---

History

• 14mar11: updated pointing model to correct for az,za offsets in 800 mhz rcvr.
  
• 11mar11: 2 screws mounting feed to under the floor were loose. This caused the pointing error vs za variation.
• fixed 11mar11 afternoon.
  
• 24sep10: 800 Mhz re installed after ao40 lowered from dome.. byu antenna version 1.0 installed.
  
• antenna versions (from ganesh)
  
• version 1: twin 50 ohm dipoles with a passive combiner adding to the insertion loss
• version 2: dipoles of 100 ohm impedance, simply connected with a T junction eliminating combiner loss. Supposedly V2 also did  a better job with the coupling between the dipoles.
  
• 14apr10:byu feed with both pols installed. (ant version 1) cold 23apr10.
  
• 22dec09: penn state feed replaced with byu single pol feed for testing.
  
• 06oct09: dana wiggled cables input to dewar, caused Tsys polB to now be lower than Tsys PolA (see tsys plots)
  
• 12aug09:  tv chan 67 789.25 turns off.
  
• 29jun09: installed cooled receiver.
  
• 15jun09: replaced cables to dewar with good cables, installed filter after dewar, retuned dipoles. Switched to 1-2 ghz as default IF
  
• 12jun09: installed new feed, filters before dewar. dewar is room temp still.
  
• 30mar09: update az offsets in model for turret position at 188.5. After updating model az offset i did a single cross (2strips) and used the az,za offsets to update the constant terms in the model.
  
• 18dec08: first spectra scan with interim correlator and 25 MHz filter.
• 05sep08: az strips at different turret positions to find correct turret center.
  
• 15aug08: installed helix antenna.

---

Calibrations

Miscellaneous:

08mar11: dual pol bandpass. rms, and  sefd vs freq

• source B0540+187 (3.32 Jy @ 750 Mhz)
• 60 sec on,off position switching with 10sec cal on ,off.
  
• za=14.
  
• mock spectrometer 172 Mhz with 8192 channels, 1 second integrations.
• the data used was limited to the 100 Mhz bandpass of the 800 Mhz rcvr.
• Pointing runs show a za offset of up to 25 asecs. This should not affect the gain much (beamwidth=390 asecs at 750Mhz).

• Top: The 60 second average bandpass for the off source position.
• red is pola, green is polB
• The Y scale in degrees K.
• I used the cals to scale the bandpass to kelvins.
• I didn't divide by the band pass shape (since that's what i'm showing) so the frequency structure is coming from the filters, etc.
• The True Tsys will be the average over the entire band.
  
• The ripples at the left side of the bandpass are coming from the cavity filter in front of the dewar (to keep out the tv station at 692 Mhz).
  
• PolA has more frequency structure than polB.
  
• This was also seen in the A2524 double position switching data  back in 30sep10.
  
• In this case the 1 Mhz standing wave from bowl to rcvr was much larger for polA.
  
• Middle: This plots the rms/mean along each channel for the 60 1 second records. The black dotted line is the expected rms/mean (from the radiometer equation). The ratty bandpass looks to be stable in time.
• Bottom: The SEFD vs frequency using 3.32 Jy for the source flux (at 750 hz).
• This was computed as posOff/(posOn-posOff)  * srcFluxJy (or TsysDeflection/ (srcDeflection/srcFlux))
• The ripple on the left of the sefd remains from the cavity filter
• The off position Power is mainly coming from the dewar (it does not pass thru the cavity filter).
  
• The on position has much more power going through the cavity filter.
• Because of this, the off/(on-off) does not cancell out the cavity filter ripple.

26sep10: calib run with byu antenna

• the cal values may have changed slightly since a cal cable was replaced on installation.
  
• Page 1: gain,tsys,sefd, bmwidth
• page 2: coma,sidelobe, eta mainbeam, eta main beam plus first sidelobe.
• page 3: pointing error
• za has a -22.5 arc second offsets.
  
• This could be from the feed mounting
• It could also be from the platform tilt from the platform reinforcement. In this case we would expect a 1 az variation in za an az. We don't see this.. the two az sections: -80, and +80 differ by 160 degs so a 1 az term amplitude show just about change sign between these two points.. but the don't so it's probably a mounting problem.

11jan10 single pol BYU antenna gain,tsys,sefd,beam testing (top)

• Page 1: the az,za track across the dish for BYU feed
  
• Page 2: gain,Tsys,sefd,BeamWidth for the BYU feed
  
• The cal values used were the same as the old antenna (measurements by ganesh 13aug09)
• Page 3: gain,Tsys,sefd,beamWidth for the Penn state feed for 06oct09->20dec09 data.
• This was limited to the same freq ranges as the BYU data.
• The Tsys vs Freq for the Penn State feed differs from the BYU feed
• BYU    : 780,755,710,730 (min to max Tsys)
• PennSt: 730,710,780,755
• The order is just about reversed. The spillover vs frequency may be different.
• (this assumes that the cal values remained the same when the antennas were changed).
  
• Page 4: coma,1st sidelobe,beam efficiencies  for the BYU feed.
• Page 5: pointing error for the BYU feed.
• There is a large az offset. This may  come from only having 1 source. The pointing error looks a little different from the penn state feed pointing errors.
• Page 6: Compare beamwidths for BYU and PennState feeds.
• Each frame is a different frequency (710,730,755,780).
• Black is the BYU feed, red is the penn State feed.
• The BYU feed beamwidths are smaller than the penn state beam width by about 4%. The BYU must illuminate more of the dish (or it's taper must be different).
  

• The BYU feed has narrower bandwidths than the Penn State feed (by about 4%).
• the Tsys vs freq increases in the opposite direction for the penn state (low to HiFreq) vs the BYU feed (hi to low freq).
• The illumination vs frequency (spillover) may be different.
  
• the SEFD for BYU is smaller than the penn state SEFD:
• The smallest SEFD for BYU is 10.5 Jy (755 Mhz).
• The smallest  SEFD for  penn state is  11.5 to 12 Jy.

12oct09 800 Mhz receiver band pass:  (top)

• 300 seconds of data averaged
• tracking blank sky at za=15 degrees
• mock spectrometer with 172 MHz bandwidth and 8k channels used

• PolA is black, polB is red.
• Top: the band passes normalized to unity (Tsys).
• A baseline was fit to each spectra (linear + 22nd order sin/cos). It is shown in green
• Bottom:  Bandpass/Fit.

• 750 MHz: This is the  dc component of the complex mixers.
• 766.25,769.75: low power channel 63.
• 789.25/792.83/793.75: channel 67 ..the station at west end of island is off the air.
  

17sep09: 800 Mhz rcvr PolA, polB ratios jul-sep09  (top)

• I only used sources with < 1% polarization (at 1400 MHz) so the polA,B ratios were not skewed too much by polarized sources.
• Color is used to differentiate between sources.
• Symbols are used to differentiate between sources.
• Page1: Tsys,Gain, and SEFD for stokes I (all frequencies).
  
• Some of the spread is caused by the variation with frequency.
• Page2: PolB/PolA Tsys,Gain,SEFD (all frequencies).
• The Tsys and gain ratio will be affected if the cal values are incorrect. The SEFD is not affected by the cal values.
• 02-06jul09 the sefd ratio is close to 1.
  
• 10aug-> 14sep09 the sefd polB/polA is greater than one.
• some of the variation is caused by the variation in frequency.
• Page 3: PolB/polA Tsys,Gain,SEFD 730 MHz.
• the 730 Frequency was used for the entire stretch of time (some of the other frequencies were switched when the low power tv station 289.25 turned off).
• From 10aug09->14aug09 SEFDB/SEFDA is greater than 1.2.

• 02-06jul09 sefdB/sefdA is close to 1.
• The new cal coupler was supposedly installed 03jul09 so this is probably not the problem (unless the date on the receiver log page i wrong).
• The cal values were changed on 03jul09. This caused the Tsys,Gain jump 02jul to > 02jul09 (purple) this had no affect on the sefd since the sefd is independant of the cals.
• 10aug09->14sep09 sefdB/sefdA is greater the 1.2
• There was a dual beam 430 run from 17jul-26jul09.
  

25aug09: azswing looking for chn 67 peak after chn turnoff.  (top)

• az spin at .38 deg/sec and za=18.
  
• mock spectrmeter sampling at 172Mhz bw, 8192 channels and .1 seconds.
  

• the plots are power in a single channel (21khz) vs azimuth
• black: 289.25 video carrier
• red: 793.75 audio carrier.
• green: baseline taken as a range of freq below the video carrier.

10aug09: beam squint in the 800 Mhz rcvr (top)

• The beam squint amplitude in asecs is the maximum value of the beam squint when the azimuth is pointing in the beam squint position angle direction. The beam squint for a given azimuth is:
• beamSquintAz=beamSquintAmp*cos(az - beamSquintposAngle)
• The different colors are taken at various frequencies.
• It looks like the major beam squint pointint error is in za (along the azimuth arm).
• The low freq beam squint 710 Mhz is large.

23jun09: rfi in the 800 Mhz rcvr  (top)

• 14jun09: an az swing 270->630 degrees was done to see which direction the rfi peaked in.
  
• Rfi with the same az peaks are probably coming from the same source.
• This was done before the filter was installed after the 1st amp.
  
• 18jun09: freq hopping.
• Data was taken with the band centered at 750 and 753.7 Mhz.
  
• If the rfi was being created by a mixer harmonic, then changing the lo by 3.7 Mhz should move the rfi by a multiple of this.

az swing on 14jun09:

• az 270 -> 630 degrees. az velocity about .38 degrees/sec.
• za=18 degrees and fixed
  
• 172 Mhz band centered at 750 Mhz.
• Sample data at .1 seconds.

• The vertical scale is -90 to + 270 degrees of azimuth.
• polA and polB have been averaged.
• the large triple rfis are video, color, audio carries for tv stations:
• 765.25 : chan 63
• 789.25 : chan 67
  

17jun09: ripple appears in polB (top)

• 15jun09: ganesh, stephen tune dipole
  
• 16jun09: 16:46 no ripple in Polb
• 17jun09: afternoon stephen tightens some cables in 800 Mhz system.
  
• 17jun09: 16:05 ripple seen in polB
• 18jun089 -> the ripple remains in polB.

• red is pola, green is polb.
• Page 1:
• top. 16jun09 600 second average spectra (16:38 ast)
• bottom 17jun09 90 second average spectra (16:04 ast)
• Page 2: blowup of 17jun09 average spectra.
• the spacing is .5 to 1 mhz (1-2 usecs). This would be a long distance if it was a reflection.
  

16jun09: calibration run 800 MHz receiver. (top)

• updated pointing model in az by 10asecs to remove final az error offset
• turret at 187.05
• dipoles were retuned 15jun09 afternoon, cables replaced.
• Switched from 260 IF to 1-2 ghz IF.
• Still using the room temperature receiver.
  

• On 16jun09 evening i switched the 4th band from 790 to 775 to see if the rfi was less... (it wasn't)
• On 17jun09 evening i switched the 4th band to 792 with 3 mhz bw..
  

• Page 1: average spectra from first pattern.
  
• 710 and 730 are 25 MHz bandwidths
• 740 and 790 are 12.5 MHz bandwidths
• 790 had a tv stations () in the band causing interferences
• Page 2: average spectra from first pattern.
  
• 710 and 730 are 25 MHz bandwidths
• 740 and 790 are 12.5 MHz bandwidths
• 775 had a tv stations (765/769) in the band causing interferences
• Page 3: gain,Tsys,sefd, beamwidth.
• The blue 790/775 results jumps around from the rfi.
• Page 4: coma, sidelobe height, main beam efficiency, main beam + first sidelobe efficiency
• Page 5: pointint errors.
• I've ignored band 4 (790/775 MHz)
• still have a small offset in az.
• Page 6: az,za coverage of these strips.
  

13jun09: tv station turn off date, rfi, and calib. (top)

•  az,za fixed at 285/10.
• 172 MHz centered at 750 MHz was taken with the 260 MHz IF (the band was centered at 325).
  
• .1 Seconds integrations were made.

• Page 1: bandpass
• Top: avg 60 seconds. red=polA,green=polB. This is a linear scale where the largest birdies go offscale.
  
• The low power tv channels 65 and 67 are flagged at the bottom of the plot (video and audio carriers).
•  Looks like they are transmitting but there isn't much modulation.
  
• The birdie at 750 is DC.
  
• The birdies at 730,734 and 711 are probably mixer images of the tv stations.
• The band is supposed to fall off below 700 MHz and above 800.
  
• On the spectrum analyzer it falls off above 800 but is pretty flat down to 600 MHz.
• The fall off seen in the plot  is probably from the baseband fitlers of the pdev spectrometer.
• The low freq junk is probably from the other sideband of the mixer.
  
• the 720 bump coming from the other side band is: 750 + 325 + 325 + (750-720)= 1430MHz.
  
• If the filter was letting these freq thru, then it would explain the increase in system noise.
• Middle: A db plot of the bandpass. The tv stations are about 35 db above the noise floor (rbw=172./8192=21 khz).
• Bottom: rms/mean be channel for the 60 seconds of data. The measured rms : .023 is close to th expected  value: .022.
• Page 2: blowup in frequency of the bandpass (db scale).

• The setup was:
•  25 MHz bandwidth centered at 740 MHz was used.
• The turret was at 188.5 degrees.
  
• 40 K was used for the calvalue
  
• Page 1:
  
• gain=6.1
• Tsys=186
• Sefd: 31
• BeamWidth: 378 asecs
• Page 2: coma,1st sidelobe, beam efficiencies. The sidelobe is low probably because the source was not strong enough for it to be seen.
• Page 3: pointing errors:
• az=-142 Asecs
  
• .. this is because the attenna was mounted in a different spot from the helix that had been there.
• za=142 asecs.
• The model used the position from the old helix. The model offset had been +72 asecs. All the other feeds have an offset of -70 -120 asecs.. So this error puts the antenna closer to the other feeds.

• low power tv channels 65 and 67 have not turned off.
• the power in the low part of the band extends down to 600 MHz.
• the za pointing error brings the model offset closer to the other feeds.
• The azimuth error is because the turret position has changed.
• TO do: track a strong source at different turret positions and match the sidelobes on each side of the main beam to find the correct turret position. I did this on 13jun09 but the source was not strong enough ..

13apr09: 2 az swings looking at 820MHz band (top)

• azswings 270-630 and back at .38 deg/sec
• za=19 degrees
• 800 MHz test receiver (not cooled) in dome was used.
• The tunable 30  MHz filter after the preamp was set to 820 MHz.
• 172 MHz Spectra and 8192 channels (rbw=21 KHz) were dumped at a 1 second rate with the pdev spectrometer.
• Data outside the 798-842 MHz range were ignored.

• The images show dynamic spectra in a az vs frequency plane.
• The top image is the clockwise swing.
• The bottom image is the counter clockwise swing.
• The dtv station at 800-806 MHz is channel 69. It will move to channel 12 after 12jun09.
• The 824 to 849 band is the cellular phone uplink (phone to tower).
  
• 828-835 has shows the 200 KHz gsm channels that are used.
•  It looks like many channels are being used at once. This is probably the frequency hopping  in gsm that is going on.

• Page 1:
• Top: peak hold spectra for the two az swings.
• Black is the CW spin, red is the ccw spin.
• The y Axis is db above the median spectral value. The rbw was  21 KHz.
• The digital tx channel gets to about 40 db above the average Tsys.
  
• Bottom: the mean spectrum for each az swing.
• Page 2: Power versus azimuth:
• Top  The azimuth dependence for the median tv station power.
• The y axis is the median dtv value/ median spectral value.
  
• The main peak is at az=22 deg.
  
• Smaller peaks occur at az= -59,-9, and 52.
• Middle: 818.81 Peak strength for 818.81 vs az (rbw=21KHz)
• Bottom: 819.14 Peak strenght for 8190.14 vs az (rbw=21khz)
  

Summary:

• 800-806: Tv station 69 will move out of the band after 12jun09
• 810-824 MHz
• These had time variable bursts. The strongest was 818-820 at 45 db above Tsys.
• I don't know whether this is representive of this band because of the time variability.
  
• Cellular phones: 824-849
• These did not get above 15 db above tsys in a 21 KHz band.
• Since this is the mobile to tower, a cell phone closer to the observatory will be stronger.
• i don't know how far away the cell phones in the data were from ao.
• From the data we saw, you probably want the filter to be 30-40 db down at 820 MHz.
  

11apr09: chan 63,67:low power tv stations in the 700-800 MHz band  (top)

08apr09: azSwings for chan62/50:Where to put the 800 MHz filter. (top)

• Cut out the tv stations below this band
• Cut out the cellular phone band 824-849 MHz (cell phone to tower uplink). 

• chan 50 (686-692 MHz).
  
• This channel is currently transmitting in digital as channel 62 (758-764 MHz)
• chan 51 (692-698 MHz). We currently see the analog version of this station at chan 52. The spectral scan of 18dec08 saw the video carrier for chan 52 (699.25) 7 db above the noise floor (this may not have been the max since it was a fixed azimuth). The transmitter is about 60 miles east of AO.
  

• azswings 270-630 and back at .38 deg/sec
• za=19 degrees
• 800 MHz test receiver (not cooled) in dome was used.
• The tunable 30  MHz filter after the preamp was set to 760 MHz.
• 172 MHz Spectra and 8192 channels (rbw=21 KHz) were dumped at a 1 second rate with the pdev spectrometer.
• Data outside the 740-780 MHz range were ignored.

• The images show dynamic spectra in a az vs frequency plane.
• The top image is the clockwise swing.
• The bottom image is the counter clockwise swing.
• The tv station around 760 is digital channel 62. This will move to channel 50.
• It has a broad peak near az=0 degrees.
  
• The tv station near 766 is analog channel 63. It is a low power station from aguada (1.79 km south of ao and 42 km west).

• Page 1: Tsys in Kelvins before the azimuth swings.
• The cal was toggled for about 20 seconds. 40 K was used for the cal value.
  
• Black trace: za = 11 degrees. 122 Kelvins
  
• Red trace: za = 19 degrees.  129 kelvins
• The za spillover 11 to 19 degrees increased Tsys by about 7 Kelvins.
• Page 2:
  
• Top: average spectra for the two azimuth swings.
  
• The y axis is spectral density in db's above the median value. For the analog stations the RBW = 21 KHz.
  
• The average value between the dotted red lines (channel 62) was computed for each second of the two azimuth swings.
• analog channel 63 (765.25 Video, 769.75 audio) has the video carrier 44 db above the average Tsys value.
  
• Middle: Average spectral density for Channel 62 vs azimuth position.
• The units are db's above the median Tsys.
• The black plot is the clockwise spin, the red trace is the counter clockwise spin
• The az peaks are at: -60, -1, 26, and 52.2. They repeat for the two swings.
• The average spectral power for channel 62 is 35 db above the median Tsys value.
• Bottom: Chan 63 Video carrier pk values vs azimuth (rbw=20KHz)
• The azimuth peaks are at -16,32.7 and 56 degrees. The two swings show the same values.
• The peak value in 21KHz bw is > 50 db above the noise floor (also using 20Khzrbw). It looks like it saturated.
  

Summary:

• chan 62/50--> bandpass filter needs to cut out the 35 db at 692 MHz.
• chan 51.
  
• If the power levels remain the same in the switch 52->51 then this should not be a problem. This assumes that chan 52 is currently broadcasting at full power. It certainly doesn't look like a 275 kw erp transmitter... Probably need to check this out some more.
• chan 63:
•  It was mentioned that all low power stations will move out of the 700 MHz band after 12jun09. We need to verify this. There have been no fcc filings stating that they will stop transmissions on 12jun09 (chan 52 has done this).
• the -16 deg az does not point at the transmitter. The station headquarters is in camuy. az=-16  degrees does point toward camuy (maybe this is just a coincidence).
  

02apr09: Emergency services usage of 700 MHz band (top)

• 763-775 (12 MHz)
• 793-805 (12MHz)

18dec08: spectral scan 620-890 MHz. (top)

• az,za=185,10 degrees. The turret was positioned to 188.5 degrees.
  
• the cal was winked on/off at one hz. during the data taking.
  
• the 25 MHz filter was tuned to a position and then data was taken for a few minutes. It was then moved manually to the next position.
• The interim correlator was used with 1024 channels, 25 MHz bandwidth, 9 level sampling, and 1 second dumps.
  
• The antenna is a helix that receives 1 circular pol.
  

• The Top plot is the 832MHz band, the bottom plot is the 839 MHz band.
• The processing was:
• Try to throw output bad frequency channels by fitting a polynomial,sine wave to The median(calon/calOff) the removes the filter bandpass.
• compute the median power over the good channels (also throwing out about 10% on the edges).
• compute (calOff)/(calOn-calOff) and take the median over the N samples. Multiply the median cal off spectra by this value to get Tsys in Cal units.
• Tsys is about 3 cal units. The cal is about 40K so Tsys is around 120K.

• (also see tvFreq)
  
• 635 MHz band
• 627: channel 40: video carrier  (pbs)
  
• 635: channel 41: digital station
  
• 639: channel 42: video carrier
• 660 Band
• 648: channel 43: digital
• 663: channel 46: video carrier
• 685 band
• 683: channel 49: digital
• 687: channel 50: video carrier
• 710 band:
• 703: channel 53: digital (yiyi avila...)
• 711: channel 54: video (yiyi avila..)
• 735 band:
• 723:channel 56:
  
• There is no license for chan 56 in pr.
  
• video carrier should be at 723.25, audio:727.75
  
• This may be intermods created by channel 54 below..
  
• 747.25: channel 60: video carrier.
  
• This is the arecibo station.
• Moving to chan 14 after 12jun09
• 735.25/739.75: chan 58 video/audio
• moving to chan 48 after 12jun09.
  
• 760 band:
• 751.75: audio carrier for channel 60. moving to chan 14 after 12jun09
  
• 761: channel 62 digital. moving to channel 50 after 12jun09
  
• 765.25/769.75: channel 63 audio/video carrier.
•  the audio at 769. is stronger than the video.
  
• low power tv station serving aguada. Office in camuy. see fcc info (.ps)
• 771.25: chan 64 video carrier. moving to chan 18 after 12jun09
  
• 785 band:
• 775.75:channel 64:  This is the audio carrier. moving to chan 18 after 12jun09
  
• 789.25/793.75: channel 67 video carrier/audio carrier
  
• This is a lower power tv  station 18.5Kw 43 km due west of AO. see fcc info (.ps)
• 795.25: chan 68 video. moving to chan 49 after 12jun09.
  
• 807 band
• 803: channel 69: digital. moving to chan 12 after 12jun09.
  
• 832 band, 839 band:
  
• 824-849 cellular phone uplink band (handheld to tower)
• 850 band:
• Not sure what this is unless it's the spill over from the downlink.
• 875 band:
• 869-894 is the cell phone down link

Things to check out:

• Check the signals 722.75-728. see if they are intemods..
• Low power stations to check on:
• 
  

05sep08: looking for the turret position. (top)

• track the source
  
• Offset the turret by a fixed amount. Offset the azimuth position to compensate for this.
  
• Do an azimuth strip using this turret position.
• Look at the sidelobes and see if they are equal.
• Repeat this for  7 positions about the turret center.
  
• When the turret is centered on the tertiary the sidelobes should be symmetric.

• rf frequency 1050 MHz
• interim correlator  1 second sampling with 25 MHz bw.
• Do strips at  187.2 + [-3.-2,-1,0,1,2,3] turret degrees. Repeat this twice.\
• Do strips at 188.45 + [-1.5,-1,-.5,0,.5,1.,1.5] turret degrees.

• Page 1: az strips thru the source (linear).
• There are 7 az strips in each frame. The center is flagged with the green line.
• The top two frames have 1 degree steps in turret centered at 187.2
• The left sidelobe is larger for 184,185,186 turret degrees.
  
• the bottom frame has .5 degree steps in turret centered at 188.45
• Page 2: az strips thru the source (db scale).
• These are the same plots as page one but using a db scale.
• around 188 turret position the sidelobes start to balance out.

• Data was taken with the az off=-346.01.
• 1 turret degree moves the pointing by 45 asecs/deg (need to check the sign..)
• moving from 182.7 to 188.5 should move the offset by 261 asecs (which direction ???).
  

home_~phil