Russ Childers's Argus Data Page

Modified: May 9, 2008

I changed the Argus frequency to 1296.000 MHz last week. Mike Murphy suggested that it would be a good test of Argus if we could try to detect the SETI League 1296 MHz moonbounce signal.

The moonbounce signal was not detected, but this frequency band is not quiet, either. There appears to be a 1296 MHz signal emitting or reflecting from satellites, causing short-term narrowband blips. These blips are relatively strong, and they make an excellent match with the "ideal beams" of the array, indicating far-field radiators.

These are definately "local" signals that we're seeing here, because the source is moving fast across the sky (about 20 degrees in 1 minute). The stars "move" at a rate of 1 degree every 4 minutes - at the fastest. Some stars like the North star, Polaris, barely moves at all.

I have created a movie of a "blip" as it occurred on May 4, 2008. The movie is put together in one-tenth-second frames, showing what the source would have looked if you had "radio eyes", only 10-times faster, for display purposes. The coordinates are not labeled on the movie, so here they are:

Left border: 230 degrees azimuth
Right border: 270 degrees azimuth
<--> 40 degrees wide

Bottom border: 25 degrees elevation
Top border: 55 degrees elevation
<--> 30 degrees high

The source started at 2008-05-04 13:24:00 UTC and was observed (on and off) for 50 seconds. The animation is ten-times faster, to enhance the satellite's movement, so the movie goes for about 5 seconds, then repeats. The frequency is approximately 1296.014 MHz.

Click on the thumbnail movie below, to see the full-size version. Note how the source flickers as it moves from the upper-left to the lower-right. You should be able to pause the movie with the key. Restart from the beginning with the "Reload" button.

(1,548,422 bytes)

Angelo suggested that a "trail" might be seen, if all images in the animation are summed. Looks like that's the case: map_average.gif

The waterfall display (color-coded intensities, frequency on the horizontal axis and the most recent data at top) shows the flickering source as a couple of red dots on the lower-right of the image: wfall_20080504.png

The full-bandwidth spectrum at the peak power: spec_0074.png

A closeup of the spectrum at the peak power: spec_detail_0074.png
Each frequency "bin" is 5 Hz wide. Note that the source is about 10 Hz wide.

broadband_spike_20080329.png (71,797 bytes)

Here is a NOAA weather satellite, captured in full, on Sunday April 8, 2007, from 11:45 UTC (bottom of image) to 11:56 UTC (top of image). [That's 15:45 - 15:56 EDT] Center frequency is 1698.050 MHz.

Look! at the central carrier. Thrill! at the sidebands. Stand in awe! at the doppler shift as the satellite whizzes overhead. Scroll down! because this is a "tall" image.

sat_20070408.png (231,542 bytes)

These spectacular images happen several times a day now. You can see them if you watch the waterfall display here.

Mike Murphy has identified the satellite which we have seen here at Argus for the past 2 1/2 years. It is NOAA 12, a polar-orbiting satellite. Apparently, NOAA 12 broadcasts data at 1698.0 MHz, but we are seeing a "spur" at 1692.035 MHz. Below is the same set of plots as seen in the previous article, but this time with the orbital data of NOAA 12 added in (blue lines).

Here is the altitude:

alt_noaa20070319.png

Here is the azimuth:

az_noaa20070319.png

Here is the frequency:

freq_noaa20070319.png

I'm sure some of the NAAPO volunteers will take a stab at accounting for the 5 kHz offset of the predicted frequency.

Here is a waterfall frequency spectrum of a satellite seen by Argus on March 19, 2007, from 11:32 to 11:38 UTC. Note the "side bands" on either side of the central signal. The center frequency is 1692.035 MHz; bandwidth is 60 kHz. The oldest time is at the bottom and the most recent time is at the top. The highest frequency is at the right; the lowest frequency is at the left.

sat_20070319.png (39107 bytes)

Here is the altitude:

alt_20070319.png

Here is the azimuth:

az_20070319.png

Here is the frequency:

freq_20070319.png

This satellite trace is significant only because it is the best data seen by Argus in about 6 months, ever since horrible-looking interference started showing up in the narrowband data. To eliminate this interference, I changed the "second local oscillator" frequency from 1 MHz to 2 MHz. This may not mean a lot to most of you, but essentially it means that there a problem with some of the receiving equipment, and my desired setting is now undesirable. It's like having a radio with a volume knob that is "crackly" at a certain volume level, and you have to turn the volume up or down slightly to get the crackle to go away. The volume may not be exactly what you want, but at least you can make out what is being said.

The changing of the local oscillator was a "last-ditch effort" on my part to find the source of the interference. There was no evidence that the interference was happening out there in the "real world", so I tried "turning the knob". (More accurately, there are many "last-ditch efforts" employed in finding the cause of problems like this. Oftentimes "remedies" come like a bolt of lightning in the middle of the night. This time, however, the remedy was an "oh, what the heck, I'll try it" one.)

Continuum data collected from 1993 through 1997 is now available to the world. I have arranged the data by "Hour of Right Ascension". Images are in small and large sizes, for those with low or high screen resolutions.

Small images: http://www.naapo.org/~rchilders/contin/html_small/contin_small.html

Large images: http://www.naapo.org/~rchilders/contin/html_large/contin_large.html

Click on one of the links above, then click on "Top declination of RA hour 00". You will see the first declination of the survey (+62 degrees 20 minutes declination). This data was collected in September/October 1993. There were three days' worth of data taken at this declination, hence you see three traces. There is a nice, low-power point-source which passed through the "leading beam" at around point 295, and through the "trailing beam" at around point 330. These "points" correspond to 10-second samples. Thus, point sources at this declination show up ~35 samples, or 350 seconds, from beam-to-beam. This spacing drops to 150 seconds at 0 degrees 0 minutes declination.

The reason why I point out the beam-to-beam spacing is because "one beam" radio telescopes (the vast majority) have trouble telling the difference between noise and true sources, unless they have several days' worth of data. A "blip" may be a celestial source or terrestrial noise. But with Big Ear's "dual beam" setup, celestial sources make characteristic "dual-humped" patterns. This dual beam is a blessing and a curse, however, because if two sources are closer than the "beam spacing" they will interfere with each other. Also, any diffuse background sources (eg the cosmic microwave background) is completely hidden in this dual-beam paradigm. But dual-beam observing gives quick "confirmation" that a celestial source was seen, and a simple deconvolution algorithm can transform dual-beam data into single-beam data (with the exception of diffuse background sources).

Three comments on the collection of images: (1) There are 540 data points in each image, corresponding to 5400 seconds of observation "per hour". "But Russ," you say, "there are only 3600 seconds in an hour." You are correct. The images I've generated actually start 15 minutes (900 seconds) BEFORE the hour, and end 15 minutes (900 seconds) AFTER the hour. 3600 + 900 + 900 = 5400 seconds. I did this so sources which fall at the "top of the hour" do not get "clipped". Thus, the "top of the hour" starts at sample 90, and the hour ends at sample 450. (2) The "first" declination shown is always +62 d 20 m. Click on the "lower" hyperlink to go to the next-lower declination: +62 d 00 m. All declinations are separated by 20 minutes of declination (1/3 degree). The lowest declination is -36 d 40 m, and was taken at the end of August, 1997. The astute observer will see sources get "brighter", and then "fade" with passing declinations. Declination -16 d 00 m is the only one "missing", and I will try to find out what happened to that data. (3) The sun passes through the beams even when the beams do not point directly at it. You will recognize the sun, because (a) it shows up in later-and-later locations in each days' data, and (b) it is broad, and makes a characteristic "large-angular-extent" dual-beam pattern, unlike the weak point-source example above. The moon also shows up occasionally. See if you can spot it. (Hint: it should only show up between declinations +28 and -28 degrees.) Also, "sun dogs" appear before and after the sun when the sun passes through the beams.

Now to the meat of the matter. Sources which show up day after day are nice to know about, but the real exciting sources are the ones which pop up only occasionally. Suppose the "weak point source" illustrated above only showed up on one of the traces. That would be VERY interesting. Now suppose that source appeared in the "leading beam" on one day and in the "trailing beam" on another day. That would be VERY VERY interesting. Bob Dixon has claimed that "if you found a source like this, you would be instantly famous".

A note to those who wish to be famous: I have looked for "spikes" in the data, corresponding to these VERY VERY interesting sources. I have come across many more "positive-going" spikes than "negative-going" ones. This indicates that there is some local interference causing biased spiking, because if the spiking sources are truly celestial, then there would be - on average - the same number of positive- as negative-going spikes.

A note to those who wish for near-fame: Sources which show up in each beam on each day may not be as boring as I allude to above, IF THEY DO NOT SHOW UP WITH THE SAME POWER IN THE ORIGINAL OHIO SURVEY. A source in the Ohio Survey may be much brighter in the 1993-1997 survey, or may have gone away. Either way, a source-by-source comparison should be made with the original Ohio Survey.

A note for you techie astronomers out there. The "declination" is the "declination at the epoch of observation" and has not been precessed to another epoch. Similarly, the "Right Ascension" is the "Local Mean Sidereal Time" and has also not been precessed to another epoch. The Ohio Survey source catalog and contour maps were precessed to epoch 1950.0, and so source-to-source comparisons of the 1993-1997 data and the Ohio Survey catalogs is "tricky".

One last note. A couple of times the data collection computer's clock differed from the true time (UTC) by as much as a minute. This shows up in the images as sources which appear shifted in time from their counterparts. No attempt has been made to correct these clock errors, however I carefully noted these clock anomalies in several log books when I encountered them at the time.

After observing the radio band at 1420.470 MHz over the winter, horrible interference and a desire for greener pastures has driven me to tune Argus back to 1692.035 MHz, where satellites abound. Click here for an awesome satellite trace, observed on March 29, 2006.

I spoke above of "horrible interference" at 1420 MHz. I was surprised to find this interference, because 1420 MHz is supposed to be "protected" for radio astronomy. Oh well. I am also disappointed that Argus was unable to locate the source of the interference: Argus specializes in locating sources with arbitrary locations.

Interference screen capture #1.

Interference screen capture #2.

When doing radio astronomy, it is essential to have on hand a calibration source, which lets the observer know that the equipment is functioning properly. Radio astronomers will often use a "noise source" to do this calibration at regular intervals. The Ohio Argus Array uses such a noise source, which turns on for one second, once per minute, 24 hours a day. The parabolic dish at SatComm does not have such a calibration (yet), so instead I have used my hand to calibrate the parabolic dish feed. I have found that when my hand is in front of the feed, the power output of the receiver system rises by 3 dB, or by 2-times the normal output. Here is a waterfall snapshot of my hand in front of the parabola's feed for about 30 seconds (the light blue stripe at the center of the image). The total time in the waterfall image is 5 minutes, with the most recent time at the top. Frequency goes from lowest at the left to highest at the right. The center frequency is 1420.470 Mhz, and the bandwidth is 60 kHz. (Ignore the 00:55 at the lower left - it is a data processing artifact.)

The real-time waterfall output from the dish is now available online at this URL: http://www.naapo.org/Argus/data/display2.html. Note that the top of the waterfall says "Argus Dish", and the page is named "SatComm Parabolic Dish Waterfall Display". This is not to be confused with the real-time Argus Array output here: http://www.naapo.org/Argus/data/display.html. Note that the top of the latter waterfall says "Ohio Argus", and the page is named "Ohio Argus Waterfall Display".

The astute observer will also note that there is just over 6 minutes of data on the "Argus Dish" page, whereas there is 6.5 minutes of data on the "Ohio Argus" page. This is because the spectra with "calibration on" have been excised from the "Ohio Argus" page; there is no regular calibration on the "Argus Dish" page: all those spectra have been left in. Since both displays show the same number of spectra per image, the time stamp of the oldest data on the "Ohio Argus" page will be older than that on the "Argus Dish" page.

Here is a meridian drift scan of the sun and galactic plane through the main beam of the 10-foot parabolic dish at SatComm on January 15, 2006. The declination of the dish is roughly -20 degrees. The center frequency is 1420.470 MHz and the bandwidth is 60 kHz. Each data point is the total power, in dB, of 16384 samples taken over 0.205 seconds. Data points are one second apart. No averaging across data points is done.

The galactic plane is seen at about 16 hours UTC, and the sun is at about 17.5 hours UTC. The Galactic Center is located at -29 degrees declination, so the beam is rather close to the "brightest" part of the galactic plane (see this image of the radio sky from Kraus's Radio Astronomy). The half-power beamwidth of the 10-foot parabola at 1420 MHz appears to be about 5.5 degrees, after taking the apparent width of the sun (0.5 degrees) and the declination of observation into consideration.

sun_and_gp_20060115.jpeg (96596 bytes)

Here is a waterfall frequency spectrum of a satellite seen by Argus on October 10, 2005, from 11:02 to 11:05 UTC. Note the "side bands" on either side of the central signal. The center frequency is 1692.035 MHz; bandwidth is 60 kHz. The oldest time is at the bottom and the most recent time is at the top. The highest frequency is at the right; the lowest frequency is at the left. The S-shaped frequency profile suggests that this is a low-earth-orbit satellite, where the frequency goes from high to low, like the sound of a race car's engine as it passes you on its way around the track. The intensity of the signal goes from low (green) to medium (yellow) to high (red), and then back down to low.

spectrum_20051010.jpg (298299 bytes)

Below is a polar plot of the above satellite. The image represents the sky as an observer on the ground would see it. The circle is the horizon; the top is North; West is on the right; East is on the left; South is on the bottom. The point "directly overhead" is in the center of the circle. The satellite track goes from close to the North horizon, to near overhead, to slightly south. This plot shows the location only of the satellite, not the frequency or intensity, which are both seen in the above waterfall image. Points (*) not on the track are either noise, or sidelobes (false location) of the satellite's location.

skytrackB_20051010.jpg (60329 bytes)

Here is a screen capture from the Argus waterfall:

outburst_20050907.jpg (318684 bytes)

...and here is a plot of "beam power" for a three-hour time period:

outburst_plot_20050907.jpg (49393 bytes)

...and here is a closeup of 15 minutes from 1700 to 1715 UTC. I don't know if the sets of spikes are local interference or are part of the solar outburst. The spikes were not caused by the calibration source, because the calibration source fired for one second each minute, and these spikes fired in groups of three or four 19 times in 15 minutes.

outburst_plot_20050907.jpg (65583 bytes)

This burst was so powerful, it showed up in all Argus elements, and in fact saturated most of them. Below is the power seen in Element #1 for the 3 hours around the outburst. Recall that the elements have roughly hemispherical coverage, which is a heck of a lot of square degrees. The sun covers a tiny part of the sky, and the outburst was a tiny part of the sun. The burst must have been VERY powerful to actually show up in individual elements. The calibration source is seen floating between 48 and 49 dB.

outburst_elem1.jpg (59764 bytes)

The following graph is a closeup of the above graph from 1700 to 1705 UTC. Note the noise source every minute up around 48 dB. Also note the six groups of spikes between 45.5 and 45.7 dB.

outburst_elem1_closeup.jpg (45799 bytes)

The sun generated two several-hour-long flareups, as observed at 1692 MHz, at 18 hours UTC on August 22, 2005, and at 15 hours UTC on August 23, 2005. The following plots show data from August 21, for comparison, and from August 22 and 23.

The first plots show the "best beam" power as seen by Argus in its entire 60 kHz bandwidth. The "best beam" points at the sun during mid-day. The sun reaches its highest altitude, and strongest power, at around 17:30 hours at Argus's longitude. Note the low hump in the August 21 plot, and the high peak around 18:00 on the 22nd and around 15:00 on the 23rd. (The data is not smoothed; power is in "dB"; artifacts are not removed; power is sometimes "negative", because the "best beam" power is subtracted from a reference beam, which is occasionally stronger.)

best_beam_outburst.jpg (68311 bytes)

The next plots show the "best match" which a 1.5 millisecond (128-snapshot) acquisition made to an "ideal" Argus beam. Argus was using 23 elements in its array on these observation dates, so the "best match" is ideally 23 out of 23 element phases. Random noise will produce 23/2 = 11.5 out of 23 element phases. I show here the 128-snapshot acquisitions instead of the full 16384-snapshot acquisitions, because the sun ALWAYS produces a "best match" of 20-22 element phases when it is high in the sky in the 16384-snapshot acquisitions (essentially a longer integration time). As you can see below, the number of elements which made the "best match" on August 21 was a random value centered on 13 element phases. But on August 22 the best match peaked at 21 at 18:00, and on the 23 the best match peaked at 20 at 15:00.

best_match_outburst.jpg (141081 bytes)

The next two plots show the altitude and azimuth of the "best match" for August 21-23. The altitude and azimuth are simply the alt/az of the ideal beam which the "best beam" was pointing at. On the 21st the alt/az are just random values spaced between 0-90 and 0-360 degrees respectively. However, on the 22nd, the altitude lingers around 62 degrees and the azimuth goes from 160 to 210 degrees. On the 23rd, the altitude goes from 45 to 55 degrees, and the azimuth goes from 110 to 150 degrees. These alt/az ranges correspond to the the motion of the sun for several hours around 18 hours UTC on the 22nd and for several hours around 15 hours UTC on the 23rd.

best_alt_outburst.jpg (190673 bytes)

best_az_outburst.jpg (130884 bytes)

Note: The sun was also seen in the strongest 5-Hz bins, and was accurately located in altitude and azimuth, even though the outburst was spectrally flat in the 60 kHz observation bandwidth. This provided a confirmation that the narrowband source algorithm (which uses FFTs as data) can also provide accurate pointing information. (Narrowband data not shown).

In preparation for attempting to observe Cas A, I observed the sun on August 10, 2005 as it transited the meridian. The data for the plot below consists of data points taken at one-second intervals. Each one-second point is actually 0.2 seconds of data captured in 16,384 analog data points, with each data point squared, summed across all 16,384 points, and then divided by 16,384. The resultant data point is 10*log10 of the mean-squared value.

sun_transit.jpg (27188 bytes)

http://www.naapo.org/Argus/data/display.html

I tried to determine if an element reported different powers when pointed toward the sun, vs pointing away from the sun. This experiment was conducted when the sun was close to the meridian, at about 72 degrees altitude.

The plot of data generated during this experiment (see below) is the power in dB. (Sampled data, with zero DC component, sum of 16384 squared samples divided by 16384; 10*log10(sum).)

A) Samples 1 -> 1000 are with the spiral "pointed at zenith".
B) Samples 1000 -> 1600 are with the spiral "edge-on to the sun".
C) Samples 1600 -> 2200 are with the spiral "pointed at zenith".
D) Samples 2200 -> 4600 are with the spiral "pointed at zenith", with a screen shielding the spiral. The screen was shielded by a non-conducting insulator.
E) Samples 4600 -> 5200 are with the spiral "pointed at zenith".

From the graph, the lowest power was with the spiral pointed at zenith, with the sun "in the main beam" (A,C,E). The spiral "edge-on to the sun" (B) produced higher power because it was seeing the "hot" roof of SatComm (I suspect) in half of its main beam. The highest power was when the shielded screen covered the spiral (D), because a "hot" object filled the entire "main beam" (I suspect).

elem23_expt.jpg (47479 bytes)

This impressive source is seen as it goes from alt 30, az 10 degrees, through zenith, and ends at 30 alt, 200 degrees azimuth. The main signal starts at the right-hand side of the screen, and ends at the left-hand side. The S-shape is caused by a change in Doppler shift of the signal: there is less Doppler shift in the rising and setting as compared to when the source is overhead (think of the sound of a race car as it zooms past a spectator close to the track). The sidebands follow the shape of the main signal. The center frequency is 1692.030 MHz, with a bandwidth of 70 kHz. The entire screen is about 8 minutes in time.

nb05may2005.gif (57933 bytes)

The sun as seen by Argus on February 18, 2005. This image was generated in "interferometry" mode, using two elements approximately 3 meters apart, roughly on an east-west line. Data was generated by multiplying each elements' I-value and summing, and multiplying each elements' Q-value and summing. A 400-acquisition moving window convolution was done over the entire 24-hour set of acquisitions, resulting in a "480-second" integration time. The x-axis is the UTC time; the y-axis is the sum of correlations centered on the 400-acquisition window. Note the interferometric "fringes" while the sun is above the horizon. Fainter fringes might be seen when the sun is below the horizon, perhaps caused by the Cas A supernova. Receiver center frequency is 1692.030 MHz; bandwidth is 70 kHz; each acquisition takes 0.2 seconds; there are 16384 samples per acquisition; time between acquisitions is 1.2 seconds.

fringe1a.jpg (74093 bytes)

The sun as seen by Argus on February 2, 2005. This image was generated by combining many "drift scan" beams into one contour plot. The drift scan beams are aligned along the meridian, at 5-degree altitude intervals. Center frequency is 1692.0 MHz, bandwidth is 70 kHz. Integration time is 41 seconds, which is derived from averaging 200 0.205-second acquisitions. The x-axis is hours of Local Mean Sidereal Time. The y-axis is degrees of declination. Sidelobes of the sun appear at various locations. These sidelobes all appeared on the meridian, because that's where the beams were pointed. Note the sidelobes close to 90-degrees declination. One would expect these sidelobes to be smeared out, because the plot is a "mercator projection". But this is not the case, because sidelobes can travel fast through the pole. There is no indication of Cass A or Sgr A (bright extra-solar sources), but this is not unexpected, because these sources passed through the meridian while the sun was up.

sun_feb02b.JPG (90112 bytes)

This source was seen by all who attended the radobs meeting on Saturday, October 2, 2004. Argus formed beams all over the sky, searched for the strongest narrowband source in each beam, then used the spectrum of the "strongest" beam to create the waterfall. The source appears to be a satellite going directly overhead from north to south along the meridian. The source shows up in the center of the screen, at alt/az 45/10, and ends up on the left-hand side of the screen at alt/az 80/200. Center frequency is 1692.030 MHz, with a bandwidth of 70 kHz. The time of the bottom spectrum is 10:07 EDT and the top spectrum was taken at 10:13 EDT.

nb02oct2004_5.jpg (213159 bytes)

The same narrowband source as above, except Argus has a "fixed" beam at alt/az 43/168. Note that without tracking, the signal seems to be weaker and fades in and out. This is because the source is passing through sidelobes of the 43/168 beam.

nb02oct2004b.jpg (231451 bytes)

A plot of the "best" narrowband frequency for the "tracked source" above. Note the smooth curve in the middle of the plot. This curve, combined with alt/az data could allow us to determine the orbit of the (supposed) satellite. It is not known, however, if the received 1692 MHz signal is steady, or if it originates from the satellite, or if it is a reflection from some other source. The x-axis is the 1.5-second acquisition number, and the y-axis is the "frequency bin" number (bin 0 -> 16383 = 1691.994 MHz -> 1692.066 MHz).

nb_bin2.jpg (98861 bytes)

nb03sep2004.jpg

sidelobes1.jpg

Acquired Saturday, August 7, 2004. The beam was switching during the 10-minute acquisition: the beam pointed at GOES (alt 43, az 168) for 40 seconds, then the beam pointed at the meridian (alt 70, az 180). One can see the strength of the signal increase in the GOES beam and decrease in the meridian beam.

goes1.jpg

Source frequency is 1692.0202 MHz, bandwidth is 25 Hz, altitude is ~49 degrees, azimuth is ~110 degrees.

argnb293.jpg

Argus sees a satellite at 13:56:07 UTC (note change from 13:55:33 on image), on 08-JUN-2004. Source frequency is 1692.0255 MHz, bandwidth is 25 Hz, altitude is ~55 degrees, azimuth is ~93 degrees.

argnb5.jpg

A waterfall representation of the above satellite. The beam is pointing at 52 degrees altitude, 100 degrees azimuth (about the middle of the observed path of the satellite). In the center-right of the screen, the sattelite can be seen changing in frequency over time (chirping). The chirping from high to low frequency, combined with the rapid change in location (shown above), is characteristic of a low-earth orbiting satellite. The radio signal may be emanating from the satellite, or it may be a reflection of a terrestrial transmitter.

nbsource1.jpg

argsun5.jpg

argsun9.jpg

argsun10.jpg

yaplet