setiQuest GOES-11

The Allen Telescope Array (ATA) collected an 8.7 MHz wide chunk of spectrum from the NOAA GOES-11 weather satellite at 1692 MHz back in September 2010.  The GOES-11 satellite has since been decommissioned and replaced by GOES-15 but its glorious signal lives on in the setiQuest data archive.  This particular quadrature dataset has a duration of 532 seconds and is 8.9 GB in size.  The baudline signal analyzer is going to take a look at some of the signals in this GOES-11 dataset.

There are an incredible number of signals hidden in this dataset.  We're going to zoom in, browse around, measure, and identify a few of them.  For fun I'm going to take the approach of doing a blind analysis of the GOES-11 signal data.  This means that I'm not going to use the Internet to determine any signal features such as modulation types and baud rates.  No external help.  Just baudline.  Let's go ...

The average spectral plot above is a map of where we're going to zoom in and analyze.  The blue lines represent wideband regions that will be focused on in different sections of this analysis.  The red dots are representative of individual signals that will analyzed in depth.  We begin the analysis on the left side of the spectrum and will work our way to the right.


-2000 ... +100 kHz
Here is the chunk of spectrum from the first blue line.  Tip: use the size and position of baudline's lower scroll bar to determine the frequency resolution and position in relation to the main spectrogram/average map.  In this example notice how the size and position of the scroll bar match the size and position of the first blue line.

It looks like a lot is going on here but in fact there are only two unique signals.  Everything else is caused by mixing problems.  Let's try to sort things out by making some frequency measurements.

The two red dots represent the signal that we'll being analyzing in depth next.  The orange, yellow, and blue horizontal lines represent the distance between the constant features seen in the previous spectrogram.

orange = 294400 Hz
yellow = 258133 Hz
blue = 36267 Hz

Here are some interesting mathematical relationships:

orange = yellow + blue
yellow / orange = 7 / 8
orange + yellow = width of the main central lobe

The 7/8 ratio is very unusual as is the coupling of the two different signal types.  Now let's zoom into the time domain and check out the two signals marked by the red dots.

This level of signal resolution clearly shows six Frequency Shift Keying (FSK) modulated signals.  Notice that the mark and space frequencies on the left side are inverted from the right side.  Also the two FSK's on the outer edges are wider with roughly twice the frequency delta between mark and space.  To get a better view here is a zoomed in spectrogram extract of them:

Baudline's periodicity helper bars measure a bit rate of 9600 bits/second.  The multiple FSK signals have different frequency widths but they have the same bit stream.  This makes them the same base signal.  One of them has been stretched by an unknown process.

Above are the stacked average spectrum of the wider FSK and the narrower FSK signals.  The upper FSK has a measured mark/space delta of 20 kHz and the lower has a mark-space delta of 8 kHz.  This 20 / 8 = 2.5 ratio is far enough away from a factor of 2 that something else other than basic harmonic mixing is occurring.

What is stranger is that the spectral shapes look completely different.  How and why is a mystery.  They are the same signal.  Straight alias mixing should not be able to do this.  Here are some modulation measurements for both FSK signals:

modulation type: FSK
spectral width: 30 and 16 kHz
mark-space delta: 20 and 8 kHz
bits / symbol: 1
symbol rate: 9600 symbols / sec

Note that when the bits / symbol parameter equals 1, the symbol rate is equal to the baud rate.  So both of these FSK signals have a 9600 bit/sec baud rate.


-1000 kHz
Next let's look at the wide spectral hump that is in-between the FSK signals discussed previously.  Centered at -1000 kHz, the red dot in the following image highlights our target signal.

Sharp filters keep the FSK signals from impairing this signal of interest.  Carrier lock with baudline's new IQ display:

The two bright points are Phase Shift Keying (PSK) symbols that are 180° out of phase.  This signal is called Binary Phase Shift Keying (BPSK).  Here are some modulation measurements:

carrier frequency: 1692 + 0.999991 MHz
modulation type: BPSK
spectral width: 350 kHz at -6 dB
bits / symbol: 1
symbol rate: 293884.01 symbols / sec

For PSK modulations, the symbol rate being roughly equal to the spectral width is a good measurement sanity check.  An interesting observation is that the accurate BPSK symbol rate measurement is approximate to the less accurate orange delta Hz measurement bar that was discussed at the beginning of this section.  An unknown mixing process must be at work for the BPSK symbol rate to be connected to the frequency positions of the FSK images.  This is very unusual.

Next let's focus on the transient horizontal bursts that are overlaid on top of this BPSK signal.  There are more than eight bursts and they appear to have random durations and spacing.  The duration of the bursts range from less than a second to more than a minute.  It's interesting to note that the IQ constellation doesn't change when mousing over the bursts.  The gain fluctuation in the spectrogram frequency space is quite dramatic yet the gain, phase angle, and jitter all remain constant in the IQ space.  How is this possible?

To answer that question let's zoom into the time axis with a smaller FFT size.

There is significant structure here. Click the image for more detail.  The middle vertical line is one of the FSK artifacts.  Baudline's periodicity bars show that these horizontal burst groups have a 56 ms periodicity.  Each burst is 48 ms in duration followed by a 8 ms off time.  Zooming deeper into the spectrogram shows a highly repetitive spectral pattern.  An autocorrelation plot verifies this.

The primary lag in red represents a periodicity of 0.3845 ms.  Note that the rightmost red line has been folded by aliasing.  The secondary lag in yellow is 1.7358 ms with 0.3845 ms sidebands.  The calculation 1.7358 ms / 0.3845 = 4.514 is an unusual ratio.  Using the BPSK symbol rate, the two lags translate to 113.00 (red) and 510.12 (yellow) symbols.  The sidebands and the strange ratio suggests two possible explanations: that the secondary lag is being time modulated by the primary lag or that each transient burst consists of a repeating bit pattern that has two periodicities.  How to construct either option is not obvious.  There must be a simpler explanation.

Since GOES-11 is a weather satellite these bursts may signify blank or repeating image lines.  An idle pattern or training sequence is another option but the complexity of the signal makes this unlikely.  Demodulating the bits would answer this question definitively but that is a project for another blog post.


+1800 ... +2900 kHz
Now we're going to jump to some signals of interest on the right side of the main spectrum (the 2nd blue line).

First we'll look at the strong vertical signal on the left side and then we'll dive in and take a look at some of the many signals shown in the right half.


+2000 kHz
The average spectrum of the strong vertical signal on the left.

Notice the middle strong tone and the mirror image spectral humps.  This what happens when the transmission carrier in not suppressed.  The delta frequency between the strong carrier tone and one of the spikes on the hump is 2 kHz.  The spectral width of the one of the main humps is 4 kHz but the harmonics of this modulation stretch out to about 500 kHz wide.  This is not an effective use of the available bandwidth.

The IQ display can be used to determine the modulation type.

Notice how the two symbols are offset from the origin.  This is caused by the non-suppressed carrier.  With some decimation, down mixing, carrier suppression, and selecting either the upper or lower sidebands (USB or LSB) the IQ plot becomes:

Here are some modulation measurements:

carrier frequency: 1692 + 1.998614 MHz
modulation type: BPSK
spectral width: 4 kHz of SSB
bits / symbol: 1
symbol rate: 4000.025 symbols / sec


+2225 ... +2750 kHz
Here is a spectrogram of the many signals section (the 3rd blue line):

It is full of signals.  Thousands and thousands of signals.  Let's start on the left and zoom into several of the interesting sections.

I call this next section centered at +2250 kHz "snow" because that is what it looks like.  The deeper you zoom into both axes the more signals you see.  Very high signal density. It is like a whole world unto itself.

Click the image above to see the fine details.  Even at high zoom levels the signals are not well defined.  The blips look like very short duration tone bursts, some look like fuzzy snowflakes.  It is like some mysterious process has shrunk down and squeezed together what you are about to see in the following sections.

There are many signals of various durations, sizes, and shapes.  The signals look like short modulated packet bursts of data.  Most of the signals appear to be lined up in frequency slots.  This is known as Frequency Division Multiplexing (FDM).  Many signals appear to begin transmission at a common 30 second time slot and to a lesser extent at 10 second intervals.  This behavior doesn't apply to all of the signals but there is a great deal of organization.

The image above is a different section that is zoomed in a bit.  Some of the signals have strong sidebands and some don't which suggests that there are multiple transmitters involved.


+2422 kHz
From the right side of the previous image, here is a deeper zoom into the frequency axis.

Four modulated packets are visible.  Each packet has a strong central carrier that is the source of the upper and lower sideband symmetry.  On closer inspection the top two packets have a slight drift to the right and to the left.  This could be due to Doppler or it could be due to a non-stable and slowly drifting transmitter clock.

Slicing through the middle of the spectrogram is a fast moving object that is being Doppler drifted.  The slope is straight at the top (overhead) and then becomes curved as it recedes (away).  This behavior follows the expected Doppler flyby curve shape.  The measured drift rate for the top "overhead" section is 90.6 Hz / 118.1 seconds = 0.767 Hz/sec.  The lower "away" section is slightly curved which makes it difficult to measure accurately.  An approximate measurement of the lower drift rate is 83.3 Hz / 161.7 seconds = 0.515 Hz/sec.  This slow moving signal looks to be AM modulated with a carrier-to-sideband delta of 10 Hz.  If this was a modulated On Off Key (OOK) data signal then the baud rate would be 10 bits/sec.

Here is the lower sideband IQ plot of the last modulated packet.

This jittery constellation is BPSK.  It looks different than the previous BPSK signal seen above.  The extra jitter and other signal instabilities suggests a lower quality modulator / transmitter source.

carrier frequency: 1692 + 2.421614 MHz
modulation type: BPSK
spectral width: 200 Hz of SSB
bits / symbol: 1
symbol rate: 200 symbols / sec


+2505.6 kHz
Here is the spectrogram display of a nearby frequency.

Five packets are visible.  Each packet begins with a brief tone and is then followed by what looks like a wider bandwidth modulation.  Notice the modulation sidebands.  Here is the waveform display of one of the packets.

Notice the constant envelope with many different types of phase shifts.  Next, let's look at the IQ display.

This modulation is 8-PSK.  Each of the eight symbols has it's own distinct phase (360° / 8 = 45°).  In this modulation scheme a symbol encodes 3 bits of information (2^3 = 8).

carrier frequency: 1692 + 2.505610 MHz
modulation type: 8-PSK
spectral width: 160 Hz at -6 dB
bits / symbol: 3
symbol rate: 150 symbols / sec
baud rate: 450 bits / sec


+2704 kHz
What is this mysterious looking stuff next to the wide vertical line?

This particular signal is featured in the found a DNA strand comment on the setiQuest forum.  We will be avoiding any erroneous time aliasing distortion here.  Let's zoom into the time and frequency axes to get a better look.

This looks like a harmonic power ramp up with the frequency carrier stabilizing after a half second.  There are some faint noise-like data bursts interspersed in the middle.

Let's zoom into the time axis some more.

This signal is a sine carrier that is FM modulated by a 71.4 Hz sine wave.  The delta of the upper and lower modulation frequencies is about 1040 Hz.

This signal looks and sounds a lot like the very strange "flying saucer" signal that was discovered a couple years ago in the setiQuest data.  The measured modulation parameters are slightly different but I wouldn't be surprised if it was generated by a similar mechanism, perhaps by spinning.  Note that this signal's whimsical nickname is a reference to classic B-movie scifi sound effects.  There is zero reason to believe that this signal is from a real flying saucer.  UFO alert canceled.


+2725.7 kHz
This next signal looked like a constant tone until zooming way into the frequency axis.  The spectral resolution of this image is 0.521 Hz/bin.

The tone dips are -15 Hz to the left.  The repeating dips have a periodicity measurement of 61.4 seconds.  The repetition timing is exact like clockwork.  The whole structure is slowly drifting to the right at a 5.73 Hz / 523.3 seconds = +0.0109 Hz/sec rate.  What it is and why it is doing this are unknown.


+3350 kHz
This spectrogram from the far right of the dataset's spectrum is an example of a Doppler flyby.

The top third of the curve is fairly straight and it represents an overhead position.  The measured Doppler drift rate is 22200 Hz / 109.3 seconds = 203.1 Hz/sec. That is fast.

The fainter middle section is curved and represents a transition.

The lower third section still has a slight curve but it is beginning to straighten out.  This motion represents that the object has transitioned to a receding (away) position.  The measured Doppler drift rate is 5067 Hz / 102.9 seconds = 49.24 Hz/sec.

No modulation is visible.  This object is likely a fast moving Low Earth Orbit (LEO) satellite.

Note that with judicious scaling, the basic shape of the Doppler curve above roughly matches the curve seen in the 8-PSK packets image.  The scaling is off by a factor of 265X but the overhead and away transitions occur at about the same times.  The two Doppler signals may be related by some extreme mixing artifacts.


Histogram
The datasets in the setiQuest archive use signed 8-bit quadrature samples.  Below is the histogram plot of the I channel.  The Q channel looks the same.

 The holes between every other bin is normal when 8-bit samples are viewed in 9-bit space (bins=512).

The large number of spikes reveals a rounding problem.  The alternating cadence is a spike every 2nd and then 3rd sample.  Somewhere in the ADC --> beamformer --> data collection chain the higher bit samples are being poorly quantized down to 8-bits.  This sort of error causes alias distortion.

The spike at 0 is interesting as is the drop at the next sample value of 1.  This signifies a truncation round to zero problem. It also will cause a DC offset.

Only about 7 of the 8 bits are being used.  There is plenty of sample headroom and clipping is not a potential problem. Not using 1 bit of resolution reduces SINAD by 6 dB.  Note that SINAD is related to SNR.

For an informative and amusing look at other histogram problems take a look at the Quantization Shop of Horrors.


Conclusion
This setiQuest GOES-11 dataset is a smorgasbord of signals.  We saw FSK, BPSK, offset-BPSK, 8-PSK, and the "flying saucer" modulations with a drifting LEO thrown in for good measure.  The mirror aliasing throughout the entire spectrum is troublesome.  Either the ATA's AGC or the 8-bit quantization is likely to blame.

Links

• http://setiquest.org/forum/topic/baudline-analysis-goes-11
• http://setiquest.org/forum/topic/evolution-vs-intelligent-design#comment-4029
• http://www.youtube.com/watch?v=wGUMcuCp9yY

Data licensed through SETI.
Software licensed through SigBlips.

setiQuest Voyager 1 redux

The NASA Voyager 1 probe has been spotted again at the Allen Telescope Array (ATA). This is fairly amazing since JPL is the only other group that can see Voyager and they use a much bigger telescope that's part of the Deep Space Network (DSN). The SETI Institute's Jane Jordon and Jon Richards were kind enough to send me this new Voyager data so that I could look at it with the baudline signal analyzer. Thank you Jane and Jon.

This Voyager signal was recorded at 8419.62 MHz and was packetized into an .archive-compamp file by SonATA. The following command line streams the 4-bit quadrature Voyager samples into baudline:

cat 2012-11-07_21-14-05_UTC.act61.dx1011.id-4.L.archive-compamp | extract_compamp_data 9 | baudline -session compamp -stdin -format s4 -channels 2 -quadrature -samplerate 711.1111 -record

This Unix command line of multiple pipes extracts the 9th channel from the .archive-compamp file and feeds that into baudline's standard input with the appropriate configuration settings.  Note the .L. in the compamp file name.  I don't know why SonATA does this but the X & Y linear polarization .compamp files are labeled as being L & R circular polarization.

X polarization
Below is the baudline spectrogram of the X linear polarization. Note the weak diagonal signal on the left side that starts on the top at around -160 Hz. This drifting line is caused by the Earth's rotation and the Doppler Effect. I call it Doppler Drift.  Click on the image below for a larger and better view of this weak signal.

Y polarization
Below is the baudline spectrogram of the Y linear polarization file. Note that the signal is bit stronger than it was in the X polarization above. Compare this with the signal strength from the July 2010 Voyager data analysis.

X & Y polarizations
The two spectrograms above were  combined to create this dual channel spectrogram (see the Channel Mapping window). The X polarization is green and the Y polarization is purple. Notice how this dual channel technique makes the weak drifting Voyager signal easier to see. It also makes it easy to visually compare both polarizations.

If you click to look closely at the larger version of this image you'll see several very weak crisscrossing lines and curved whistler-like shapes that are much weaker than the main drifting signal. These very weak features could be real or they could be just random shapes in the noise floor.  Need more data to know for sure. If they are real then they're likely artifacts caused by distortion somewhere in the signal path.


Auto Drift
Baudline's Auto Drift feature was enabled and each of the X & Y polarizations were pasted into the Average window. The Y polarization (purple) is about half a dB stronger which confirms our spectrogram strength observation mentioned above. The drifting signals are about 3 - 4 sigma above the noise floor. The auto drift rate measurement window reports that the signal is drifting by -0.5770 Hz/second.

Next Auto Drift is enabled with baudline's spectrogram display. Think of this as a 4-dimensional version of the above Average display where the additional dimensions are time and vector space. Each pixel has 20865 possible vector paths and that works out to about 10 billion possible vector paths seen in the spectrogram image below. Good thing that Auto Drift is a quick O(n log n) algorithm so that image rendering is rather fast.  For reference the Drift Integrator's settings were: beam width = 326 slices (16 seconds), optimum overlap = 400%, Auto Drift quality = 8.

Notice how the weak drifting signal stands out and is easier to see. Also note how the drifting signal fades in strength between the X (green) & Y (purple) polarizations. This image reminds me a lot of ice crystals on a window pane. It isn't that far out there since the Auto Drift algorithm that generates the spectrogram vectors is somewhat similar to how ice crystals grow.


Histogram
The Histogram window shows 5 vertical lines (bins) that make a Gaussian-like shape (see AWGN). This sparse histogram is not surprising since the input signal has 4-bit quantization. Note that only slightly more than two bits are being utilized.

Waveform
Baudline's Waveform window shows a time series view of what the 4-bit quadrature signal looks like. This sample distribution matches what is seen in the Histogram window above.  Do you see the signal? It's the magic of DSP that allows such a weak drifting signal to be encoded in so few bits.

Conclusion
No modulation characteristics are visible in either the baudline spectrogram or Average window views. The Voyager signal looks like a non-information baring pilot tone. This could be due to the weak nature of the signal or because Voyager was not transmitting at the time the signal was recorded.

As mentioned above, this setiQuest Voyager 1 redux capture is much weaker than the July 2010 Voyager 1 signal I previously blogged about. A good question is why is it weaker?  Here is a list of possible explanations:

• 28 months have passed and Voyager is now 8.3 AU farther away.
• Voyager has entered the Heliosheath and the compressed turbulent solar wind is increasing signal attenuation.
• Voyager's Plutonium-238 power plant has a half-life of 88 years and is producing less power for signal transmission.
• Voyager's dish antenna could be slightly off axis and is not pointed directly at Earth.
• The ATA dishes could be slightly off axis and are not pointed directly at Voyager.
• Reduced collection area due to some ATA dishes being off-line.
• Some cryo feed units require maintenance and are resulting in a higher ATA system temperature.
• An error in the calculation of the ATA beamformer coefficients.
• Increased local RFI mixed with Walshing artifacts are effectively reducing ATA system gain. 
• Rainy cloudy weather in Hat Creek, California.
• The compamp 4-bit sample quantization is pushing up the noise floor. The first Voyager data collection used 16-bit samples.
• Aliens have found the Voyager space probe and are fiddling with its innards.

Some of these explanations are unlikely or not significant enough to account for the reduction in signal strength. What do you think? Please discuss in the comments section.


Links

• http://setiquest.org/forum/topic/baudline-analysis-voyager-1-redux
• http://baudline.blogspot.com/2011/03/setiquest-voyager.html 
• http://voyager.jpl.nasa.gov/
• http://www.nasa.gov/vision/universe/solarsystem/voyager_agu.html
• http://twitter.com/NASAVoyager
• http://twitter.com/NASAVoyager2

Data licensed through SETI.
Software licensed through SigBlips.

setiQuest Voyager

The baudline signal analyzer looked at the setiQuest Voyager data set that is part of the SonATA source code package. The NASA Voyager 1 probe is about 106 AU from the Sun which makes it the most distant object humans have ever sent into space. Voyager is very far away and it's telemetry signal is very weak.

This Voyager data set was collected from the multicast UDP packets output by one of SonATA's channelizers. The format is different from previous baudline/setiQuest analyses. It is 16-bit quadrature samples but more importantly it is packetized and contains two polarizations. A simple strip_header demux program was created for this analysis.

The following command line was used to stream the Voyager data file into baudline:

cat vger-2010-07-14-406.pktdata | strip_header | baudline -session voyager -stdin -format le16 -channels 2 -quadrature -samplerate 546133


8419.53 MHz
The Voyager signal is so weak that the spectrogram looks like noise. Accumulating spectral curves with the Average display (green Fourier below) to reduce the variance of the noise floor doesn't help. Enabling baudline's Auto Drift algorithm (purple below) changes the situation and the drifting signal pops out of the noise.


The standard Fourier transform found a flat featureless spectrum. Auto Drift found a strong signal near +11200 Hz about 10 sigma above the noise floor that has a -0.473 Hz/sec drift rate. Let us zoom in and inspect this signal is greater detail.


+11689 Hz
Decimation by 2048 and a 2048 point FFT for a 0.26 Hz/bin resolution gives a better view. Here is baudline's spectrogram display:


The Voyager telemetry signal is Doppler drifting at a very linear -65.4 Hz / 139 seconds = -0.471 Hz/second rate due to the Earth's rotation. This manual drift rate measurement matches the Auto Drift rate value to within an error ±0.002 Hz/sec.


Conclusion
Baudline's Auto Drift feature made finding the drifting Voyager signal extremely easy. After it was found in the Average display the spectrogram display was zoomed and scrolled into position. A -0.47 Hz/sec drift rate seems a bit high. Need to check with the SETI Institute if this is a reasonable value. Next step demodulation ...


Links

• http://setiquest.org/forum/topic/baudline-analysis-voyager
• http://www.nasa.gov/vision/universe/solarsystem/voyager_agu.html
• http://twitter.com/NASAVoyager2

Data licensed through SETI.
Software licensed through SigBlips.

setiQuest Deep Impact

The baudline signal analyzer was used to analyze the setiQuest Deep Impact satellite telemetry data file. Deep Impact was a 2005 NASA mission to crash a high velocity probe into the surface of comet 9P/Temple and make numerous measurements of the resulting explosion. That mission was successful and now the FlyBy section has been re-purposed to study other space phenomena. The Deep Impact satellite telemetry was captured by the ATA for this data file.

The quadrature data has a sample rate of 8738133.333 samples/second and since the meta data text file was missing the base frequency is unknown but believed to be in the X-band. The duration of the data file is 13 minutes and 23 seconds.

The following command line was used to stream the Deep Impact data files into baudline:

cat 2010-01-22-deep-impact-* | baudline -session setiquest -stdin -format s8 -channels 2 -quadrature -flipcomplex -samplerate 8738133 -fftsize 65536 -pause -utc 0

A Welch windowed 65536 point FFT for a 266.67 Hz/bin resolution was used to create the image below.


The main features are the suppressed carrier (+797199 Hz), the lower and upper sidebands (+597205 Hz and +997194 Hz @ +3.5 dB), and their harmonics (±400000 Hz). It is interesting that the separation between the center carrier and its primary sidebands is ±200000 Hz which is half the spacing of the harmonics. The rest of the band looks clean.

Note that the Histogram display has a nice Gaussian shape with even/odd holes because of the signed 8-bit sampling.


Zero Gaps
The Waveform window shows a gap of zeroes:


The zero gaps happen 3 times in this data file at these time locations with these durations:

• 2:54.9, ∆0.177 s
• 5:19.9, ∆0.115 s
• 5:45.0, ∆0.099 s

The zero gaps are believed to be caused by errors in the data collection process. They likely are dropped UDP packets where the collecting process did not get a full buffers worth of data so it filled in the missing gap with zeroes.


+597205 Hz
Decimating by 512 for an effective rate of 17066.666 sample/sec and down mixing into the main lower sideband at +597205 Hz.


The Average window shows the center subcarrier with some tight modulation products that will be explored in detail later. The center subcarrier has ±800 Hz lower/upper sidebands and their harmonics have spacings of ±1600 Hz. This is the same frequency spacing scheme seen with the main +797199 Hz carrier but the frequencies have been scaled by a factor of 200000 Hz / 800 Hz = 250. [What is the significance of 250?]

Also of interest is the spectrogram display which shows that all the sidebands and harmonics are stationary in frequency at this zoom level. The Histogram has a small spike at the zero value that is explained by the zero gaps mentioned above. Other than that it has a nice Gaussian shape.

Here is the Average window zoomed in with the Hz axis set to 4X.


The three main tones are separated by ±800 Hz. The harmonics next to the center subcarrier have a spacing of ±39.098 Hz. The left/right sidebands have harmonics with ±39.098 Hz spacings. They also have a second set of ±39.098 Hz harmonics that have an offset of 18.262 Hz that likely originated from the center subcarrier (800 - 18.262 / 39.098) = 19.994 which is very close to 20. This means that harmonics of the center subcarrier are interspersed with harmonics of the lower/upper sidebands and they all share a spacing of ±39.098 Hz.

The spectrogram below is the same view as above but the frequency axis has been zoomed in with a Hz=1X setting. The subcarrier and harmonics all have a slight curvature.


Let us take a closer look at the strongest tone in the center by using a decimation by 4096 factor for a 2133.3333 sample/second rate.

This signal has a curved drift. The first half descends in frequency for a -3.06 Hz / 449 seconds = -0.00682 Hz/second drift rate. The second half ascends in frequency for a +2.08 Hz / 351 seconds = +0.00593 Hz/second drift rate. [Describe the possible motion that could cause this Doppler shape.]


+797200 Hz
Decimating by 512 for an effective rate of 17066.666 sample/sec and down mixing into the main suppressed carrier at +797200 Hz. Below is the Average display with a Hz=32X scaling.


The main center carrier has the same ±800 Hz lower/upper sidebands with ±1600 Hz harmonics that was seen above at +597205 Hz. What is missing are all the ±39.1 Hz harmonics. The main center carrier is much cleaner.

Below is the decimate by 4096 spectrogram view.


This spectrogram looks identical to the curved-drifting-random-walk seen above.


+997194 Hz
This is the upper sideband. The modulation and harmonic structures look identical to analysis that was done for +597205 Hz (lower sideband). The data for the spectrogram below was decimated by 4096.

This signal looks identical to the ones shown above at different frequencies. The same drift shape is not a surprise but I would of expected the content to be frequency inverted.


Modulation

In this section the signal modulation of the +597205 Hz main tone of the lower sideband will be examined by down mixing and decimating by 524288. The effective sample rate is 33.3333 samples/sec.

Below is the Fourier spectrogram in magnitude space of an interesting section.

The middle section looks a lot like ten symbols of FSK with an alternating (01)^* pattern and a delta between mark/space of 0.49 Hz. The periodicity bars measured a 1 /3.34 seconds = 0.299 baud rate. Below is the spectrogram of the same section of data but using the blip Fourier transform in phase space instead.

Directly below the FSK section mentioned above looks to be a region of periodic phase changes. Two distinct phases are visible which suggests BPSK modulation. Using the periodicity bars a 1 / 1.68 seconds = 0.595 baud rate was measured which is very close to double what was measured above.

Here is the blip Fourier phase spectrogram of a section a little farther down:

Take note of the numerous phase changes in the center section. Four distinct phases are visible which suggests a QPSK-like modulation scheme. Using the periodicity bars a 0.595 baud rate was measured again.

So we started with what looked like FSK and then jumped to BPSK and then to QPSK. Why? The clues are that the distance between FSK mark/space frequencies were very small and half the baud rate of the PSK measurements. Consecutive 00 and 11 symbols of BPSK can appear to look like very tight FSK. Then what appeared to be BPSK doubled its number of phases and became QPSK. Again QPSK can look like BPSK if only two symbols are being alternated. I believe the FSK / BPSK sections are actually QPSK of a sync header string such as (00001111)^*(0011)^* before the data sequence begins.

The only explanation I have for the extremely low baud rate is that the Deep Impact satellite is in a low power mode to conserve energy. A slower baud rate is also stronger and more tone-like which makes it easier for the Earth based telescopes to find it. This is only a guess since I wasn't able to find any information on-line about Deep Impacts modulation or telemetry schemes.

Self Similarity
Below is the Autocorrelation transform of the magnitude operation.

The horizontal and vertical features were most unexpected. The 3 vertical lines are periodic with a 572 sample (65.4 microsecond) separation. It was a difficult measurement to perform but the periodic bars proved to be very helpful in measuring a horizontal line periodicity of 1 / 0.01672 s = 59.82 Hz. So what is the significance of the 65.4 us and 59.82 Hz orthogonal lines?

Changing the point of view might help. Let us deconstruct the Autocorrelation transform down into its most basic primitive, the sample Raster transform. Below is a display of the sample Raster transform with its overlap width set to 572 samples in the Scroll Control window.

Three chirp-like whistlers are visible and they have a very LC discharge-like shape. There are thousands of these chirps and they are all slowly drifting to the left in time which signifies that the overlap width is fractional and not an exact integer. Assuming that this phenomena is in fact stationary a rough measure of the error is 13 / (145926.8 / 572) = 0.051 samples for a corrected overlap width of 571.949 samples (1 / 65.4469 us = 15279.6 Hz).

The periodic bars measured the chirp-to-chirp periodicity to be 0.01670 seconds which translates to a 59.88 Hz frequency which is suspiciously close to 60 Hz.

This temporally encoded chirp with a 15279.6 Hz line frequency and 59.88 Hz repeating periodicity is a surprising discovery that gets even stranger. Decimating by 2 and moving the down mixer frequency shows that this chirp signal is stationary. This is true all the way down to a decimation factor of 256 at which point the signal disappears. This means that this chirp signal passes through a tuner filter mostly unharmed. Note that this is the same "quadrature magnitude elephant" phenomena that was first seen in the Exoplanet 060 analysis report but the quadrature magnitude operation isn't required to see this particular LC chirp. Another difference is that the Exo 060 sighting focused on 1/3 Fs harmonics and different 25600 Hz subharmonics but 60 Hz was present there too.

Something interesting happens as you continue to decimate down while tuning into one of the sidebands. Here is the Autocorrelation transform when decimating by 4096 and down mixed to the +997 kHz upper side band.


This again was not what was expected. The peaks have a periodic spacing of 54.611 samples which works out to a repetition frequency of 1 / 25.534 ms = 39.064 Hz which was also seen above in the sideband harmonics. Changing our point of view helps show us what is going on. Here is a spectrogram of the sample Raster transform applied to the magnitude operation with an overlap width of 764 samples. Click on the image to see the fine details.

The 14 vertical bars have a thickness of 8 samples (3.750 ms) and a very slight drift to the right with a slope of 9 / (612267 / 764) = 0.011 samples. This is the error term for a corrected overlap width of 764.011 samples (358.1 ms). The 3 longer horizontal black lines have time stamps that correspond to the 3 zero gaps mentioned above. Also of interest are the 40+ shorter horizontal lines that are bounded by the vertical bars. They are not zero gaps but have decreased amplitude. Their origin is unknown and their spacing appears random.

This vertical bar effect is not a function of the magnitude operation. Here is a decimation by 2048 view with the green and purple colors representing the I & Q channels. The changing wavy pattern is the main sideband frequency that is slowly drifting. So the vertical bars are stationary while the sideband signals drift.


Why the vertical bars? Is the 25.534 ms spacing significant? What do the 40+ bar-to-bar amplitude dropouts represent? I don't know these answers but some very complicated signal artifact-ing is taking place.

Filter ExtractionThe impulse response transform was used to compare the I and Q channels. To inverse the positive lag symmetry a Hilbert filter was applied to the I channel prior to calculating the impulse response. This effectively undoes quadrature. Note that the horizontal axis should be time lag (not Hz) and the vertical axis should be a linear scale (not dB).


Compared to all the other setiQuest Filter Extractions I've done the filter shape and the sharp notch at zero are upside down! Another unusual feature is the strong periodic ripple throughout the whole filter. This file had strong carrier tones which is not a good stimulus source like noise is. So this strange periodic ripple could be due to the carrier tones or to a very broken quadrature filter which is possible since this data file is from January 2010.

Conclusion
The Deep Impact signal has a very rich self-similar harmonic structure at multiple levels of zoom. At the top level a slightly suppressed carrier has sidebands that have a spacing of ±200000 Hz with weaker ±400000 Hz harmonics. Zooming in on one of the sidebands reveals a sub-sideband spacing of ±800 Hz with weaker ±1600 Hz harmonics. What is the significance of 200000 Hz / 800 Hz = 250 ratio of the harmonics? Zooming in further shows 39.1 Hz harmonics. Is the rich harmonic structure due to Deep Impact having a non-linear amplifier for radio transmission which is common in deep space probes for power efficiency reasons or are the rich harmonics an ATA collection artifact?

Zooming into any of the strong sidebands or harmonics reveals a drifting signal that looks like a mixed FSK / PSK modulation scheme at a low 0.595 baud rate. I believe the FSK section is actually QPSK of a repetitive header sequence. Deep Impact had a downlink speed in the Mbps range so why do I measure such a very low baud rate? Is Deep Impact in a low power sleep mode?

The Autocorrelation transform found something unusual and very unexpected. Further investigation with the Sample Raster transform shows abundant whistler-like chirps with a ~60 Hz periodicity. Zooming in with decimation by 4096 shows an Autocorrelation periodicity of 39.1 Hz that the Sample Raster shows to be vertical bars with encapsulated dropouts. Note that the finest harmonics of the upper/lower sidebands had the same ∆39.1 Hz spacing. What is the significance of the bars, the dropouts, and the relationship to the sideband harmonics?

This Deep Impact data is an excellent example of what a SETI signal might look like. It is unknown which analysis components are caused by the satellite and which are artifacts of the collection process. This is an important distinction but it really doesn't matter for the purposes of this analysis report because the rich harmonic structure, modulation, and raster features all have an equally fascinating complexity.

Links

• http://setiquest.org/forum/topic/baudline-analysis-deep-impact

Data licensed through SETI.

Software licensed through SigBlips.