Initial stuff ------------- z>6 QSOs - March 2005 - optical z's - need MgII z's SDSS0836+0054 5.82 24.4 08 36 43.85 +00 54 53.3 i21.04 z18.74 SDSS1030+0524 6.28 24.4 10 30 27.10 +05 24 55.0 i23.23 z20.05 SDSS1048+4637 6.23 99 10 48 45.05 +46 37 18.3 i22.4 z19.86 SDSS1148+5251 6.35(.42) 99 11 48 16.64 +52 51 50.2 i23.9 z20.12 SDSS1306+0356 5.99 24.4 13 06 08.26 +03 56 26.3 i22.58 z19.47 SDSS1630+4012 6.05 99 16 30 33.90 +40 12 09.6 i23.4 i20.42 1411+12: z=5.93 1137+354: z=6.01 0840+56: z=5.85 1250+31: z=6.13 0002+25: z=5.80 0005-0006 z=5.85 1602+4228 z=6.07 1623+3112 z=6.22 J1436+5007 z=5.83 z>5.8 QSOs z=6 => 202.915MHz 6.2 => 197.279 6.4 => 191.947 6.6 => 186.895 --------- January 14, 2005 -- talked to Durand: tech staffing is for installation and testing of 1 module. includes 3 guys (1 runs cherry picker, 2 to crawl around) for 3 days for initial installation. the 2 guys for 1st day of testing (1 to remove/install dipoles, 1 to run cherry picker). than a second day of testing. total is 13 man-days. and 2 days for good luck and we get 3 weeks. Could be that the 3 days for installation is over-kill, but steve also thought we'd want to do some bench testing, to make sure it works, then he is preparing for potential difficulties in installation, which might require machine-ing. and I think that we will probably need at least 2 days of testing, and perhaps a third. overall -- 3 man-weeks is probably about right for 1 element testing. With lincoln -- Review discussion see Design rev. notes check for antennas with holes impact -- 1st installation and tests: 3 tech-weeks, 2 CC-weeks, 1 Mert-week 2nd installations: 1 tech-weeks, 2 CC-weeks, 1 RP-week Full installation: 2ants/day, 2techs/ant => 28 tech-days = 6 weeks plus 1 week to fix problems? then testing full array: 2 CC-weeks, 1 RP-week observing time: like any large proposal jan 23: -got go-ahead from fred/jim for review -organize by me/lincoln --------------- Testing Notes - Carilli 2005 ---------------------------- March 3,4: with perley, hayward, mertley, carilli ant 6 - west arm pre-tests: set-up total power monitoring system check - power level, cal signal generator (0.1 Hz?) note: power monitor sends out records at 100/s, labview only reads at best 10Hz => get 1/10 of records (10%duty cycle) measured - K band stuff, tips and sources, hot/cold L band - effect of 4 band using hot/cold + sources + dip? taking down 4 band only takes a few minutes (hokey system) find: 2 - 3K from scattering, plus 5% or so efficiency lose cal signal is 12K? sources: must do E-W-N-S pattern for 'off' P band - tips are meaningless overall - can get to 0.1 K accuracy with TP system March 8,9,10 Perley, Merteley, Smith, Leiker (SAO), Kimber (SAO) devices show up in lab - intital test check noise cal temp, daisy chaining stuff with P band, ... to antenna - check bandpass with antenna out in field looks good. RFI - know sources - TV chan 13 at 210 MHz, some communication stuff at 176ish, 180 - 200 looks clean install boxes in barrel mcclindon, smith, merteley, leiker, kimberk, greenhill, carilli, perley mount feeds, cables - boxes first, then thread cable through tubes, attached from feed problem - caps screws - won't go in/out brass ruined? keep pipe from falling out. need to drill out to get space for SAO connector before mounting dipole -- REMOVE 4 BAND!! first potential tests: -check power levels through signal path from barrel to vertex room -total power measurements -efficiency - go to source -rfi - set up monitor -spectral sweeps at various places -might even try on/off on thursday? MOU - input frmo Lo, need input from Alcock. no timeline? dipole installation -- 1. box in barrel takes some time, and thread cables through pipes first 2. cherry picker - take off 4band!! takes about 1hr otherwise Notes from Rick: To convert SA scans x to usable units (Tsys...) and get Trx, conversion factor to Kelvin 1. Convert dB to linea: T* = 10^(x/10) 2. calibrate on hot load -using Thot*, Tcold* linear values from above -Tld = 275 -K = conversion factor for linear units -Tsky+spill = 10Kish Thot = Tld + T_rx = KThot* Tcold = Tsky+spill + Trx = KTcold* 3. If you know cal diode value in real K, then falls out more easily. 4. can't do hot load at <= P band March 10 -------- check bandpass with antenna hooked up: RFI!! narrowish 20db + = TV stations 181 - cha 8 video 185.5 - chan 8 audio 186.5 = chan 9 video 192 - chan 9audio 193.5 - chan 10 video 197.8 - chan 10 audio 199 - cnah 11 video 204 - chan 11 audio other birdies, but probably off at night except every 10MHz - internal biggest problem - 186 - 192 MHz broad band RFI 10db ugh probably not TV related Clear range: 193.5 - 199 (except for 2 TV birdies) overnight monitoring: TV all constant broad band - varies broad band - not seen in sysquick Nov 14, 2004 (but sensitivity may be very difcferent) Balance - Rx looks ok relative to P, 4 band, for power/BW perhaps few dB low -- find 13.8 V in barrel. need 13.1 to drive amp min, want 15V, when heater comes on could be problem - easiuly fixed to get us where we want to be wider SA sweeps - lots of stuff down to 160. ouch tests for today: - test with VLA correlator - 3MHzbandwidth - see if backend filters are good enough to avoid ringing? -10deg sweep in AZ using W8 monitor -inject reference tone ? -noise floor/sys 'eff' - Cas A test plots needed - W8, SA, sysquick - last 150 - 200 MHz got list - all z> 6 QSOs 9 at z>6, evenly distributed up to 6.42 ------------------------ reduce single dish data - need 2 ants to get into correlator, but xcorr data doesn't have to be useful (eg. can use dead ant) Scan Source Qual Calcode Sub Timerange FrqID START VIS END VIS 1 3C461 : 0000 1 0/17:57:30 - 0/18:00:50 1 1 1920 Source summary Velocity type = ' ' Definition = ' ' ID Source Qual Calcode RA(2000.0) Dec(2000.0) No. vis 1 3C461 : 0000 23:23:25.8000 58:48:53.400 1920 ID Source Freq(GHz) Velocity(Km/s) Rest freq (GHz) 1 All Sources 0.1941 0.0000 0.0000 IF( 2) 0.2006 0.0000 0.0000 Frequency Table summary FQID IF# Freq(GHz) BW(kHz) Ch.Sep(kHz) Sideband 1 1 0.19410000 12500.0010 390.6250 1 2 0.20060000 12500.0010 390.6250 1 LISTR2: Appears to have ended successfully --------------- can Source Qual Calcode Sub Timerange FrqID START VIS END VIS 1 3C461 : 0000 1 0/18:01:20 - 0/18:15:00 1 1 6238 2 3C461 : 0000 1 0/18:15:20 - 0/18:24:20 1 6239 10646 Source summary Velocity type = ' ' Definition = ' ' ID Source Qual Calcode RA(2000.0) Dec(2000.0) No. vis 1 3C461 : 0000 23:23:25.8000 58:48:53.400 10646 ID Source Freq(GHz) Velocity(Km/s) Rest freq (GHz) 1 All Sources 0.1894 0.0000 0.0000 IF( 2) 0.1959 0.0000 0.0000 Frequency Table summary FQID IF# Freq(GHz) BW(kHz) Ch.Sep(kHz) Sideband 1 1 0.18941250 3125.0002 24.4141 1 2 0.19591250 3125.0002 24.4141 1 ------------- problem - flagged all data? cparm 0 1 16 - turn off corr flags possm 0 0 0 0 0 0 1 1 0 ante 6 IF 1 = broad band crud - ugly timer 18:21 - 18:24 - on 3C461 - lower RFI, can see 190MHz birde timer 18:10 - 18:17 - stowed - higher RFI (factor 2-3) azimuth tests - every 10-20deg at 10deg elev broad band RFI - strong signal to NW, goes below noise floor (down by 6dB) at other az's Cas A - on/off efficiency continuum 3 MHz at 196 (clean) want - AIPS sweep of band at 3.25 MHz, 1 pol SA sweep in AZ at 10 deg elevation: big rfi - blocked at -40 AZ ------------- Sweep of band 4PM: fluke settings in obs file 1.56 MHz BW observe file w. proper LO settings: VLATEST/VHF.Mar10.obs track CasA start (4:16MST) - 23:17 IAT on Cas A 1min per freq, 10s int, 512 channels, 1B mode, 306 vis, takes 20min ID Source Qual Calcode RA(2000.0) Dec(2000.0) 1 CASA : 0000 23:23:25.8000 58:48:53.400 ID Source Freq(GHz) Velocity(Km/s) Rest freq (GHz) 1 All Sources 0.1846 0.0000 0.0000 Frequency Table summary FQID IF# Freq(GHz) BW(kHz) Ch.Sep(kHz) Sideband 1 1 0.18463125 1562.5001 3.0518 1 2 1 0.18613125 1562.5001 3.0518 1 3 1 0.18763125 1562.5001 3.0518 1 4 1 0.18913125 1562.5001 3.0518 1 5 1 0.19063125 1562.5001 3.0518 1 6 1 0.19213125 1562.5001 3.0518 1 7 1 0.19363125 1562.5001 3.0518 1 8 1 0.19513125 1562.5001 3.0518 1 9 1 0.19663125 1562.5001 3.0518 1 10 1 0.19813125 1562.5001 3.0518 1 11 1 0.19963125 1562.5001 3.0518 1 12 1 0.20113125 1562.5001 3.0518 1 13 1 0.20263125 1562.5001 3.0518 1 14 1 0.20413125 1562.5001 3.0518 1 15 1 0.20563125 1562.5001 3.0518 1 possm - TV and other birdies reproduce at exact channels => frequency conversion to correlator is OK. TV stations can 'saturate', such that drop to 0 (quantization/ACL?) and they can ring over +/- 0.6 MHz even after hanning smoothing can ring into other settings if <= 0.3MHz or so from edge? broadband RFI - looks like spead specctrum radar - clear modulation with frequency .ugly. starts 186.2 - 191.9 CasA efficiency test - failed due to weak cal signals and strong RFI - full band is getting into correlator? or correlator baseband filters not enough to mitigate? need front end 3 MHz filter - ask ridgeway to build by wed. to do next week: 1. check rfi at night 1a. find source of broadband RFI? 2. effic test 3. first fringes 4. L/P on/off test RFI mitigation options - -find broad band source -filters around TV stations -observe with narrower bands (hunt and peck) -lower 90deg sidelobe via directional element -obs at specific az - not effective --------------- greenhill web site: cfa-www.harvard.edu/~lincoln/vlavhf --------------- broad band RFI 186- 192: http://svartifoss2.fcc.gov/cgi-bin/ws.exe/prod/cdbs/pubacc/prod/sta_det.pl?Facility_id=84215 >> Here's your station: >> >> http://www.fcc.gov/fcc-bin/tvq?list=0&facid=84215 >> >> -W >> frmo Greenhill: Here is another link for KNMD DTV. http://www.fcc.gov/fcc-bin/tvq?list=0&facid=84215 My reading of fcc documents (tell me if this is yours as well) is that UMN has a 200W DTV station licensed (and using an antenna manufactured by Dielectric Comm.) and a construction permit for a co-located analog 316 kW station w/ an unspecified antenna. The DTV station does not have programming on air, according to the FCC - and my google search. Despite the low power, this must be what we are seeing in our monitor. The planned and current broadcast areas http://www.fcc.gov/fcc-bin/FMTV-service-area?x=TV233499.html http://www.fcc.gov/fcc-bin/FMTV-service-area?x=DT1018041.html show fairly nondirectional service area at a fixed received power. Presumably, we can notch filter the carrier for the analog station and pray that we can get close to a thermal noise floor. The DTV signature is a bigger problem obviously. I agree completely with your assessment that if we have to deal with DTV, an educational station as partner is the best of all worlds. I might imagine that the station programming will not be impinged if they turn off the transmitter at night. http://www.fcc.gov/fcc-bin/tvq?list=0&facid=60793 (KCHF, evangelical station in Santa Fe, 30 kW construction; 1.7 kW not seen by us but licensed.) DTV (KOFT) Channel 8 in Farmington has a service area footprint that is directed away from us w/ their temporary 2.4 kW station, but their 40 kW station under construction, will have a nondirectional pattern. --------------- March 16 Carilli, Perley, Greenhill, Mertely, Kimberk, Blundell install second dipole on 17 - 1hr w. cherry picker, but still took most of a day with box etc... established rf signal in vertex room some tests on ant 6 - L band on measurement - RFI from WSMR 2 band - Cas A - Tsys looks like 200 K probably OK in Gal plane beam maps - holography mode get file from RP) fringes - no. why not? ant 6 says it is in L band talked to KNMD -------------- All, Following is a summary of my discussion with the station engineer at KNMD - UNM digital 186 - 192 MHz. The engineer is Jim Gayle at KNME/D - cell 505 379 3594, office 277 2049 Some general points - He said the broadcast is from Sandia Crest at 200 W. They plan to increase power to 1.5 kW in a year. - They broadcast 24/7, some PBS stuff. Under auspices of UNM. - They hope to move to IP datacasting of eg. UNM lectures for out-lying students. (why this needs to be 24hr/day is beyond me?) - He said the digital signal has roughly equal signal across the 6 MHz with modulation that looks like spread spectrum radar -- that's exactly what we see. - he said there is another digital station - chan 10 out of Jemez - no name mountain or something. they do not boradcast all the time. - he said UNM does 1.5 B$ of business in the broadcasting arena every year! - he said the fcc mandated that all analog commercial channels go digitial a few years ago, and turn off their analog, which would then be auctioned to eg cell phones - he said the only way they could shut off is if we made an agreement with the President or Board of regents of UNM. I do not think the last statement above is completely out of the question. UNM is making a major push into low frequency radio astronomy, and has expressed desire to have closer ties to NRAO, and they are hiring 2 professors this year to work on LWA. I think we have an opportunity for a deal. We will have to talk to the people in the department, and this guy Gilfeather, who is head of research of something for the university. It may be that we could get some of their students to even work on the fabrication of the devices, if that's what it took to make this work. cc --------- March 17 -------- 1. established cals on ant 17 and 6 2. L band on/off Cas A, hot cold Cas A - effic hot/cold - Tsys to get Tsys: - plot up data in TEST.DATA/MarchXX -get mean hot/cold -calculate using Tsys.f 3. reinstall 4 band RFI sweep with VLA at night March 17 (during AW605) - source = 1148+5251 -compare night vs day (10th) -compare 6 vs 17 -compare auto vs xcorr? -compare HA's large sweep - freqid = 15, 1B, 512chans prtuv - just like D10 display - amp, phase No Fringes 6-17: phases random amps wander from 0.1 to 8? to make plots - lots of individual, then data file with all sweep - then use SWP.com edit 1148.6.swp - remove header stuff, and replace FLAGGED with 0.0 0.0 in SWP.com - add arbitrary number to y (100) before taking log problem- lose alot of info on single plot Ante 17 - really ugly - lots RFI. impossible. ouch. Ante 6 - March 10 - definitely worse than at night, but night still isn't brilliant Ante 6 - other hour angles 1048 - 0948 - 0848 - 0748 - forget it note - March 10 - clean sweeps had mean value of like 50, and flat baseline. March 17, there is broad bump(s) in baseline, and mean value is about 1? Summary, All, I've looked at the VLA RFI sweeps from March 10 and March 17. Attached are some plots: CASA.6.ps: Sweep on Cas A for Ant 6 on March 10, day time 1148.6.ps: Sweep at 1148+52 for Ant 6 on March 17, night time 1148.17.ps: Sweep at 1148+52 for Ant 17 on March 17, night time 1048.6.ps: Sweep at 1048+52 for Ant 6 on March 17, night time 0948.6.ps: Sweep at 0948+52 for Ant 6 on March 17, night time 0848.6.ps: Sweep at 0848+52 for Ant 6 on March 17, night time 1230.XCORR - Xcorr 6-17 on Virgo A at 196.6 MHz 1230.17.PS - Autocorr ant 17 on Virgo A 1230.6.PS - Autocorr ant 6 on Virgo A Notes: - I've added 100 arbitrarily to avoid taking log(-x). - 1148 is a 'blank' field - The different sweeps have different frequency ranges, so be careful when comparing - I have used hanning smoothing off-line - the FITS data are in: ftp.aoc.nrao.edu/pub/staff/ccarilli - ignore the mJy scale on the AIPS plots - these are uncalibrated - The ant 6 - 17 baseline = 0.8 km A few comments: 1. Day/night comparison: it looks like some of the lower level 'hash' during the day goes away in the regions away from the TV stations, although for Ant 6 March 17 on 1148 there was something funny going on in some of the bands (low amplitude, large 'bump' in band - not seen on March 10 - see eg. the band around 185. These return to 'normal' later in the run). 2. Ant 17 shows crap throughout spectrum. not seen on Ant 6. unclear what is happening here. 3. Changing Hour Angle: the TV stations go up and down, but never away. I've played this HA game before at 140', and it is a losing battle -- it's just the sidelobes bumping up and down, but they never go away. 4. Virgo A - we get phase coherence across the 1.56 MHz band, which suggests that we have 'fringes'. The source flux density is 515+/-5%Jy at 408 MHz and 1100+/-50 Jy at 178 MHz, which means spectral index = -0.9 and S_196 = 1008 +/- 70 Jy. If our efficiency is 50%, then the rise in Tsys = 100 K or so. Note there is a significant ripple across the band. The individual autocorrelations show very different behavior - with ant 6 showing the large 'bump' and ant 17 showing 'modulation'. We don't see coherence in the Amp and phase of the chan 0 visibility. Note: Virgo Gal latitude = 74deg 5. Looking at some of the AIPS autocorr spectra, we may start seeing modulation even between the carriers due to the broader band TV 'signal' in the analog stations. Ugh. ---------------- Hi, Suggestions below. Refinement requested. Monday day -If 4m dipole on ant6 can be removed/replaced, then expanded activity on Monday is possible. Transfer tasks from Tuesday. -arrange logistics for 3 MHz filters if at all possible. -confirmation of fringes -check fringes w/ VHF LCP and RCP cables reversed -short term monitor FE AGC on ~ 1s time scales pointed toward Sandia -install 16V SAO power supply on antenna 6 in the vertex room. If another is available for ant 17, same. -restore 4m dipole on ant6 at end Tuesday day -pull down 4m dipole on ant6 & ant17 -ant 6 total power L-band tip and efficiency w/ VHF dipole in place -P/band efficiency, Tsys, beamcut w/o VHF short (ant6) -P/band efficiency, Tsys, beamcut w/ VHF short (ant6) -remove VHF short -VHF hot load measurement in prep for nighttime measurements ant 6 & 17 Tuesday night -interferometer map of VHF beam response to many tens of degrees off-axis (mid elevations pointed away from ABQ and the control bldg) -VHF cold sky, nighttime Tsys via total power -deep VHF interferometer integration Wednesday day -VHF efficiency and cold sky, daytime Tsys via total power Thursday day -put up 4m dipole Lincoln J. Greenhill ---------------- All: I'll talk with Ken S. when he comes in (shortly, I think -- although the roofers are busy on his house today, [to repair the Oct hailstorm damage] so that may be a distraction) about making use of his software time. If we can get it, I suggest: 1) On/off tests of antenna efficiency and Tsys, from pointing on Cas A, elevation dips, etc. (On-Off means the 4-band dipole here, as I'd like to compare these results to those I did 2 weeks ago without the 196 MHz system. With a bit of luck, these new results will be completely consistent with the old ones, which would argue for not having to take off/put on the 196 MHz feed). 2) Same, for 327 MHz. The above should take only a couple of hours. By far the most importanta activity (IMHO) is to get this filter issue resolved. Without a good filter, no useable interferometer tests can be obtained. I'll report back when I get the necessary information. Rick ---------------- Rick's report on effect on L band: Gentlemen: I've reduced the data taken last week on the 20cm performance loss caused by the 2-meter dipoles. The data were at first confusing, as the Y-factor from the hot/cold load (used to calibrate the scale) was far too small -- it was 8.22 dB, when it should be near 10. Furthermore, midway through the test -- when the antenna was moving to Cas A, the cold-sky power dropped by 1.54 dB. Fortunately, Bob Hayward was nearby when I was puzzling about this -- he pointed out the obvious explanation that the subreflector was clearly not in the right position during the hot/cold calibration -- a legacy of the last observation when clearly was done at a different band. This is an important factor to remember! Fortunately, the hot load value is not affected by the subreflector position, and the cold sky value can be taken from the (off) Cas A observation. When this is done, I can properly calibrate the data. The cal value comes to 2.3 K, compared to 2.2 K (rough value) determined from early March. Independent of the calibration is the G/T ratio -- and this is what we really want. From the early March `baseline', (with the 4m dipoles off), the Cas A antenna temperature increment, divided by the cold sky value is 5.82. From the data taken last week, with the 2-meter dipole on (and the 4-meter dipole off), I find 5.84. (Formally, a little bit better!). In fact the errors are larger than 0.02, so the result is consistent with there being no effect at all of the 2-meter feed on 20cm performance. For the record, I measure a 173 K antenna temperature increment due to Cas A from last week, vs. 172 from earlier this month. The cold sky temperature was 29.7 last week, vs. 29.5 earlier this month. So this is good news, I figure. Rick ------------------- March 20: sent rfi note to Gilfeather, meet 23? ------------------- March 21 - P band beam cuts - ant 6,17 vs. 7, 19 6,7 look narrower than 7, 19 => 'better' ------------------- March 22 perley mertely greenhill carilli sowinski Cherry picker - 1hr, bob + john 1.5 hr 0. remove 4 band 1. P band tests: w, w/o short -- beam cuts on 3c147, efficiency Cas A cuts w./o short: presence of 2 band REDUCES skirts to P band beam by few dB. ant 6 2. new ant 6 power supply: up to 16V to get 14.8 in barrel => better cal signal? -- no can't find cals. weak at P. strong at 4? 3. Night - if get cals, then go deep on Virgo A? also redo VLA spectral sweep - Ant 17? On/Off Cas A in 196.6 cont 1.56MHz w. S/A - power looks about right (3dB = 2x T_sys = 150 K) but... funny stuff at beginning and end - S/A filter off? interferometry - P band On/Off - funny source strength w/w/o short? CasA 2band beam cuts w. 6/17 as ref ants 3c147 RFI sweep deep virgo A - 2-3hrs 0542+498, 1230+124 ----------- cross correlation plots: Ant. 6-17 (all Hanning smoothed) VIRGOA.ID8.PS (196.6MHz) - 1 hour integration VIRGOA.ID8.PS (195.1MHz) 3C147.ID9.PS (196.6MHz) - 2 min integration 3C147.ID3.PS (187.7MHz) 1230.UVPL.PS (196.6, 195.1MHz) - phase vs time 1. Fringes on Virgo: 1.56 MHz spectra cross correlation in 'clean' bands (196.6, 195.1 MHz) look good. Also, phase vs. time shows coherence over 3hrs. Note that we altenated between 195.1 and 196.6MHz every 30 min. Virgo is about 1000 Jy. 2. 3C147 - about 100 Jy a. In clean band (196.6 MHz), see phase coherenece, but funny structure in both phase and amplitude -- looks almost like modulated signal, eg. the 'content' of a TV broadcast in-between the carriers. b. In digitial TV channel 9 band (187.7MHz), see excellent phase coherence, wrapping very quickly as function of frequency. This is clearly related to the modulated digital TV signal, and has nothing to do with the source. 3. Scalar vs. vector averaging -- lots of RFI in scalar, none in vector! ------------ Beam cuts - March 22 Cas A (elevation = 20-30 deg) //OF 0 -16 90 5.0 10 33 33 point raster at 90deg in Az 5.0 = spacing between points in beams = beam * (1/5.0), where beam = Lam/D (in rad or convert to arcmin) [apr18 - No - Step size = 0.69 /(nu_GHz * 5.0) - pretty much = below] 10 = timer per point in sec => 33 points * beam * (1/5.0) = 6.7 beam for full raster time = 33 * 10s = 330s = 5.5min per cut 10s = (1/5.0)*(Lam/D) = (1/5)*(1.53/25) =0.0122rad=0.70deg (beam = 3.5deg) timerange: 0 1 10 0 0 1 43 0 => ant 6 raster, 17 reference timerange: 0 1 45 0 0 2 15 0 => ant 17 raster, 6 reference Ant 17 as reference => ant 6 raster 25 SCANLR :40000 1 0/01:17:30 - 0/01:23:10 26 SCANDLUR :40000 1 0/01:23:20 - 0/01:29:00 27 SCANDU :40000 1 0/01:29:10 - 0/01:34:50 28 SCANDRUL :40000 1 0/01:35:00 - 0/01:40:40 Ant 6 as ref => ant 17 raster 30 SCANLR :40000 1 0/01:49:00 - 0/01:54:37 31 SCANDLUR :40000 1 0/01:54:50 - 0/02:00:25 32 SCANDU :40000 1 0/02:00:40 - 0/02:06:20 33 SCANDRUL :40000 1 0/02:06:33 - 0/02:12:10 can plot: freqi 1, bif 2 (or 1 - same) antenna 6 beam looks very very strange - multiple peaks? ant 17 looks ok in 3 of 4 dimensions could be: RFI? or ionospheric refraction? note Cas A = 23:23:25.8000 58:48:53.400 Sun = 0hr 0dec observed at sunset different stokes: no major effect, LL noisier, but basically same. Tsys - listr: RR looks OK, LL bad for ant 6 try plot Tsys for ant 17: good: scandlur bad: scanlr both look pretty good. the 'bad' one has some peak at beginning, but not seen in xcorr. no idea what is going on, except that the 'structures' we see are at the scale of the beam size - ionosphere? but Rick say's the P band phases were great. Ante 6 - T_sys looks basically OK for RR plot notes: -interferometric measurements on antenna 17 beam, with ant 6 as 'reference' (ie. cross correlation amplitudes with rastering of 17 while ant 6 stares at Cas A) -Since this is cross correlation amplitude, with ant 6 not moving, it corresponds to voltage pattern, so should be squared for power. -X axis scale: for this raster 1 milliday= 86.4s= 6.0deg, or each point corresponds to 0.70deg (1/5 beam) -I think Cas A is of order 1e4 Jy at 196 MHz -all rasters use 1.56 MHz band centered at 196.6 MHz ANT17.DRUL.PS - raster from down right to upper left ANT17.DU.PS - raster down to up ANT17.LR.PS - raster left to right TsysLR.17.ps - ant 17 Tsys during above scan ANT17.DLUR.PS - raster down left to upper right TsysDLUR.17.ps - ant 17 Tsys during above scan Conclusions: 1. Some cuts look not-so-bad. FWHM = 3.5 deg or so 1st sidelobe = 10dB (in power). see SCANDLUR, SCANDRUL 2. Some cuts make no sense, eg. SCANLR. This is why I include the the Tsys plots, to see if the total power was fluctuating wildly. But the Tsys plots don't look crazy, ie. one can see Cas A with FWHM = 4 deg or so. 3. For antenna 6 all the cuts were of type (2) above, that's why I don't show them. cc --------- As addendum to my report on beam cuts yesterday, attached are plots for Antenna 6 raster, with Antenna 17 acting as reference. ANT6.DRUL.PS - Cross correlation amplitude TsysDRUL.6.ps - Tsys measured by VLA The raster just looks noisy, although the Tsys plot looks like a beam. Again, 1 milliday= 86.4s = 6.0deg. From the Tsys plot the beam FWHM = 3.9deg. I've put the FITS data for beam cuts and Virgo A and RFI sweeps in anonymous: ftp.aoc.nrao.edu/pub/staff/ccarilli also, in response to Kim's questions: 1. the cross correlations in the digital TV band (187.5MHz) show perfect correlation with a fringe rate vs frequency exactly as expected for the signal coming from Sandia crest (see the 3C147 plot from yesterday - 3C147.ID3.PS). 2. we don't see any spike at 196.83 MHz. There is a weak spike at 197.1? cc ------------- Meeting with Frank Gilfeather March 22: 1. appraised him of project and problem with KNMD 2. he agreed to approach KNME 3. send note detailing issue 4. goal - get 2 hrs at night in april where they turn-off for testing. Gilfeather - 856-5680, 272 7042, 269-0346 -------------- RFI plots - do all in linear!! All, I have processed the frequency sweep data from March 22 on 3C147 0542+498. Attached are some plots and notes. plots: 0542.17.ps - autocorr on ant 17 0542.6.ps - autocorr on ant 6 0542.XC.ps - cross corr 6-17 notes: - all used 1.56 MHz with 512 channels - all were on 3C147 about 100Jy - even in cross correlation, quiet bands (196.6MHz), we see modulation... - note that in previous RFI sweep plots I did something mathematically ill-defined in the scaling. these new plots are all linear. cc modulation -- possibly confusing source? NVSS - closest bright source is 2 Jy (not strong enough) at 1deg radius. wider search for 10Jy sources (to 4deg radius) - nothing. --------------- archive passwd - ACTST 3UY5kdeK APEVLA is VNHRuguP. ---------------- March 24: 1. reduced RFI data from March 22, report to folks to VLA: carilli, perley, greenhill, sowinski 1. P band on/off Cas A w/o short - morning Tsys in AIPS - ant 6 isn't sensible, ant 17 looks ok 2. in ant 6 - new 15V power supply (remove 17.2V) 3. short 2 dipole - 2PM then P on/off on Cas A + 22IAT - start good raster at P band 4. remove short and repeat 5. during C/P work - checked out 183.5 and 196MHz 3MHz bands with S/A - fairly clean, but not empty (and could be TV signal below S/A noise level) 6. 3c123 - 2 band - 3MHz 256 chan at 183 and 196MHz 7. Cas A - pol 2 band continuum mode ----------- Perley report on (1 - 4) above Lincoln: I've reduced some of yesterday's P-band data. The results are decent. You'll recall that the tests were of the effect of shorting the 2-meter feed dipole. There are two tests: a) Effect on P-band G/T b) Effect on P-band beamshape. The first was measured by on-offs using Cas A. The second from holographic beam cuts through 3C147. Results. A) G/T. This is very easy to measure, using the back-end Tsys measurement. The Tsys is calculated from the sync detector voltage, Vs, by: Tsys = 45*Tc/Vs, where Tc is the noise diode temperature (as measured in the lab). The factor 45 is entirely mysterious to me, and nobody has ever been able to explain its origin. However, since we are determining a ratio, we needn't worry for now about this. Similarly, the value of Tc is immaterial (and it is clear from the results from the other antennas that these values are far from correct, if we are to believe the antennas are even approximately similar!). We calculate the metric Tcas/Tcold, which is a good indicator of the G/T (antenna gain/system temperature). Tcas is the increment in Tsys when going onto Cas A, and Tcold is the system temperature on cold sky -- taken as towards 3C147. The data are of good quality. Although we did 4 'offs', each 4 degrees from Cas A in the cardinal directions, I used the Tsys values from 3C147 (Test B) as the 'cold' value. From the G/T data from all antennas, it is immediately apparent that antenna 17 is amongst the best, if not the very best. Unfortunately, we don't have a similar measure taken before the 2-band dipole was installed. (But we will for the next two systems). Typically, G/T is 2.5 to 3.0. For antenna 17, G/T = 2.89/3.17 (R/L Poln) with the 2-band feed in normal operating condition, G/T = 3.09/3.33 with the 2-band feed shorted. So we have about a 6% improvement in G/T from shorting the 2-band feed. It is unlikely that the presence of the 2-band feed degrades P-band performance by any significant amount, as this antennas is the the best of all of them. It is clear from the data that the improvement in G/T is due to an increase in G, rather than a decrease in T -- as expected. B) Beam Pattern. I have reduced only the Down-to-Up cut (it's the easiest). The plot is attached. You'll see that there is no believable change in the pattern. A beam squint is certainly notable. DU.ps is the full cut, DUZ.ps is a zoomed-in view. I'll reduce the other cuts later today, if I have time. Rick I got the Down-Left to Up-Right scan reduced before time ran out. The plot is attached. Effectively no change at all, and no squint. I recovered my other plots (from earlier data), showing the improvement in sidelobe structure caused by the 2-band feed. This is shown by comparison with other antennas which are absent the feed. I hope the titles are self-explanatory. All are attached. -------------- March 29 Summary -- a. VLA measurements suggest ant 6 RF signal path is not working a. war games make these TP measurements very difficult, at all bands b. square wave voltage pattern driving cals is seen in Rx box c. there is evidence that the 2band cals are firing in the SD voltages measured on A rack, but factor 10 fainter than they should be and not seen in millhouse Notes: I. Morning - tests with VLA to check 2 band system, with K. Sowinski. Observed 'blank sky' and Cas A in nawrrow band quiet frequency range (196MHz). results: ant 6 - ALC 'pegged' at -1 => getting no power in RF path, such that ALC is working at max to restore level? (should be -4) Total power - just noise, both on blank sky and Cas A (value = 0.1) Synch Detector - nothing - value around 0 ant 17 - ALC also seems 'pegged', and TP, SyncD were 0 on blank sky on Cas A looks like something is getting through - TP = 5, at least for IF B,D. II. At site: Carilli, Mertely, Bob Smith, John M. goal: check cals in ant 6 -- note: war games at white sands => RFI at all bands, very bad -- checked bandpass with S/A - looks same as before, but lots of intermittent 'hash' across full band at >=10dBm due to white sands. the balance between the different bands actually looks closer now than before (4 , 2, P) -- used VM to check Total power (TP), synch detec (SD), etc... on A rack. TP = 3Vish - OK, varies more than usual. this means the ALC is doing its job, but less stable than usual(?) SD - can see at P band, but at 2 band - barely seen. with cals turned on we get 0.2 (A) 0.7V (B), should be 3V. with cals turned off we get 0 -- used oscilliscope to check if cals are firing (15V square wave) yes in vertex room Rx box [-- swap out F11] -- used Millhouse to search for slow cals in total power. can be seen at P band and 4 band at 0.3 to 0.4dBm, but noisy (very) due to RFI cannot see at 2 band convincingly (maybe something?), tried LPC and RCP -- could not do barrel work due to high winds (30 MPH) to do tomorrow (thursday): -- retry millhouse/A rack cal searches if RFI environment is better -- check 15V square wave in barrel -- adjust power supply to get 15V in barrel ------------------------ Gilfeather - has got KNMD to turn off for testing. hurrah. exchange -- Great news. Some responses. > I talked to the engineer and he is willing to work with me to arrange > a midnight to 6 am quiet time on specified nights. > > I need a couple of things first. > > 1. Why can you not move the frequency - I have asked several > astronomers and don't have a good answer yet. The main purpose of the system is to perform a specific observations of the Cosmic Stromgren Spheres around 3 know quasars in the near-edge of the epoch of reionization. These observations should provide the first limits on the neutral fraction of the IGM at the end of reionization. The redshifts are set by the quasars, and hence dictate the observing frequency. In particular, the most interesting source is the highest redshift, z=6.42, which puts the HI line within the channel 9 digital band (191.4MHz). > > 2. How many days do you need and when - so can you give me a > tentative schedule of when you would want the time so I can try to > negotiate it. Right now we are in the middle of debugging some hardware. I think we will be ready for the night tests toward the end of April, when we have 4 systems up and other problems sorted out. Can I get back to you with a specific date next week? Also, we might prefer 2 or 3 hours for 2 nights as opposed to 6 hours for 1 night -- that would fit better with other VLA scheduling. many thanks, Chris Carilli NRAO -------------------- March 31 testing 1. with Sowinski: tried different front end filters to limit RFI before modem ALC. problem - 'fixed' frequency in IF centered on like 1025MHz => narrow filter could preclude specific band we want - ken will check results: found that Cal was turned off, so that all previous VLA tests had no cal switching. ugh. fixed - looked at things very quickly, so notes are not very reliable/detailed. main thing is that we see power in the synch detectors now. no FE filter (=> 50 MHz): FE measurement: ant 17 0.5 +/- 0.25 TP=3 ALC = 5 ant 6 - jumps around BE measurement: 17 -- TP is ok, synch detect=6,2,10,9 (ABCD) 10/9 => 'pinned' 6 -- synch det=0.2 - 0.8, no TP (0.7, 0.1, 0.3, 0.2) FE filter = 25 MHz BE 17 - looks ok, ALC=-2 on B,D, SD - 'pegged' 6 - TP still 0 12.5MHz filter FE - getting cal signal both ants BE - TP and 6 = 2 to 4 17 - TP OK, SD=10,2,10,5 Overall -- above numbers went by too fast. need to redo systematically/slowly. 2. in AIPS: a. Continuum obs on 3C123 (0437+2940 = 180Jy) -- Tant 6R,L 17R,L no filter = 15,15 15,15 K (all exactly 15) 25 MHz = 23,90 28,14 12.5 = 30,66 84,91 b. Line observations: Tant 6 7 183.7 42 30 stable - Clean Narrow band 1.5MHz 195.7 250 230 stable - Clean Narrow band 189.1 340 520 stable - Chan 9 Narrow band 190.6 330 350 stable - broad band 12.5 MHz c. Line obs: spectra broad band - xcorr: without hanning smoothing, just get ringing right across, with smoothing - get phase coherence across band autocorr: we see what we expect - TV stations 184, 196MHz: 3MHz BW autocorr: clear RFI. some looks like spread spectrum - related to war games? cross corr:184 - possibly coherence, but lots of variation across band, plus RFI 196 - looks ok - fringes. 189 MHz - 1.5 MHz in chan 9 digital autocorr - ouch! cross corr - RFI, phase winding as expected --------- Front end filters: reject RFI before ALC/FE mixing onto modems nothing useful for F6 setttings = -810 (190-240), -860 (140-190) at -840 => 160 to 210 25MHz => 172.5 - 197.5 12.5 => 180.75 - 193.25 might be useful for testing --------- April 1 testing: 1. Turn off Channel 9 ?? - 10 min 2. Beam shape using M87/holographic cuts 3. Effec, T_sys - Cas A or ?? Attached is file. We do many things: 1. holographic beam cuts for ant 6, then 17 2. On/Off G/T tests on Virgo (1000Jy), then 3C295 (100Jy) then Cygnus A (10000Jy, to order of magnitude). i fiddle with spectral modes to get high resolution, then both polarizations, then for cygnus cross pol. 3. a short RFI test to see if Channel 9 is there. also, talked to Gilfeather. he's with us. he also said that KNMD engineer was willing to switch off pretty much any night midnight-6AM. seems like they don't do anything then, although request has to come from the university. to start with, Frank will ask them to shut off for our next late night test April 16. we need to send him some stuff: 1. web site with info (he has that already, but wants it again) 2. schedule of late night test time 3. vla proposal to justify the frequencies 4. recent powerpoint presentations ---------------- April 1 - tests 1. RFI: high resolution - chan 9 still there in middle night low res: something at 193.5MHz wopping! 2. Fringes, sensitivity, bandpass: virgo, cygnus, 3c295 3. Beam cuts: noisy - virgo too faint? Xcorr: ant 17 - look fair. noisy. a few look 'fair', some are really bad - RFI? ant 6 - really noisy. RR real bad, LL looks fair. messy messy - sometimes RR looks better. sometimes LL. need stronger source - Cygnus A? or something else outside Gal plane. peak = 160-180 'Jy', baseline = 40 assume 40 = 'off' voltage, delta = 130 = 'on' voltage note - this is xcorrelation, so noise is NOT Tsys. G/T on Cygnus: look at R only, narrow channels Tant: 7 1959+407 1 0/09:04:50 - 0/09:09:40 1 8 1959+40W 1 0/09:10:10 - 0/09:10:40 1 9 1959+40S 1 0/09:11:10 - 0/09:11:40 1 10 1959+40E 1 0/09:12:20 - 0/09:12:40 1 11 1959+40N 1 0/09:13:10 - 0/09:13:40 1 Note: values are STABLE - to within 5% Cygnus A = 9006 Jy Narrow channels - only 1 pol (R) 6 17 On cygnus 327 563 W 190 383 S 145 222 E 205 201 N 172 188 broader channels - R+L pol 6R 6L 17R 17L On 345 200-69100 530 81 W 190 7000 380 100 E 208 3500-6300 210 47 S 174 3500-5220 224 80 N 182 3700-7800 193 90 Analysis for 'rational data' eg. ant 6 R pol: 1. Assume G = 0.1 K/Jy => Tcyg = 900K Measured: On=327units, = 178units => Tcyg = 327-178 = 149units. Scale factor = 900K/149units = 6.0 K/unit Tsys = 178unitsx6.0K/unit = 1070 K ! 2. Or assume Tsys = 200 K => 1unit = 200K/178unit = 1.12 Cygnus: 149 units = 167 K => Gain = 167K/9006Jy = 0.02 K/Jy Cygnus Gal coordinate: b=76deg, l=5.7deg Could the sky really be 1070 K? ---------------- Note to guys April 1 efficiency/Tsys I've looked at the Cygnus A data from April 1. The observations were of cygnus for a few minutes, then 4 off positions in cleanest band with 1.56 MHz BW. I used both a single pol, and dual pol with few channels. for reference: Cygnus = 9006 Jy at 196 MHz, Gal coord: b=76deg, l=5.7deg I've looked at Tant as reported in AIPS. Also, the 4 band system is up, so results might not be optimal, although I don't think that is an issue right now. Summary: 1. Some data makes sort-of sense, eg. pol R for ant 6. The Tant values are stable over 10min to within few %, both on and off source. 2. If we look at the 'sensible' data, then G/T is 10x too low, ie.: if we assume G=0.1 K/Jy we get Tsys = 1070 K (G/T = 1e-4 /Jy) or if we assume Tsys = 200K we get G=0.02K/Jy could the sky really be 1070K at Cygnus position? 3. Ant 17 - R pol - one off position is high in both data sets. but the measureme 4. L pol gives garbage on both ants. bottom line - we got a long way to go... cc ---------------- Virgo A - April 1 On/Off = 1006 Jy high lat. 1 1230+124 1 0/08:26:20 - 0/08:27:50 1 2 1230+12W 1 0/08:28:20 - 0/08:29:50 1 3 1230+12S 1 0/08:30:30 - 0/08:31:50 1 4 1230+12E 1 0/08:32:30 - 0/08:33:50 1 5 1230+12N 1 0/08:34:30 - 0/08:35:50 1 6 1411+522 1 0/08:38:50 - 0/08:43:40 1 7 1411+52W 1 0/08:44:20 - 0/08:46:40 1 8 1411+52S 1 0/08:47:20 - 0/08:49:40 1 9 1411+52E 1 0/08:50:20 - 0/08:52:40 1 10 1411+52N 1 0/08:53:20 - 0/08:55:40 1 Dual pol: R - stable to within few %. L still not sensible. Note G = 0.1 K/Jy => efficiency = 0.57 6R 6L 17R 17L On 176 1660-3800 146 79 W 93 1750-2690 122 80 S 85 200-46000 132 80 E 98 980-4900 134 79 N 107 1500-2700 123 80 again: 1. only 6R makes any sense. 2. Using 6 R: if G = 0.1 K/Jy => 100.6 K for Virgo off = 96 on = 176 diff = 80units => 1.25 K/unit => Tsys = 96units x 1.25 = 121 K => G/T = 8.3e-4 /Jy Seems OK? For ant 17 get: diff = 18units => 5.5 K/unit => Tsys = 700 K single pol, R: 6 17 On 145 157 E 100 142 W 115 155 N 104 135 S 130 115 ant 6: if G = 0.1K/Jy: off=112, diff = 33units=> 3.0K/unit => Tsys = 336 K and G/T = 3e-4/Jy 3C295 - stable but drifts, can't see source. Overall from AIPS -- no consistency. -Bright sources are seen in some polarizations, not others -get different answers for different sources for different pols, different ants, even different spectrometer settings. -need a source out of Gal plane that is bright! (no such thing - virgo is still best) ---------------- April 4 tests 1. 'deep' integration on Hydra A - in South - 230 Jy, Gal lat=25d, 1hr at RR, 1hr at LL in cleanest 1.56 MHz band 2. some fiddling with front end filters to see if they do anything useful. if they can be used at certain clean frequencies, we might use them during further testing to help with front end ALC issues relating to RFI. ken and i have gone through the frequencies they cover, and the ranges are commented in the observe file. ------------- April 6 on site Notes from April 6 2005 ----------------------- at site - Carilli, Smith, Merteley (some), Hayward (some), McClendon (some). We learned alot today. To quote bob smith - "the engineers should be happy, the astronomers not" Summary: 1. voltages all look ok in vertex and barrel. P band cals look fine. 2. 2band cals are barely, or not, seen, in synchdetect or millhouse TP monitor in vertex room with antenna hooked up 3. 2band cals are easily seen with total power monitors (SynchD and millhouse) when antenna is disconnected and input terminated. => This means that (a) the electronics is all working (ie. the cals are firing and are strong), and (b) the test/monitor setup is working fine. It also means that attaching the antenna somehow corrupts the system. This could be due to radical modulation of total power by RFI when antenna is connected completely swamping everything else. Or some bad mismatch of antenna to system? Or? Note that we were using the S/A 1MHz filter to stop-down the band into millhouse. The word is that this filter has significant skirts, so we may need a better filter for testing. 5. On/off Cas A - have not had time to analyze in detail, but looks like we get 3dBm deflection on/off Cas A. If G=0.1 K/Jy (=> efficiency=57%), then we should get 1000 K from Cas A, which means the Tsys off source = 500 K. This may be reasonable - the sky in this region is about 450K, and Rx adds maybe 50 K. See notes below. I will send various TP traces and broad band S/A spectrum later today. cc ---------------- Notes: Ant 6: 1. Arrival - power supply was on 'off' -- does not recover from a power cycle; set up slow cals 2. Checked voltages with power supply at 15V - in vertex room: Vcal (switched) = 14.93 then 0 - OK - In barrel: 13.98 to frontend, switched cal = 14.88/0 adjust power supply to 16V - In barrel 14.93 V to frontend, switched to cal = 14.88 - OK note: switched cal voltage is coming from signal generator, not power supply. overall - looks fine. 3. Hookup millhouse: a. set up S/A filter, RF path - S/A as 1MHz filter at 196MHz b. set up power meter - zero a/calibrate a c. millhouse - icon = AGE..., ^R - run etc... 4. series of tests: Cas A On/Off + beam cut RCP then LCP Cas A = 10,000 Jy 3C123 On/Off first look: see Cas easily, 3dBm over noise floor (factor 2) if G=0.1 K/Jy, implies Tsys = Tsky+Trx=500 K. assume Tsky=450K and Trx=50 K - looks about right? Rick says around Cas A: "5 degrees to the north, it's 450K. 10 degrees north, it's down to 360K. 5 degrees to the south, it's 380 K, 10 degrees south, it's 220K. To the east or west, you're moving along the galactic plane, and the temperature stays near 450 to 500 K." don't see 3C123 in on/off TP - too faint? there are no good strong sources at high Gal latitude! Note this does not explain rick's very early cross cut, nor why the cals look so strong in his earlier cross cuts. 5. Checked for P band cals - very clear: SynchD = 3.2, 2.4V Millhouse = -25.55 to -25.25 => 0.3 dBm 6. checked 2 band cals with antenna attached: barely/not detected SynchD: 80mV R, 0mV L Millhouse: nothing 7. In barrel, terminate inputs from antennas, ie. unplug antenna and put terminators -- see CLEAR cal signal on Millhouse: LCP: -35.75 dBm to -36.15 dBm RCP: -35.57 to -34.45 [Note: this is not a matched load] 8. Save broad band S/A sweep of band. Lots of highly variable spikes due to white sands in afternoon. ------------------------------ Millhouse plots - Apr 6: Combined output of 4, 2, P bands to 2 MiniCircuits amps (G=20 dB ea) to Spectrum analyzer input, set to: Fc= 196 MHz RBW= 1 MHz VBW= 1 MHz Span= 0 span (not sweeping) Trig= external (no trigger) IF output of SA to power meter power head to power meter to milhouse running the custom Labview DAS program. RCP: \Delta (dBm) (On-Off) = 3 => T(on)/T(off) \sim 2 T(on) = Tsrc + T(off) T(off) = Tsky(\sim 450K) + Trx(\sim 50K?) \sim 500 K => T(src) = 500K; CasA= 1e4Jy => G = 0.05 K/Jy => 28% Efficiency? But -- also see cals clearly: intermittent! (why did we not see them in afternoon?) and Delta(cal on/off) off source = 0.5dBM = 12% On source = 0.2dBm = 5% If Tsys = 500 K => Tcal = 60K! consistency - ugh! conversely, if Tcal = 10K => Tsys = 83 K (can't be right). ----------------- April 11 - Rick and peter think the low efficiency for on/off Cas A is due to impedence mismatch of dipole with antenna -- ie. all the stuff at the prime focus. same was found at 4band, and needed 'trimming' but this doesn't explain: 1. why we don't see cals for L on Cas A, but do see them with dipole disconnected? 2. why cals on R we see sometimes, but sometimes not? 3. why the total power measurement of Cas A on/off gave reasonable Tsys (but unreasonable cals values) - note that 'off' did not change much going away from Gal plane? test -- vector volt meter to measure return loss when transmitting (before FE amp). ------------------ For cold sky, you have plenty of choices. The absolutely coldest part of the sky is at: 09h15m +30 d. Anything north of +30 d at an RA between 07 and 13 h is nearly as cold. ------------------ I asked Ken to give me a table of the frequencies selected by the narrow (12.5 MHz) FE filter. Here they are: -760 is 260.75 -- 273.25 -790 is 230.75 -- 243.25 -810 is 210.75 -- 223.75 -840 is 180.75 -- 193.25 -860 is 160.75 -- 173.25 -890 is 130.75 -- 143.25 ------------------ Tests April 12 - 1. freq sweep of high end 220-240 2. set FE filter to 12.5 MHz and observe at 234 MHz in narrow band, observe on/off virgo + very cold sky 3. then again with no FE filter check: a. record ALC values at 20s - check for large fluctuations? b. Tany values in AIPS All, I am looking at the April 12 test data. Attached are some plots of Tant as reported by AIPS (backend). The next step is to look at ALC data (after I get my car fixed). summary: 1. the total power reported by AIPS is well behaved with 12.5 MHz FE filter, and wildly fluctuation with no FE filter 2. ant 6 Tsys shows no variation on/off Virgo A, while ant 17 shows nice behavior on/off Virgo A 3. We get fringes on Virgo A, both in frequency and in time, with and without FE filters in both polarizations. Plots: L.Tant.ps = AIPS antenna temperature for L pol R.Tant.ps = AIPS antenna temperature for R pol VIRGOA.R.FEF.PS = spectrum of Virgo A R pol, with FE filter VIRGOA.R.NOFEF.PS = spectrum of Virgo A R pol, without FEfilter VIRGOA.R.PHASE.PS = phase vs. time Virgo R pol with FEFilter notes: 1. R and L were observed seperately to get more channels 2. The observing sequency was: cold sky, then On Virgo A, then Off Virgo A 3. This sequence was repeated, first with a 12.5 MHz Frontend filter before the A rack covering 230.75 to 243.25MHz then with no frontend filter. cc -------------------------- I have looked at the FE ALC data sampled at 20s for April 12 data. Attached are plots. The basic conclusions are much the same as for the AIPS Tant data: 1. with FE 12.5 MHz filter the data on/off Virgo A and cold sky look reasonable, and ant 6 behavior is not as good as 17. 2. without FE filter all hell breaks loose. ALC.FE.R.ps = Front End ALC values, R polarization ALC.FE.L.ps = Front End ALC values, L polarization -------------------------- April 14 2005 Folks, I have now also looked at Back End ALC sampled at 20s from the April 12 observations. To remind you, we have 2 ALCs: 1. One in vertex room (FE) which adjusts the levels before transission down the waveguide. This sees the full bandwidth (12.5 or 50 MHz), as set by the FE filter. 2 One in the back end which sets the level before going into the correlator. The backend ALC only sees the 1.6MHz (or whatever bandwidth is specified) set by the backend filter. I sent the FE ALC plots yesterday. Attached are two plots for the BE ALCs: ALC.BE.R.ps = Back End ALC values for full time range, including with 12.5MHz FE filter, and then no FE filter ALC.BE.R.2.ps = Back End ALC values, but a blow up of time range when 12.5 MHz filter was in place. I had a long talk with Ken about the meaning of all this. Some conclusions: 1. With the 12.5 MHz FE filter, the ALC values for both the FE ALC and BE ALC look fine. The value of -5 for the BE ALC is in the normal range (-4 to -6), and the lack of strong variations on/off Virgo is due to the fact that the FE ALC has already taken out these power changes. 2. Without the 12.5 MHz filter, both the FE and BE ALC values fluctuate wildly. The FE variations mean that interference is modulating the system total power. The BE variations mean that the FE ALC is not doing it's job completely. cc ---------------------- Amplitude vs. time here are a couple interesting plots. this is cross correlation amplitude vs time for the April 12 data. Again, the sequence is: 12.5MHz FE filter - cold sky, Virgo A On/Off then some delay and repeat with no FE filter. The two plots are for RR and then LL polarization (note, I observed each separately in order to get more channels, so the times are slightly different). we clearly see Virgo, and cold sky has lower noise than the sky around virgo. For R polarization the Virgo A amp is 300 milliunits 12.5 MHz filter, 70munits with no filter. For L polarization the Virgo A amp is 900 munits with and without filter. cc ---------------------- RFI scans - 217 - 243MHz Attached are the autocorrelation spectra for Ant 6 and 17 from April 12 in 217 to 243 MHz range. Unfortunately, I left a gap at around 230 +/-1MHz due to change in L6 setting. The range 227 - 239 looks pretty clean, although S/A shows possible spike right at 230. The 12.5MHz FE filter sets a range of 230.75 -- 243.25, which includes one large birdie at 240 MHz. Ostensibly, this birdie is not strong enough to affect the total power too badly, since the ALCs and Tant all looked OK, but when we open the full band up (50 MHz) we get reamed. The bandpass for Ant 17 has a big curve in it. cc ---------------------- To do list: week of April 18 return from CV - April 19, to site April 20 w. mertely, smith 1. Setup huron with millhouse labview clone 2. on/off cas a + COLD SKY - check 'eff' - TP monitor 3. move total power monitor to Ant 17 - check for cals/eff/... 4. return loss - ant 17 - antenna/dipole impedence match (problem - 50ohm connectors?) 5. on/off TP monitor - at 234 MHz ------------------ April 16 VHF testing - holographic beam cuts at 234 MHz using FE filter. 20sec per point, 10sec per record => 2 records per stopping point //OF -16 0 0 5.0 20 33 33 point raster at 90deg in Az 5.0 = spacing between points in beams \sim beam * (1/5.0) where beam = Lam/D (in rad or convert to arcmin) KS says: Step size = 0.69deg /(nu_GHz * 5.0) = 0.60deg 20 = timer per stopping point in sec => 33 points * beam * (1/5.0) = 6.7 beam for full raster time = 33 * 10s = 330s = 5.5min per cut 20s = 0.60deg 1 1230+124 1 0/05:32:10 - 0/05:34:10 1 1 2 1230+124 1 0/05:34:30 - 0/05:35:20 1 3901 3 SCANLR 1 0/05:35:30 - 0/05:46:40 1 5701 4 SCANDLUR 1 0/05:46:50 - 0/05:58:00 1 7333 5 SCANDU 1 0/05:58:10 - 0/06:09:20 1 8965 6 SCANDRUL 1 0/06:09:30 - 0/06:20:40 1 10597 7 1230+124 1 0/06:20:50 - 0/06:29:10 1 12229 8 SCANLR 1 0/06:29:20 - 0/06:40:30 1 27529 9 SCANDLUR 1 0/06:40:40 - 0/06:51:50 1 29161 10 SCANDU 1 0/06:52:00 - 0/07:03:10 1 30793 11 SCANDRUL 1 0/07:03:20 - 0/07:14:30 1 32425 12 1230+124 1 0/07:14:40 - 0/07:29:00 1 34057 same strange results - some look ok, some not. stokes RR has much weaker amplitudes than LL, and RR looks 'flat topped' or lousy. LL looks ok, but sometimes broad wings, sometimes flat wings? sometimes 'clips' at peak - IF2? ant 17 - stoke I - looks like ll due to different scales 70mJy vs 1 Jy IF 2 - clips, or 'flat' - but same freq? - ignore IF 1 - RR - 'flat' LL - finally looks OK ant 6 - same thing IF 1: -stoke RR double peaked/flat top - stoke LL - OK IF2: noisy, funny behavior both RR, LL just concentrate on IF 1, LL for now -- both antennas show same pattern for all cuts, eg. UD vs LR .... each look similar. ant 17 reachs 2Jy, 6 reaches 1.6 -- sidelobes vary alot with cuts: DU have broad wings LR is 'rounded' Don't understand RR and IF 2 to do: get scale of time vs angle => FWHM I am getting 5deg FWHM. we expect 3deg. out of focus? or ? Ken says: check other cross polarizations RL, LR. perhaps the polarizers are no good? ACK! -- LR and RL look GREAT! and narrower beam (but still a bit too wide). but LL also looks ok. conclude: 1. polarizations are crossed AND 2. X or Y is busted in one antenna? Beam FWHM = 4.8deg still Report: All, I've looked at the holographic beam cuts at 234 MHz on Virgo A (1000Jy) from April 16. We used the 12.5 MHz FE filter, so total power fluctuations should not be critical. Attached are some plots. Summary: 1. Something interesting is going on in the polarizations: -- RR is noisy, and down a factor 20 from LL -- RL and LR are both as strong as LL! Hard to figure how this could be. Suggests we are cross polarized (ant 6 vs 17), but the fact that LL is also strong (while RR is not) suggests an additional factor. 2. The beam FWHM = 4.8 deg. We expect 3deg at this wavelength (1.3m). But this wavelength is well outside the nominal range of the system, and defocusing etc... will affect this. Better to measure beam at 196 MHz when we have the narrow filters. 3. If we assume the cross hands are the correct ones (which I have reason to suspect), then the first sidelobe is about 20%. Notes: 1. We did 4 cuts: up-down, left-right, then 2 more at 45deg. 2. We used continuum mode, 1.56 MHZ BW, centered at 233.4 3. First set of cuts used ant 17 as reference (not moving) and 6 as moving. Second set was reversed. 4. The timescale can be converted to sky angle with: 1 milliday = 2.6 deg. PLOTS: these are all the left-right cuts. other cuts yield similar results. 17.RR.PS -- ant 17 rastering, ant 6 fixed, RR pol 17.LL.PS -- LL pol 17.RL.PS -- RL pol 17.LR.PS -- LR pol 6.RR.PS -- ant 6 rastering, ant 7 fixed, RR pol 6.LL.PS -- LL pol 6.RL.PS -- RL pol 6.LR.PS -- LR pol ---------------------------- All, Given the strange polarizatin results at 234 MHz, I went back to a continuum observation we had on Cas A at 196 MHz with 1.5MHz bandwidth. Attached are 4 plots: stokes RR, LL, RL, LR on Cas A on March 22. Notes: 1. The amplitudes are strongest for LR, then RL, then LL, then RR. but.. the SNR doesn't look better for the higher amplitude plots. I think this means that the AIPS amplitudes may not be telling us about the true (relative) strength of the different polarizations. I also think the Tsys correction is currently turned off. I'll ask Ken what this may mean. 2. The amplitudes show clear modulation with time. This should not be on Cas A, unless the source ie beating off another source, such as an RFI signal? I think proper testing in this frequency range requires the filters. cc ------------------------------------ April 21 at site: Carilli, Mertely, Smith - Lothar installed labview/millhouse clone on my pc - setup TP system in ant 17 - try return-loss in barrel - do some TP on/off Cas A on 17 Dan installed TP monitor in Ant 17. ---------------------------------- Simulations -- PS to Furlanetto Steve, I have put a simulated uv data set for the VLA-VHF in: ftp.aoc.nrao.edu/pub/staff/ccarilli/EOR.UV The file is just ascii, and is labeled in straight-forward way. The key things are the u,v coordinates (in kilo-wavelengths), and the visibility amplitudes (in Jy). Note that the w coordinate is removed through standard on-line delay tracking. What could be done is to turn these data into a noise power specutrum, to get an idea of our sensitivity. Someday we might even include the power specutrum of fluctuations, but lets first figure out how to convert from visibilities in Jy to noise power spectrum. Simulation Description: 1. The simulation is for 1 day observation with VLA (12hrs) 2. Center freq= 190 MHz 3. bandwidth = 1 MHz channel (only 1 channel) 4. averging time per visibility = 120s 5. I include a 'Stromgren sphere' in the center of diameter 15arcmin (3.2mJy total flux density). should not affect the power-spectrum too much. I could redo the simulation if it does. 6. I use Tsys = 160K, Ant Eff = 50% 7. This is D configuration (1km max baseline) 8. The primary beam FWHM = 3.6deg 8. For your calculations, you can assume we will get 2x more data, due to second polarization. 9. And we can integrate for many days - as necessary 10. And we will have like 8 channels potentially ------------------------------ Next week (apr 25): 1. install filters - ant 6, 17 2. TP tests, efficiency, beam maps - with millhouse clone on 17 - with correlator - check cross pols as well 3. Tant, ALC FE, BE tests 4. Impedence return-loss tests? 5. KNMD turn off test - wed night - REMOVE FILTERS!! Apr 26: On site: Smith, Mclendon -- return loss measurements 150-250 sweep, 180-205 sweep, both pols (both ants?) At AOC - Greenhill tune filters to 195.1 MHz 15 dB looks like +/-3MHz -- might cascade to get more attenuation? RFI test - 200MHz birdie not in S/A sweep of RF, but clearly seen in autocorrelation => birdie in IF? test - short antenna and do spectral sweep. ---------- Reduce RFI data from April 22 1B, 1D, time variable? R and L look much the same. lots of spikes and birdies etc... ugly, but 17 doesn't look that much worse than 6. ----------- April 26 - RFI sweep test without antenna strange ant 17 (no antenna) shows just noies bandpass, ie no birdies ant 6 - shows mostly noise bandpass, plus 190 and 200MHz birdies and NOTHING ELSE! but the antenna is hooked up?? ack! => Ant 6 problem solved: SAO power supply was off. must have cycled in storm over weekend. it has been removed and old power supply attached. Apr 27 - short antennna 6 Report: All, Following are some plots of what the interference looks like for antennas 6 and 17 with the dipoles disconnected and inputs terminated. This would tell us about internal birdies. Antenna 6 -- We see birdies at 190 and 200 MHz. Ken says these are known. The first frequency setting looks strange -- probably just glitch. what are these birdies? - ask ken Antenna 17 -- We don't see any birdies. The shape of the bandpass is different than Ant 6. cc ----------- Apr 27 ----- 2. Ant 6 short RFI sweep: a. Interesting result from ant 17 test: Don't see 190,200 birdies on ant 17 when Rx is working, but dipole disconnected, but DO see them on ant 6 when Rx is powered down! b. Ant 6 ant disconnect RFI sweep: see above 3. Return loss tests in barrel, ant 6 (+ ant 17 day before). Results: Bob Smith says 'it looks like antenna.' get minimum return at 205 MHz (45dB), then pretty steep? wings to either side - NEED TO HAVE EXPERT LOOK AT THESE SWEEPS - REDUCE 4. Installed cascaded filters: sort of 5 MHz wide or so -checked with S/A - look ok - really knock out KNMD etc... Ant 6: At filter peak = -92.0 dBm at 195.25 MHz Far out wings = -109.5 remaining birdies: (do not dominate power) 193.28 = -85.5 195.05 = -87.5 197.82 = -97.3 199.33 = -100.3 Ant 17: At filter peak = -93.8 wings = -110 193.28 = -97.8 197.8 = -105 199.32 = -104.7 5. Fast ALC logging was started 6. Cas A obs with filters on L polarization a. On/Off ALC/Tant data b. Beam cuts 7. KNMD Off test - 3C295 a. Standard L, R sweeps b. High FREQ/TIME Resolution Sweeps - REDUCE c. 1 12.5MHz sweep ------ (7) KNMD Off test: 1:30 AM - 3:30 AM (14:30 LST) Observe - 3C295 = 1411+522 a. Standard sweep: Still there? Also, lots of Birdies on Ant 6, and Ant 17 noisy - white sands? **No clear change in digital signal in 186-192 range** Note to Gilfeather and rest -- did they really turn off? jgale@knme.org response: Chris, Thanks for your note and attached data. I just spoke with Franz Joachim, our operations manager. The on-air operator tried to turn it off several times per the schedule, but encountered a remote control problem. I will have our technical staff look into it and let you know when it is working, so that we can re-schedule at test. Jim Gale Director of Engineering KNME-TV ------- (6a) Cas A - beam cuts Ant 6 - Both RR and LL look good, and RL, LR do not - Xpol down by factor 3- 15% FWHM = 7 steps = 4.2 +/-0.3 deg 1st sidelobe = 15% Ant 17 - much the same Polarization - R ALC could be affecting this Report: All, I'll be sending a series of reports today on the test data from the last few days. The main effort was on installing the filters at 195.4 MHz and re-doing some of the old tests. This first report is on the beam cuts using the narrow filters. Beam cuts, summary: 1. Beams looks reasonable, with FWHM = 4.2+/-0.3 deg, first sidelobe = 15% at 5.4deg from peak. 2. Both RR and LL look OK, with LL having higher peak, which may be the effect of the filter? 3. The cross polarization is about 3% in RL, 18% in LR. But these look pretty noisy -- could be the lack of filter in R polarization affects this measurement? Or there is significant ellipticity of the beam? The plots: DRUL.17.LL.PS -- Antenna 17 beam cut LL polarizations DRUL.6.LL.PS -- Antenna 6 beam cut LL polarizations DRUL.17.RR.PS -- Antenna 17 beam cut RR polarizations DRUL.6.RR.PS -- Antenna 6 beam cut RR polarizations DRUL.17.RL.PS -- Antenna 17 beam cut RL polarizations DRUL.6.RL.PS -- Antenna 6 beam cut RL polarizations DRUL.17.LR.PS -- Antenna 17 beam cut LR polarizations DRUL.6.LR.PS -- Antenna 6 beam cut LR polarizations Notes: 1. The 2MHz wide filters were centered at 195.4 MHz and installed on the L polarization on both antennas 6 and 17. The R polarization was wide open, as before. 2. We observed Cas A, which = 1.0e4 Jy with size=3arcmin. The 1km baseline resolution = 5.3arcmin, so there could be some marginal resolution effects. 3. We observed in continuum mode with 1.6MHz bandwidth centered on the filter. 3. Holography was done in steps of 0.6deg, with 2 records per step and 5 sec/record. For the X axis, the units are: 1 milliIAT day = 5.2 deg Note that the Y axis is uncalibrated (xcorr amplitude). 4. I have similar plots for an orthogonal cut, but these look much the same. ------- (6b) Cas A - On/Off - AIPS/correlator Need Tant, FE ALC, BE ALC Look at both stokes LL - filters, RR - no filters Tant - ant 6 looks ok -- RR noisy, LL good ant 17 - strange -- single valued off? - may never have got set up right? (off source for first 5min) FE - clearly fluctuates alot without filter BE - in this case, looks like FE is taking out much of the problem. Report: All, Following are some notes on the ALC monitoring data with and without a front end filter. Lincoln already sent a summary looking at all the data. I just look at a shorter time range with Cas A On/Off, and I also have look at the antenna temperatures reported in AIPS (basically, a measure of back end synch detector power, scaled by whatever the ALCs are doing). For simplicity, I just send the plots for Antenna 6 Front End ALCs and AIPS Tant. Summary: 1. As Lincoln already showed, the FE ALC fluctuates alot without the filter (R polarization), to the degree that one cannot even see the effect of On/Off Cas A. 2. The FE ALC is well behaved with the filter (L pol), and one can see On/Off Cas A clearly, and even modulation due to the different sky brightnesses in the 4 different Off positions. 3. For this particular test, it looks like the FE ALC for Ant 6 was able to remove the the total power modulations both for L and R polarizations, such that the back end ALC values look essentially constant. This was not true for Antenna 17. 4. The antenna temperature reported by AIPS is clearly affect by FE total power fluctuations due to RFI, although in this time range the RFI was as dramatic as some others. 5. For kicks I include phase vs time plots on Cas A. Clearly we have strong fringes, and the phase stability is pretty good, even for RR (without filter). plots: Tant.6.ps -- Ant 6 antenna temperature as reported by AIPS ALC.FE.6.ps -- Ant 6 Front End ALC value notes: 1. Tests started on Cas A, then had 4 Off positions, then back to Cas A 2. L polarization had 195MHz filter, R did not 3. The observations were at 195MHz with bandwidth = 1.6MHz ---------------- April 28 At VLA - Mert, Carilli, Greenhill, Smith, McClendon 1. retry return loss -- optimal subreflector setting - perhaps not -- nust retry? 2. Setup filter + total power monitor in Ant 17 On/off Cas A + cold sky ----------- Plan for next week: 1. repeat return-loss with proper subrefl setting (smith/Mcclendon) 2. total power On/off Cas A - effic with filters: ant 17 RR, ant 6 RR and LL (carilli/mert) 3. RFI - turn off KNMD - tues night - No, May 18 RFI - ATF turn-off it would be nice to incoroporate into tomorrow night's test an observation of the pole w/ 1 sec integration times over on the order of 10-20 minutes. 4. fringes on 3c295 or 3c147 with filters - check rms in phase => G/T -- best done on May 11 overnight with filters in. 5. check xpol - filter R, L and do Cas A beam cuts 6. TP with ant 17 inputs terminated, ie. 300K input. with cals firing. to do: 1, 2, 3bc, 5 ----------- Terminate dipole with 50ohm matched load => Tsys = 300K (?) From April 6 in Ant 6 "In barrel, terminate inputs from antennas, ie. unplug antenna and put terminators -- see CLEAR cal signal on Millhouse: LCP: -35.75 dBm to -36.15 dBm RCP: -35.57 to -34.45 [Note: this is not a matched load]" if this is 300 K Tsys, then Tcal = 88 K and 30 K ---------- Report: May 4 - ATF turn off RFI test Summary: There was no obvious change to the character of the RFI 'fingerprint' of antenna 17 with the ALMA Test Facility turned off, vs. previous measurements with the ALMA antenna turned on. Chris Carilli --- plots: Ant17.May4.L.ps Ant17.Apr28.L.ps Notes: 1. I confirmed with Nick Emerson at the ATF site that he was able to shut off essentially everything in the US and Euro antennas, plus most of the electronics in the (shielded) trailers, except one PC. He has no access to the Japanese antenna, nor does anyone here right now. (Nick is Darrel's son, by the way) 2. Pointed at 3C295 (1411+522 -- high Gal latitude) This was toward the ATF site for Ant 17 for this LST range. 3. Standard spectral sweep for 1 hr. First stokes RR, then stokes LL 4. Antenna 6 looked perhaps a little better May 4, but these things come and go... --------- Tests May 4 Carilli, Merteley, Smith at VLA Ant 17 and 6 - short dipole and install 1 MHz filter then: a. return loss with subreflector in right place b. cal on/off measurement L and R pol with filter c. cal on/off L, R with dipole back on d. Cas On/off - R pol with filter 00:37 LST - terminate dipole on Ant 17 01:00 - re attached dipole and do cal on/off (-28deg Gal lat) 01:30 - on/off cas a ant 17 05:00 LST - ant 6 dipol cal on/off tests (-4deg Gal lat) ant 17 on/off cals without dipole -- 0.2 dB R, 0.4 dB L over noise floor = 14K and 30K if terminated Tsystem = 300 K ant 6 on/off cals without dipole -- 0.1 dB L => 7K cal if Tsys = 300 K with dipole reattached ant 6 - clearly see power modulating at +/- 0.5 dB over few min due to RFI, even with SAO filter attached. We did not have frequency generator, so had to turn cal on and off by hand from A rack. Interestingly, the cals can be seen at 9.6 Hz beating against TP monitor time (which is close to 10 Hz). But don't quite get to peak or min. ---- Report: Cal On/Off with dipole shorted and 1 MHz filter in. May 4, 2005 Following is my report on the cal measurements with the dipole shorted (terminated) on Ant 17 and 6. The main result is the dBm change cal on vs cal off. The Tcal values need to assume a Tsys, which we don't really have, so I guessed. cc Summary: Dipole terminated: Ant Pol Cal On-Off = dBm Tcal** 6 LCP -24.775 - (-24.875)= 0.10dBm 7K 6 RCP -24.47 - (-24.17) = 0.30 dBm 21K 17 LCP -24.85 - (-25.35) = 0.50 dBm 37K 17 RCP -25.25 - (-25.38) = 0.13 dBm 9K Dipole attached: drift scan, Gal lat=30deg 17 LCP -22.6 - (-23.7)= 1.1 dBm 43K 17 RCP -23.4 - (-23.7)= 0.3 dBm 11K Notes: 1. **Tcal above assumes that the system temp (the noise floor) = 300 K for terminated dipole; = 150 K with dipole attached at high Gal latitude 2. We did not have frequency generator, so had to turn cal on and off by hand from A rack. 3. The 1 MHz SAO filter was installed for all tests. 4. On ant 6 there was 0.05dBm drift in noise floor on LCP with dipole terminated. 5. When we attached dipole on Ant 6, the total power fluctuates by +/- 0.5 dBm on 10's sec timescales, such that the cals are barely noticeable. This happens even though the filters are being used! I think there was alot of RFI this day (white sands). Or it could be Sky moving through beam? But 1min = 0.25deg => only small fraction of beam? LST = 05hr => Gal lat = -4deg ---- Report: Cas A On/Off with filter Ant 17 R, Ant 6 R, L This is my report on the Cas A on - off tests using the SAO 1MHz filter set eat 195.5MHz. Summary: Same conclusion as always: The Cas A deflection is consistently 2.5 to 3 dBm over the sky. If the Tsys on the sky is 500 K, that means on Cas A we are getting 1000 K, half of which is sky. Then if Cas A = 1e4 Jy, that means the aperture efficiency = 28% or so. plots: A6.CasA.RCP.filter.ps -- Cas A on/off RCP Ant 6 from May 4 A6.CasA.LCP.filter.ps -- Cas A on/off LCP Ant 17 from May 4 A17.CasA.RCP.filter.ps -- Cas A on/off RCP Ant 17 from May 4 Ant17.CasA.L.filter.ps -- Cas A on/off LCP Ant 17 from Apr 28 Notes: 1. We did not have frequency generator, so the cals were turned off. 2. Even with filters on, the TP seems to fluctuate alot. I think this was a particularly bad RFI day (white sands), although we don't have any S/A sweeps. ---------- Next week: 1. organize all reports onto web 2. install 2 new receivers - lab test tuesday - install wednesday ----------- LO settings //* *** A = 195.2MHz center, B=193.7center BW=1.56 1411+522 $00 06 00 14 11 20.6477 +52 12 09.141C 22 5555 //DS 5 //LO 0.0 0.0 -840 -840 //FISF 134.400000 232.900000 That said, the center frequencies I get are: 195.18125 = 160 + 34.4 + 0.78125 193.68125 = 160 + 32.9 + 0.78125 [previous stuff off by 0.1MHz] ---------- May 10,11,12 1. Install 2 new Rx/dipoles -- with independent cals etc... Ant 8 - currently on W4, remaining inner array for C Ant 26 - currently on N16 and remaining there Ant 17 will move in for C Ant 6 will stay 2. Filters + interferometry -- 3C295: check G/T 3C295 - 90 Jy at 178 MHz - small, Gal lat = 60 Virgo A - 1100 Jy at 178 - 15arcmin diam at low freq Gal lat = 74 3. test new electronics in lab - looks ok, hot cold, Tcal... note: now have independent cal signals. --------------- May 11 at site: Carilli, Mertely, Smith, Greenhill, Kimberk 1. Cal On/Off on Ant 6 with input terminated -- check for varying baseline from before - looks ok 2. Install electronics (dipoles?) ant 8, 26 - check electronics with dipole terminater - other things? 3. reduce test data from previous night -- 3c295 rms L pol - has filter, R pol - no filter. 1.5 MHz at 195.5 MHz if was 327 MHz, 1pol, 3kHz/ch, 18min, 2 ant => rms_theory= 1.6 Jy/chan S_src = 85 Jy => SNR_theory = 53 /chan Lpol - spectral P-P phase = 35 to 40 deg => rms = 8 to 10 deg if random rms = 180 / 53 = 3.4 deg ?? or 360/53? or 57/53?? P-P amp = 70munits=> sigma = 15-20munits, = 85munits S = 85 Jy => rms = 15 Jy this is 10x too high? Rpol rms phase = 2 deg BUT clearly see modulated signal across band => RFI dominated? Lpol - temporal ave all chans, time series: P-P=50 deg => rms=10-12deg rms_theory = 0.9Jy (400chan, 3kHz/chan, 10s records) => factor 2 worse? overall - depending on what rms phase = random, we are either right on or factor 2 low, or factor 8 off? -------------- Greenhill Cell phone: 617-851-8918. mertley data base: /home/evla/techdocs/fe/working_files/merts-working-files/196mhz-dat -------------- May 11 Lband Ant 8 - before dipole Cas A on/off, problems: 1. LST 4hrs => lowish elevation => high Tsys 2. not enough absorber => T_hot = 230K, not 285+30K 3. we do not see the cals firing on Cas A or hot load -- could it be scaling? I don't think so, but what else? really need to do this in controlled situation ------------- May 12 At site - greenhill, kimberk, mertely, smith install dipole Ant 26 various tests etc... ------------- May 18 Test time 1. KNMD turn-off test a. normal RFI sweep to 197 MHz b. high res RFI sweep around 180-193 MHz 2. 3 antenna fringes - virgo? 3. ant 26 - night time vertex room tests 4. Suite obs files: RFI sweep - May 18 Cas A on/off - Holography beam cut -------------- May 17. 2005 Lincoln, Attached are three observe files. RFI.obs = standard RFI sweeps with LO settings specified. Includes standard and high spectral resolution. Plus observation of NCP (+88deg). Note that the LO settings apply to IF2 (modes 1B and 1D = Stokes R and L, respectively). Ignore the settings for IF1. CasA.OnOff.obs = Cas A plus 4 off positions in continuum mode with 1.56 MHJz BW in 'clean' bands. BeamCuts.obs = holographic beam cuts. This takes some explaining. Here's an example: 6DLUR $00 06 00 12 30 49.4236 +12 23 28.044C 22 5555 52.30 //DS 5 //LO 0.0 0.0 -840 -840 RAPPROT //FISF 134.600000 233.000000 //OF -11 -11 45 5.0 10 33 This is a cut through Virgo A from down-left to upper right (DLUR). //DS card => 5 sec records //OF card: -11,-11 = x,y starting position in 'units' 45 => 45 deg cut direction 5.0 => unit step size = Beam/5.0 = 0.69deg/(nu_GHz * 5.0), ie. Beam = Lambda/D 10 => 10s per step (ie. 2 records per step) 33 = total number of steps cc ----------- May 18, 2005 All, We repeated the KNMD turn-off test last night, and this time it was successful. Attached are three spectra from April 22 and 28 (with KNMD transmiter on) and May 18 (with KNMD off). The spectra are composites of a series of narrow bands at high spectral resolution from 184 to 197 MHz. Note that the baseline level in each sub-band can be different, if there is very strong interference in the band. We clearly see the KNMD digital broadcast on April 22 and 28, and it is gone on May 18, as expected. There are also lots of RFI spikes across the bands, which become more evident in the KNMD band when it is turned off. Most of the weaker spikes seem to be transient, in that they come and go on different days. Still unclear what they are, perhaps local to VLA, or White Sands, or ?. We will track them down. Off course, other TV carriers, like channel 10, are still seen. Many thanks to the folks at KNMD for their help on this. Chris Carilli NRAO Socorro NM USA --------- May 18 At VLA - Carilli, Mertely, Greenhill 1. at Ant 26 -- check cal signal during day (at night = 0.5 dB => 12% of Tsys, Tcal = 20K so Tsys = 180 K). seemed to fluctuate more, and whole baseline fluctuates alot when antenna is tipped. 2. Found channel 10 = kob frmo ABQ 3. moved TP to ant 8 for night tests (cal stuff) =================== Some AIPS loops: TO PLOT AUTOCORRELATION SWEEP SPECTRA and make data file with sweeps in them: [get ccpsm] aparm 0 0 0 0 0 0 1 1 0 - autocorr aparm 1 0 0 0 0 0 1 0 -xcorr edit: anten, source, outfi source '0542+498' anten 6 plver 0 freqi 1 outfi 'ant6.1' go possm;wai possm outfi 'ant6.1.ps' go lwp;wait lwp freqi 2 outfi 'Ant6.2' go possm;wai possm outfi 'Ant6.1.ps' go lwp;wait lwp freqi 3 outfi '' go possm;wai possm outfi '0542swp3-XC.ps' go lwp;wait lwp freqi 4 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp4-XC.ps' go lwp;wait lwp freqi 5 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp5-XC.ps' go lwp;wait lwp freqi 6 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp6-XC.ps' go lwp;wait lwp freqi 7 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp7-XC.ps' go lwp;wait lwp freqi 8 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp8-XC.ps' go lwp;wait lwp freqi 9 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp9-XC.ps' go lwp;wait lwp freqi 10 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp10-XC.ps' go lwp;wait lwp freqi 11 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp11-XC.ps' go lwp;wait lwp freqi 12 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp12-XC.ps' go lwp;wait lwp freqi 13 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp13-XC.ps' go lwp;wait lwp freqi 14 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp14-XC.ps' go lwp;wait lwp freqi 15 outfi '0542.XC.swp' go possm;wai possm outfi '0542swp15-XC.ps' go lwp;wait lwp ------------ source '1411+522' ante 8 freqi 1 outfi 'Ant8-1.swp' go possm;wai possm freqi 2 outfi 'Ant8-2.swp' go possm;wai possm freqi 3 outfi 'Ant8-3.swp' go possm;wai possm freqi 4 outfi 'Ant8-4.swp' go possm;wai possm freqi 5 outfi 'Ant8-5.swp' go possm;wai possm freqi 6 outfi 'Ant8-6.swp' go possm;wai possm freqi 7 outfi 'Ant8-7.swp' go possm;wai possm freqi 8 outfi 'Ant8-8.swp' go possm;wai possm freqi 9 outfi 'Ant8-9.swp' go possm;wai possm =============================