx1, tx6 on antenna. xmit 100KW,
tx4 fil delay problem (top)
tx1 on antenna. xmit 100KW
- we transmitted using tx1 on antenna.
- az=360,za=15, alfa bm0 polA,B, .1 sec sampling with mock
spectrometer. 172Mhz bw, 8192 channels, cfr:1450Mhz
the plots show the mock data
while transmitting on tx1 (.ps) (.pdf)
- Page 1: total power vs time
- top: total power 1390 to 1520 Mhz vs time.
black=polA,red=polB, green is xmitted power in KW.
- There were only 3 or 4 glitches in the 40 minutes we took
data
- bottom: blowup in time of largest burst. the samples (*) are
at .1 second spacing.
- page 2: spectra at largest bursts and stw
- top,middle: polA,B spectra around largest burst
- bottom: stw vs delay. peak at 1 usec.
tx4 stuck in filament delay
transmitter wouldn't leave filament delay
tx6 on antenna. xmit 100 kw
With tx on antenna we first took data with
alfa bm0. xmitting up to 100 KW.
the plots show alfa bm0
data while xmitting on tx6 (.ps) (.pdf)
- Page 1: total power vs time for the 9.6 minutes of data
- black is polA, red is polB, green is the xmitted power in
KW, the dashed blue line is blown up int he bottom frame.
- there are only 3 or 4 bursts in the 9.6 minutes
- bottom: strongest bursts.
- Page2: spectra around the strongest burst. polA,B
- bottom: stw strength vs delay
- peak around 1usec and smaller bump at 2 usecs (double
bounce
tx6 430 dome receiver. Strength of harmonics
the plots show the 430 dome signal while
transmitting with tx6 (.ps) (.pdf)
- page1: total power vs time
- computed total power 421.8 to 443.1 Mhz
- black is polA, red is polB, green is the xmitted power in
KW.
- the dashed blue lines are blown up in the bottom frame.
- we saw no bursts in the 9 minutes we took data with 430 Mhz.
- Page 2: spectra and harmonics
- top: spectra around largest total power spike
- you can see the 4 harmonics of 5.125 Mhz
- bottom: strength of power in the 4 harmonics channels vs
time/
- each color is a different harmonic
- the light blue line shows the transmitted power in Kw.
- the harmonics don't seem to get larger with higher xmitted
power.
the last plot is a dynamic
spectra of the 9 minutes of gr430 data (.gif)
- the dashed lines at the bottom show where the harmonics occur.
- at 17.4 we started to raise the power (starting at 2KW)
- the strong harmonics are mixing with other rfi in the band
Summary:
- we transmitted 100 KW with tx1 on antenna. There were only 3
or 4 bursts in 40 minutes
- tx4 would not come out of filament delay
- we transmitted 100 KW with tx6 on antenna while taking data on
alfa
- only 3 or 4 bursts. max strength 3*Tsys during the 9.6
minutes of datataking.
- we transmitted 100 KW with tx6 while taking data with the dome
430
- the harmonics got to 500*Tsys (in 1 channel)
- the harmonics did not seem to increase with power once we
got the power above 10KW.
- we didn't see any bursts in the 9 minutes
processing:x101/150602/hftest_1.pro,
hftest_2.pro
150603: tx4,tx3 on antenna (top)
We transmitted on tx4 and then tx3 while on
antenna. The setup was:
- az,za: 360,15
- alfa beam 0
- mock spectrometers: .1 sec spectra, 172 Mhz centered at 1450
Mhz, 8192, channels
tx4:
The plots show the alfa data
was transmitting with tx4 (.ps) (.pdf):
- At 50 KW, tx4 kept shutting down with reflected power fault.
- I stopped recording the xmitted power values after 15. 3
hours.
- We continued debugging looking at the front panel
outputs of tx4 on the 1 ghz scope.
- We could see the bursts on the rf output (through the 70Mhz
hipass filter).
- We didn't see them on the Anode output of the final power
stage.
- We did see a 4 ns glitch every power 5.125 Mhz cycle..
this did not make it through the output filter stage
- At first we thought this pointed to the output filter,
matching network as the source of the bursts.
- After inspecting the antenna, we saw that it had been
arcing...
- page 1: total power vs time
- after 15.53, is stopped recording the xmitter power
changes.. it was probably off till 15.75.
- The bursts around 15.85 were also seen on the oscilloscope
monitoring the tx4 front panel outputs.(rf output sample)
- This is where the arcing started at the antenna. It was
continuous until 15.3 hrs (we must have brought down the
power).
- The reflected power was coming from the arcing,.
- bottom: blowup of burst at 15.32
- Page 2: spectra and stw amplitude vs delay for the 15.32
burst..
- Page 3: blow up of arcing
- top: total power vs time during arcing.. it was continuous
- middle: spectra at start of arc
- bottom: stw amplitude vs delay
- the spike at 34 Ft may be the height of the antenna above
the dish
- primary wave dipole to dome
- delayed wave: dipole, dish, dome
tx3:
The plots show the total power vs time
while we transmitted with tx3 on antenna (.ps) (.pdf)
- page : total power vs time 1390:1520 Mhz
- we transmitted up to 100 KW
- there were not bursts.
Summary
- tx 4 had reflected power faults around 50KW.
- we saw the bursts on the front panel outputs using the scope
- We had arcing at the antenna.
- The bursts, reflected power probably came from the arcs.
- tx3: transmitted 100 KW with no bursts.
processing:
x101/150603/hftest_1.pro,hftest_2.pro
<- page up
home_~phil