How to extract useful data from two-element interferometer fringes
Apeirohedron
Marcus D. Leech
-A two element interferometer doesn't have to be a *fixed* interferometer. You can change the spacing every few days and get new fringes for the same objects. Convolve those
measurements in "stack", and you end up with something approaching a point response. The resulting *amplitude* of that response tells you something about the
brightness of the object along a single axis.
You can also take the (complex) fringe response and compute the complex envelope, which will yield something that is an approximation of the aggregate (convolved)
beam-response of your two antennas to the object in question. Basically, like a radiometer, but MUCH more stable.
The main reason for an amateur building a simple, fixed, two-element interferometer is that it will operate at much closer to theoretical sensitivity, and it tends to
reject or reduce many of the non-idealities of a single-dish radiometer. That *in turn* means that you stand a better chance of getting meaningful science results
out of it.
But I would caution. 99% of amateur radio astronomy is primarily an *engineering* exercise. The average back-yard radio astronomer, using average back-yard
budgets is unlikely to be doing "new science". For back-yard *optical* astronomers, this is also true, but to a lesser degree. Patient back-yard optical astronomers
have detected new comets, etc.
Apeirohedron
Marcus D. Leech
Hi Marcus,
Thank you so much for the response (as always!). I hadn't really given enough consideration to moving the dishes between observations: I was thinking about changing the spacing as being something only feasible for advanced setups that can measure time very precisely, but for a shorter separation like the one I would have it wouldn't be an issue. Similarly, I hadn't considered that even just measuring the amplitude of the complex output from the interferometer is more stable than a single receiver: I keep forgetting that interferometers have advantages outside of higher spatial resolution, but that definitely seems useful.
https://www.cv.nrao.edu/~sransom/web/Ch3.html
IF it were *my* problem to build a two-element "advanced amateur" interferometer, I'd arrange for there to be a rigid structure that spans the entire East-West baseline along
its maximum extent, and then move the dishes to precise locations along that baseline and slowly build up my collection of complex visibilities with unique spacings.
This obviously works best with quite-large baselines. At CCERA, if we ever get the budget, we'll likely build an interferometer that would span about 350m at its greatest
East-West extent, and then put in permanent geo-markers that we move our two dishes to. A lot of unique baselines would be possible that way.
Also, I really do appreciate that new science is basically impossible for amateur astronomy, and especially so for amateur radio astronomy. That's never been my goal: it's just that I want to make sure that, since I'm basically doing a much lower-quality version of research that's already been done, I still have results that are independent enough of the instrument I used that I can see if my data lines up with real experimental results. I'm not trying to find anything new, but I do think that attempting to replicate results from other experiments is a worthwhile scientific exercise, even if there are hard limits on how scientific I can actually be and the experiments I'm trying to replicate are much more advanced than I could ever achieve.
I hope that clarifies my goal somewhat. I really do appreciate what you mean about amateur radio astronomy being mostly an engineering exercise and I'm content with that, so thank you for giving such a clear answer.
Thanks,
Aidan
On Tuesday, January 11, 2022 at 3:23:30 PM UTC-5 Marcus wrote:
On 2022-01-11 14:46, Apeirohedron wrote:
Hi all,
So as I've mentioned several times previously I'm working on a two-element interferometer for the 21cm line. Through people's (incredibly helpful) responses on this forum and trying to find resources on my own, I have a pretty clear idea as to how to build a two-element interferometer high-quality enough for amateur radio astronomy, but one thing I haven't found or heard much information on is how to get useful data out of the fringes such an interferometer would produce.
I know that in professional radio astronomy, aperture synthesis techniques, where the correlation for each baseline can be combined into a 2D image, but these techniques work on the principle that with a large amount of baselines, the fringes will only line up around a single point. As such, it seems that these techniques would require a large number of individual telescopes and are impossible even in theory for a two-element interferometer with only one baseline.
To be clear, this was not my goal, since producing a satisfactory image would require an unreasonable number of dishes. However, I assumed from the fact that so many people on this forum encouraged people to build interferometers that there was something useful that could be done a two-element interferometer if properly constructed. However, after searching for a long time now I haven't found much in terms of advice on processing those fringes into something meaningful about the source being viewed. I know there are algorithms for deconvolving data to recreate the original source, but with only two I'm not sure how accurate that data would actually be.
So my question is this: is there in fact some form of analysis that could yield meaningful information from such a simple setup, or is it the case that on an amateur level, the only reason to make an interferometer is because it's an interesting challenge? I really don't mean to belittle anyone's work in the slightest, but my goal is to find results that can be meaningfully analyzed and compared to 'real' results, and if an interferometer at this scale isn't compatible with that goal I'm not sure I would want to pursue it. This is a weird question to ask, but I think it's important so I can understand what it is I'm getting into, and if anyone does have any advice about analysis techniques for two-element interferometers I really appreciate it.
Thank you,Aidan
-A two element interferometer doesn't have to be a *fixed* interferometer. You can change the spacing every few days and get new fringes for the same objects. Convolve those
measurements in "stack", and you end up with something approaching a point response. The resulting *amplitude* of that response tells you something about the
brightness of the object along a single axis.
You can also take the (complex) fringe response and compute the complex envelope, which will yield something that is an approximation of the aggregate (convolved)
beam-response of your two antennas to the object in question. Basically, like a radiometer, but MUCH more stable.
The main reason for an amateur building a simple, fixed, two-element interferometer is that it will operate at much closer to theoretical sensitivity, and it tends to
reject or reduce many of the non-idealities of a single-dish radiometer. That *in turn* means that you stand a better chance of getting meaningful science results
out of it.
But I would caution. 99% of amateur radio astronomy is primarily an *engineering* exercise. The average back-yard radio astronomer, using average back-yard
budgets is unlikely to be doing "new science". For back-yard *optical* astronomers, this is also true, but to a lesser degree. Patient back-yard optical astronomers
have detected new comets, etc.
--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/3d5d0470-6676-42c0-b3b7-87c627632e7en%40googlegroups.com.
Jim Abshier
I have to agree with Marcus that you would be unlikely to discover
anything new with an amateur radio interferometer. The interferometer is
particularly suitable for detecting weak point-like radio sources. The
challenge is to see how weak a source you can detect. It is sort of like
the ham radio operator who chases DX. Most amateur radio astronomers
don't seem to ever get beyond detecting the 4 strong radio sources. With
a couple of 3 meter dishes many more sources can be detected. Detecting
the really weak ones requires averaging multiple days worth of data. I
have detected about 70 sources including 12 quasars with my
interferometer. The weakest quasar that I detected is 3C298. It has a
flux density of 6 Janskys and a red shift z of 1.438. That means that
the light travel time was about 8.9 billion years. That is better than
half way back to the Big Bang. 25 days worth of data were averaged to
get a good clean fringe plot of 3C298. I didn't discover anything new,
but the thought of capturing energy that had been traveling through
space at the speed of light for 8.9 billion years was rather mind
boggling. The SARA Journal for November-December 2016 has my article on
detecting quasars.
Aperture synthesis imaging with an amateur radio interferometer has been
done. The SARA Journal for September-October 2012 contains my article
describing several earth rotation aperture synthesis experiments.
Because of the limited size of my back yard, the resolution of the
images is not very fine. It is nowhere near what professional
observatories produce. The whole purpose of these experiments was first
of all to see if I could do it, and second to work out the details of
how to do it. Actually doing something leads to better understanding of
the process. These aperture synthesis experiments were done at 400 MHz
so that sufficiently long fringe data records could be obtained without
having to track the source region. For one experiment, two full wave
dipole antennas were used to get data over about 8 hours without having
to track. To get multiple baselines, the antennas were moved to new
positions on successive days. Of course, these experiments did not
contribute to the science of aperture synthesis, but they were fun to do.
In operating an interferometer, one sometimes sees fringes that don't
correspond to what is expected. These might be signals from passing
satellites. Although they may not be of scientific interest, they can be
interesting to analyze. The fringe rate can provide an estimate of the
source declination. This is useful to confirm that you have detected the
intended source. The fringe data can also be compressed by Fourier
transformation for improved signal to noise ratio. With system
calibration, fringe amplitude can provide an estimate of flux density.
Jim Abshier
On 1/11/22 3:23 PM, Marcus D. Leech wrote:
> On 2022-01-11 14:46, Apeirohedron wrote:
>> Hi all,
>>
>> So as I've mentioned several times previously I'm working on a
>> two-element interferometer for the 21cm line. Through people's
>> (incredibly helpful) responses on this forum and trying to find
>> resources on my own, I have a pretty clear idea as to how to build a
>> two-element interferometer high-quality enough for amateur radio
>> astronomy, but one thing I haven't found or heard much information on
>> is how to get useful data out of the fringes such an interferometer
>> would produce.
>>
>> I know that in professional radio astronomy, aperture synthesis
>> techniques, where the correlation for each baseline can be combined
>> into a 2D image, but these techniques work on the principle that with
>> a large amount of baselines, the fringes will only line up around a
>> single point. As such, it seems that these techniques would require a
>> large number of individual telescopes and are impossible even in
>> To be clear, this was not my goal, since producing a satisfactory
>> image would require an unreasonable number of dishes. However, I
>> assumed from the fact that so many people on this forum encouraged
>> people to build interferometers that there was something useful that
>> could be done a two-element interferometer if properly constructed.
>> However, after searching for a long time now I haven't found much in
>> terms of advice on processing those fringes into something meaningful
>> about the source being viewed. I know there are algorithms for
>> deconvolving data to recreate the original source, but with only two
>> I'm not sure how accurate that data would actually be.
>>
>> So my question is this: is there in fact some form of analysis that
>> could yield meaningful information from such a simple setup, or is it
>> the case that on an amateur level, the only reason to make an
> --
> You received this message because you are subscribed to the Google
> Groups "Society of Amateur Radio Astronomers" group.
> To post to this group, send email to sara...@googlegroups.com
> To unsubscribe from this group, send email to
> sara-list-...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/sara-list?hl=en
> ---
> You received this message because you are subscribed to the Google
> Groups "Society of Amateur Radio Astronomers" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to sara-list+...@googlegroups.com
> <https://groups.google.com/d/msgid/sara-list/7fd31780-eaec-0317-02fc-ec4ce258d6dc%40gmail.com?utm_medium=email&utm_source=footer>.
Dan Layne
Jeremy Waller
B. J. Rickett
Department of Applied Physics and Information Science, University of California, San Diego
Received 1974 August 30
Marcus D. Leech
> Aidan,
>
> In operating an interferometer, one sometimes sees fringes that don't
> correspond to what is expected. These might be signals from passing
> satellites. Although they may not be of scientific interest, they can
> be interesting to analyze. The fringe rate can provide an estimate of
> the source declination. This is useful to confirm that you have
> detected the intended source. The fringe data can also be compressed
> by Fourier transformation for improved signal to noise ratio. With
> system calibration, fringe amplitude can provide an estimate of flux
> density.
>
> Jim Abshier
correlation/multiplying interferometers is that constant narrowband
interferers simply create a DC offset in the output.
You can simply DC-block the output of the correlator, and those
offsets just go away.
Similarly, gain drift isn't as big a problem with interferometers. It
still happens, but only really affects the amplitude (slightly) of the
fringes. It *cannot*
"fool" you into thinking that you're seeing an object, since if there
aren't fringes, there isn't an object within the detection limits of the
interferometer.
You don't get fringes unless the object is a point-source (with
respect to the interferometer grating response), and the object is
moving relative to
the baseline.
When Ken Tapping and I built the "differential interferometer" in his
back-yard all those years ago, we got lovely fringes from passes of
Orbcomm satellites.
This was a 137MHz instrument--one end was a Yagi, the other was a
simple dipole (from what I recall--it has been many years now!).
Michiel Klaassen
--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/03397085-ae13-848f-a002-bf4b64703070%40gmail.com.
Larry Mayfield
I love it! I happen to have two identical dishes like that. So top of my list for the future when health is a bit better and it is not so cold for an old geezer like me. I will start gathering the necessary parts now….
Larry
Pahrump, NV
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/CAABBjRBExah2%2BEPifRj0yeNiPK%2B3Ri8XxSiq51eS4GsxaXWRwA%40mail.gmail.com.
Marcus D. Leech
If your garden is not so large you also can build a interferometer on 12GHz with thesame $10 LNB'sseeIf down then see att.Michiel
device when modifying them, due to the quite-small circuitboard, and the PTFE substrate doesn't "hold on" to the traces very well.
Pointing accuracy becomes an issue at these frequencies with anything other than a quite-small dish. Expect to spend a lot of
your time and $$ getting that right.
I bought a number of *C-band* PLL LNBFs a few years back, and they are MUCH easier to modify for a common (25MHz in this case)
reference clock. Unfortunately, with C-band being re-structured to include terrestrial services, it may be a lost-cause for radio
astronomy work now, sadly. But the pointing requirements are more-relaxed (still not "casual"), and sources are a bit brighter.
I used the Titanium C1W-LIte units, and modified a few of them for external clock.
Larry Mayfield
Is there a detailed parts list for this interferometer? Sure would save a lot of time of so. What are the feeds for the two antenna? Home built or ? Any help out there? This looks to be a great project?
And some questions regards construction. I know zip regards these methods: so dumb question #1) Is there any benefit to have the antenna mounted to a moveable truss so that the reflectors could always be perpendicular to the target? And with tracking of the dish in both azimuth and elevation? Does that help anything? Just a few words needed here – like what a dumb a$$ question, or yeah that would work great or no benefit… Dumb question #2 would longer recording times be of any benefit as this kind of mount might give a lot more data to work with per session time: record for 30 minutes, repeat every half hour as long as target was visible to apparatus. Repeat for days on end if necessary. Make any sense at all? Dumb question #3. What is a rule of thump for antenna spacing if there is one? Wider better? Point me to a simple tutorial or spreadsheet?
Or just ignore me, lol…
Larry
Pahrump, NV
From: sara...@googlegroups.com <sara...@googlegroups.com> On Behalf Of Michiel Klaassen
Sent: Wednesday, January 12, 2022 8:02 AM
To: sara-list <sara...@googlegroups.com>
Subject: Re: [SARA] How to extract useful data from two-element interferometer fringes
If your garden is not so large you also can build a interferometer on 12GHz with thesame $10 LNB's
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/CAABBjRBExah2%2BEPifRj0yeNiPK%2B3Ri8XxSiq51eS4GsxaXWRwA%40mail.gmail.com.
Marcus D. Leech
Is there a detailed parts list for this interferometer? Sure would save a lot of time of so. What are the feeds for the two antenna? Home built or ? Any help out there? This looks to be a great project?
feed-horn.
And some questions regards construction. I know zip regards these methods: so dumb question #1) Is there any benefit to have the antenna mounted to a moveable truss so that the reflectors could always be perpendicular to the target? And with tracking of the dish in both azimuth and elevation? Does that help anything? Just a
considerably in keeping the two dishes pointed in the same direction.
few words needed here – like what a dumb a$$ question, or yeah that would work great or no benefit… Dumb question #2 would longer recording times be of an
The larger the spacing (in wavelengths) the higher the spatial resolution (the faster the fringes are produced). Oncebenefit as this kind of mount might give a lot more data to work with per session time: record for 30 minutes, repeat every half hour as long as target was visible to apparatus. Repeat for days on end if necessary. Make any sense at all? Dumb question #3. What is a rule of thump for antenna spacing if there is one? Wider better? Point me to a simple tutorial or spreadsheet?
your resolution starts getting smaller than the objects in question, the fringes start to disappear.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/001901d807db%243808e950%24a81abbf0%24%40mayfco.com.
Jim Abshier
I managed to dredge up a draft copy of the article on aperture synthesis
that was published in the SARA Journal. A copy is attached.
On the topic of what can be done with an interferometer, I have attached
a couple of examples. The first example is an earth rotation aperture
synthesis image of Cygnus A and Cassiopeia A. This image was produced
using a _single_ interferometer baseline. The right half of the image
shows the real part of the visibility in the uv plane. The image is a
polar projection, so the uv tracks are in a circle. I don't recall if
the fringe data for both sources were collected in a single pass or if
fringe data from separate passes were projected to the uv array. The
image on the left is a whole sky polar projection. Right ascension is
portrayed as an angle increasing clockwise from the top. The circle
represents zero declination, and the radial distance from the center
represents the cosine of declination. The image is perhaps not very
exciting, but it can represent a first step in building a capability for
earth rotation aperture synthesis. Adding more data from additional
baselines would make the image more interesting.
The second example is a spectrogram image produced from fringe data of
galaxy 3C295. Fringe data from various levels of processing are shown on
the right side of the image. The lower trace is fringe data from a
single days collection. This raw data is collected with 10 second
averaging. The middle trace shows the average of 16 days collections.
The top trace is a low-pass filtered version of the 16 day average. The
spectrogram on the left is produced using a sliding FFT. A window about
the size of the antenna beam is slid over the fringe data, and at each
data point, an FFT is taken of the fringe data falling within the
window. The frequency data is displayed as columns that are stacked side
by side in the horizontal direction. The horizontal axis is then right
ascension and the vertical axis is cosine declination. By taking the FFT
of the fringe data the fringes are compressed to a point and the signal
to noise ratio is improved.
These examples show some additional things that can be done with a
single baseline interferometer.
Jim Abshier
Marcus D. Leech
*HAD NOT* seen before.
(1) Wow! This is really really excellent work.
(2) With a modern SDR-based approach, you get complex fringe data
automatically. Which makes downstream processing easier.
(3) What bandwidth was actually used? At 400ish MHz, you're bandwidth
limited by RFI more than anything else.
(4) I'd love to see more of a description of the software steps (in
Octave, I guess) that were used to take the several-hour-long
interferograms at each baseline setting and
first produce the 2D visibility array, then the fourier (2D FFT?)
step.
(5) One could likely do this at 21cm without tracking using a small horn
antenna, and have shorter recordings--without tracking, but perhaps with
higher absolute bandwidth
to improve sensitivity. Even 20Mhz bandwidth at 21cm is not that
large in terms of fractional bandwidth, so bandwidth decorrelation will
be a very minor problem.
(6) Again, this is truly excellent work.
fasleitung3
Hi all,
Since this work has been performed at our observatory I can add some
detail.
First lat me say that the "mastermind" behind this Horst Thum (DK2KA)
who is our expert for anything at higher frequencies (Ku- Ka-band). The
LNBs used in this setup were PLL based devices from Sharp. These have a
25 MHz quartz oscillator from which the LO frequency is derived. The
quartz has been removed and instead a 25 MHz external reference signal
has been injected. This reference signal has been provided to both LNAs
via two phased matched cables. The same scheme can also be set up with
other PLL based LNBs such as an Octagon LNB. In case of the Ocatgon the
reference osciallator is at 27 MHz. A description on how to modify the
Octagon LNB has been described by Andy Talbot:
http://www.g4jnt.com/OctagonExtLo.pdf
There may be quite a few other types of LNB around which allow the
same kind of modification.
The setup at our site uses two 1.2-m offset dishes with a baseline of
about 10 meter in East-West direction. It is a pure transit instrument.
Both dishes are aligned to look due south. The declination, however,
can be adjusted via linear actuators. The declination angle is measured
using inclination sensors of the type ADXL 203. These are sufficiently
accurate as the beam width is about 1.5° with the 1.2-m dish. As
pointed out by Marcus, this will become a bigger challenge with larger
dishes.
A scan of the sky visible to this instrument has been performed. I have
attached a plot of this survey with the clearly identified sources
annotated. It still needs to be investigated whether the other signals
visible are real and, if so, what these sources are.
Best regards,
Wolfgang
Jim Abshier
Thank you for your complementary remarks. The bandwidth I used at 400
MHz was about 200kHz. More recent attempts to use 400 MHz have been
discouraging because of increased interference. There are several
versions of the octave programs (scripts) that were used to do the
aperture synthesis processing. Of course, none of them are commented, so
they might require a lot of explanation. I will have to go through the
various scripts and see if I can come up with something that explains
what processing was done. There are 3 separate programs. One processes
the fringe data into a dirty image. Another uses my version of the
Hogbom CLEAN algorithm to produce the cleaned image from the dirty
image. A third program transforms the cleaned image from l,m coordinates
to right ascension and declination. There may be programs available from
NRAO to do CLEAN and transformations, but you would have to figure out
what data formats etc. they require. I prefer to develop my own
programs. That way I know what they are doing.
Jim Abshier
Hamish Barker
--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/fafe124f-359e-8f86-5675-934ceec553ba%40ameritech.net.
Marcus D. Leech
> Marcus,
>
> Thank you for your complementary remarks. The bandwidth I used at 400
> MHz was about 200kHz. More recent attempts to use 400 MHz have been
> discouraging because of increased interference. There are several
> versions of the octave programs (scripts) that were used to do the
> aperture synthesis processing. Of course, none of them are commented,
> so they might require a lot of explanation. I will have to go through
> the various scripts and see if I can come up with something that
> explains what processing was done. There are 3 separate programs. One
> processes the fringe data into a dirty image. Another uses my version
> of the Hogbom CLEAN algorithm to produce the cleaned image from the
> dirty image. A third program transforms the cleaned image from l,m
> coordinates to right ascension and declination. There may be programs
> available from NRAO to do CLEAN and transformations, but you would
> have to figure out what data formats etc. they require. I prefer to
> develop my own programs. That way I know what they are doing.
>
> Jim Abshier
processing, but I've had to fight with them quite a bit, and it took me
quite a while to figure
out how the very-simplest input formats worked. The SIGPROC
filterbank format is used quite a bit, but the technical descriptions
are either inadequate,
or often plain *wrong*. Or you find out that one of the features
is just plain not-implemented, etc. One of the hazards of letting
astrophysicists write
(and, worse, publish) code :) :)
I have another question specifically about your 2009 Cygnus A image.
How did you, given the relatively-airy nature of your primary beam, avoid
significant contamination from Cass. A ?
CCERA is currently focused almost entirely on the FRB project, but we do
have "build another interferometer" on our list of things to accomplish
at the new site--
the site is large and flat, and, well, just begs for an
interferometer. If I can apply the relatively simple (hah!) techniques
you've outlined in your paper, it should
produce some really nice results...
fasleitung3
Thank you very much for providing this. It is excellent work and we can
learn a lot from that.
Wolfgang
Jim Abshier
I think that the reason Cassiopeia A does not appear in the image is
because interferogram data outside the window of interest is edited out
in each record as a preprocessing function. My records are standardized
as files with 8640 samples covering a full 24 hours at 10 seconds per
sample. A days run typically had a lot of other stuff in addition to
what I was interested in, including strong solar fringes. Each fringe
data set had a separate script that was executed to load the data,
define limits of good data, and set parameters specific to that
particular data set. These scripts were called by the main program when
processing the individual fringe data sets. I examined one of the data
sets and found that some fringes from Cassiopeia A were there, but they
were mostly outside of the defined data boundaries. Also, the Cassiopeia
A fringes were of lower amplitude than the Cygnus A fringes because Cas
A was down near the half power edge of the antenna beam. Although the
antenna beam was very broad in right ascension, it was only about 40
degrees wide in declination.
Jim
Marcus D. Leech
> Marcus,
>
> I think that the reason Cassiopeia A does not appear in the image is
> because interferogram data outside the window of interest is edited
> out in each record as a preprocessing function. My records are
> standardized as files with 8640 samples covering a full 24 hours at 10
> seconds per sample. A days run typically had a lot of other stuff in
> addition to what I was interested in, including strong solar fringes.
> Each fringe data set had a separate script that was executed to load
> the data, define limits of good data, and set parameters specific to
> that particular data set. These scripts were called by the main
> program when processing the individual fringe data sets. I examined
> one of the data sets and found that some fringes from Cassiopeia A
> were there, but they were mostly outside of the defined data
> boundaries. Also, the Cassiopeia A fringes were of lower amplitude
> than the Cygnus A fringes because Cas A was down near the half power
> edge of the antenna beam. Although the antenna beam was very broad in
> right ascension, it was only about 40 degrees wide in declination.
>
> Jim
I had assumed that the beam was uniformly "airy", but that was a wrong
assumption...
Marcus D. Leech
>
> I think that the reason Cassiopeia A does not appear in the image is
> because interferogram data outside the window of interest is edited
> out in each record as a preprocessing function. My records are
> standardized as files with 8640 samples covering a full 24 hours at 10
> seconds per sample. A days run typically had a lot of other stuff in
> addition to what I was interested in, including strong solar fringes.
> Each fringe data set had a separate script that was executed to load
> the data, define limits of good data, and set parameters specific to
> that particular data set. These scripts were called by the main
> program when processing the individual fringe data sets. I examined
> one of the data sets and found that some fringes from Cassiopeia A
> were there, but they were mostly outside of the defined data
> boundaries. Also, the Cassiopeia A fringes were of lower amplitude
> than the Cygnus A fringes because Cas A was down near the half power
> edge of the antenna beam. Although the antenna beam was very broad in
> right ascension, it was only about 40 degrees wide in declination.
>
> Jim
So, just going through the paper again, and other treatises on aperture
synthesis using interferometry.
If we assume a perfectly east-west baseline, and a perfect phase-center
on the resulting interferometer:
o We have a set of fringe data covering all our baselines
o We assign the transit-time of the object as the notional origin
of our u,v matrix/array
o For any given point in a given fringe data set, we place that
point within the u,v matrix according to
u = the baseline associated with this dataset
v = the calculated baseline rotation amount relative to our
origin (transit time)
o For point read "complex point" -- the fringe data are complex
correlations with an I/Q, Real/Imag, COS/SIN component.
o repeat for all baselines
o zero out anything we dont want in that dataset
o apply a 2D FFT
o scrub and CLEAN until happy
Purely coincidentally, I worked and filed a patent with David Steer in
my last year at Nortel. He did his graduate work in Radio Astronomy,
and came up with
significant improvements to CLEAN, many of which are used at
observatories all over the world now...
Larry Mayfield
List, be gentle as you read through my scribbles below:
I need your help! I have been staring at the interferometer diagram in the pdf file that Michiel sent to the list sent to us, second page of the file, and am stumped on several components. I am gathering the components in the electronics 2nd page schematic to build an observatory exclusively for 1420 MHz with maybe a +-50 MHz band width. I have a pair of DISH offset dish hardware that I am going to use. But I am having some trouble understanding some of the items and how to use them. The DISH components I have, use I think, LNA, rather than LNBF pieces for getting the signal and passing it along. However I do not know that for sure! What else is not clear is the brand used in this schematic since I will need to find and replace the internal Xtal with an input from the local 25Mhz oscillator shown on the schematic. I have no clue as to what kind of LNBF would be able to use with the DISH reflector: I ask because some of the reviews of available LNBF units have indicated that they are not useable with Dish equipment. This might mean simply that an LNBF works differently than DISH receivers are expecting to see. I just do not know, lol. Any one have ideas as to a brand name LNBF that might work with the dish reflectors as that is the only part of DISH I use? The next issue, for me, filters used to reduce the bandwidth of the signal being collected and processed. The schematic show a low pass filter in series with what really appears to me to be a high pass filter and the combo seems to make a bandpass filter. The filter 3 db corner points appear to be 950 MHz for the low pass and 656 MHZ for what I call the high pass filter. I thinks that makes the 3 dm bandwidth of the pair about 803 MHz. Would that make the LNBF downconverter output frequency be around that frequency? If so, then in that case could I be looking to find a Bandpass filter with values of 3 db points at 656MHz and 950 MHZ points? Seems reasonable to me, but as I said, theory of electronics is not something I am very familiar with, lol. Any advice for me in this regard? And finally, the A/D converter: I am planning on using an Arduino Mega 2560 microcontroller which has an ESP32WRM wifi on board and it will do all the on platform signal collector for my GPS Timing Signal for UTC time, latitude and longitude, a weather station module for atmosphere pressure, temperature and humidity; the pointing angle of the dish(s) as outputs from two high resolution rotary encoders, as well as the schematic output signals for processing as digital signals. These and probably more will be passed to a Raspberry PI 4 (I may use my Pi 3 B+ however as it now collects the GPS and weather data in my office already), The Pi will be used to host a VPN system (I have the basic understanding of that) for building an html system after processing the data even more. My question is what resolution of ADC do I need to add to the arduino Mega because it is only a 10 bit ADC but that doesn’t seem like enough. I have a 12 bit, 14 bit, a 16 bit and I think a 24 bit ADC modules kicking around unused right now for signal processing.
I have already purchased some of the items but the above items stump me. So, any thoughts or ideas for me?
In any case, I thank you all in advance for any help and light you might be able to shine on my questions!
Larry
Pahrump, NV
From: sara...@googlegroups.com <sara...@googlegroups.com> On Behalf Of Michiel Klaassen
Sent: Wednesday, January 12, 2022 8:02 AM
To: sara-list <sara...@googlegroups.com>
Subject: Re: [SARA] How to extract useful data from two-element interferometer fringes
If your garden is not so large you also can build a interferometer on 12GHz with thesame $10 LNB's
see
If down then see att.
Michiel
.
Larry Mayfield
Jeremy, and all ya’ll others who might find it interesting…
I do all of my own work in building my structural hardware such as the mounts, alt and az drives design and fab work as well as the electrical & electronics work. I am retired and use as much recycled components as I can. As to the electronics, involved, I shop around. Most all of the electronics, except for the feed horns which are available via either Banggood or Amazon and I have purchased most of these already. For instance the 25 MHz clock, is a standard clock used for PLL work and the one I purchased (well, I buy a second unit for replacement in case of failures), cost me about 7 bucks. Same with the rest of the components on the schematic (yeah, it is pretty poor in information) I have the logarithmic rectifiers as modules, the Ad8203 and the summing mixer also which has I forget the number, but is an ADxxxx also. The splitters are about 10 bucks apiece. I have a pile of Arduino microcontrollers from the nano to the mega versions all with either com port connections, ethernet or wifi capabilities. I have a Raspberry pi 3B+ that may become the web host for the system. I have 2 dish 500 parabolic offset reflectors with feeds that will be replaced with single feed ones and they are about 15 bucks each. They have to be modified to remove the 9MHz – 10 MHz Xtal, and then the input from the oscillator clock added in its place. This 25 MHZ clock can be tuned from 8Mhz to 160 MHz, either integer or fractional Hz capable, with input control from the microcontroller. I build my own cables and for the high frequency and I use semi rigid coax which has a flexible but stiff outer shielding. I am mounting both of my offset dishes to a common, but moveable beam that is the pointing part for the pair of them as this is going to be a tracking system. I am currently in the design phase of my project for the alt az drive and have devised a way to fab large drive “gears”. The alt part of the system will use a periodic pulse to move it and will be guided in time by the period that the signal will be in the main part of the receiver dish reflector. I have a pair of high resolution rotary encoders to tell me and the system what it is pointing at. Probably, all in all, I have a couple hundred bucks into my system right now if spares are included. So, I am cheap, lol. Will it all work? Well the last system I built worked first time out of the box so to speak. I will use an FFT to look at the channel that comes from Summing Mixer which is actually a multiplication function (ie it is really modulation and when using the log rectifier produces this effect) when examined. I am looking at the signal characteristics only and have used NI’s Multisim software to first help me define what I am searching for, building a test signal, and then using multisim and FFT to tell me what it decodes as. All is good for what I am going to do. I have planned on installing the on board electronics suite in a weather tight metal enclosure between he pair of dish mounted feeds to shorten the connections and eliminate any spurious RFI, if I can. I plan on using TCP over wifi to get he signals through my VPN dedicated router to my main computer for doing all the collection, display and analysis work if any.
The bottom line for all of this is for me to have fun and learning new things before I step off the mortal coil. Mostly as cheap as I can. I am also gong the have a full power point set of charts describing my system as well and hopefully will host it along with my commercial free web site showing all of the charts and info pertinent to the is observatory operation. Not there yet however.. But you can look at my auto portion I built when I was racing my land speed record holder car. It is http://www.mayfco.com, just click on the first page active link and get the index page. Then surf around. It is commercial and user free and has a lot of car related information on it, just no astronomy related stuff yet. Oh, I built a magnetometer using active IC pieces from Texas instruments some of that data may be included as well. I have an Arduino UNO connected to a BME 280E module weather station as well and a GPS module that are providing both temperature, humidity and atmospheric pressure in psia as well as UTC time, latitude and longitude for my basic location and operating parameters. I am currently peeved at myself for inability to find canned dashboard data visualization items to display such things as pointing angles, signal levels, etc… But I will get there, hopefully.
Still seeking the data visualization and process control elements for building a control system and its dashboard…. Free stuff if possible… things like analog meter displays the pointing angles, time, active buttons, and displays for things like voltage levels, status, ect. Things that can be manipulated by the mouse or keyboard as entries and display.
If anyone has questions, you can use the email link on the web site mentioned above, and pending health and other pressing issues I will get back to you!
Thanks all ya’ll
Larry, Pahrump
Yup, that’s me… who woulda thought…
From: JRW <wall...@bigpond.com>
Sent: Sunday, February 13, 2022 3:37 PM
To: Larry Mayfield <drm...@mayfco.com>
Subject: Re: [SARA] How to extract useful data from two-element interferometer fringes
Thank you for your reply Larry.
This Radio Astronomy is a very expensive business so from my point of view it is best to fully understand exactly the system before buying components. My response to you is to pose some questions you may ask of the authors.
Re: For one thing this particular design, and I did not come up with it, so don’t shoot the messenger.
Oh no ! Please do not read me wrong. I am merely saying that the author(s) of that block diagram need to or should provide much more detail to understand exactly what's going on. For example those filters are not specified - just 2 numbers (?).
Question: (That needs to be asked) What is the input frequency ? Given that the authors have said in the title "Observational Results with a Ku-band Interferometer" where as Ku Band is centred on 10GHz. in slide 3 the authors mention 10GHz. - well removed from 1.42GHz.
About the Analog to Digital Converter.
Those signals (F1, E, F2 etc) these will need to be suitably band limited before sampled in the ADC - notice, no mention is made of this. As for the bit depth a 10 bit ADC will give a dynamic range (volts) of 30db. High sampling rate(say 20 MSPS) 12 or 14 bit ADC's are very expensive.
Perhaps the bandwidth of the signals from the log amplifiers are quite low say less than 16KHz. in which case a USB sound card (like : Volans VL-UA01) will do the job very well - 2 channels, 16bits , *.wav format etc.
So, from my perspective, I would ask the Authors to provide considerably more detail on their project before I would start buying components and start construction.
Kind Regards,
Jeremy. (VK5WJ).
On 14/02/2022 09:03, Larry Mayfield wrote:
As I mentioned, I am not the sharpest pencil in the pocket protector, however, and so the reason for the questions I asked. I may make a lot of dumb responses here so forgive me?
This is an interferometer first off. So things are done a bit differently I suspect. For one thing this particular design, and I did not come up with it, so don’t shoot the messenger. The feed horn and internals are an LNBF which is an antenna that has its low noise amplifier and the frequency down converter built in. And since in an interferometer one data stream is mixed with the other, each data stream must of some necessity use the same clock rate and phasing. Else when mixed to get the multiplied ( the by the summer) to get the overall signal output to be accurate, they have to be precise and in tune. So the built in local oscillator Xtal is removed and the single source 25 MHZ signal used instead for each feed. Yes, fiddling with it is going to be required to get what I want to do. A couple of the reasons for doing an interferometer are irresolution and increase in sensitivity. And yes, I am limiting this set up to watching and dissecting the Hydrogen line. I do want some bandwidth because of doppler shifts from far away signals and also our local system doppler effects. So, some adjustments will be needed.
As far as I know the filters are to remove unwanted noise and signals from being processed downstream. For me that means getting rid of unwanted signal junk on each side of the pass band that is not wanted and to do it early in the process stream. It also helps to reduce the size of recorded files if any are recorded. From what little I may think I know, the low pass filter and the one called bandpass are used to make a PassBand filtering system. The low pass filter lets all signals up to the cutoff point, defined as a the 3dbm point and go through and what is above that frequency is reduced severely; 0 to 950 Mhz is good and above 950MHZ is eliminated so to speak. The Bandpass filter lets only signal from 656 MHZ pass. The result is a Passband that is essentially letting signals and anything else that’s between 656 MHz and 950 MHz though into the processing area of the circuit. The down converted signal from the LNBF must of certainty fit into the is filter parameter range so tweaking will in all likelihood be required. In my plan I will probably put in a switch selectable bypass around the filters to process the entire signal, if for nothing else, than chuckles.
The job of any ADC is to convert analog signals into digital ones for processing by computers or microcontrollers. The AD8302 recives analog frequency signals and process those as analog outputs. The signals that the Logarithmic rectifiers receive are analog signals as well and the Log rectifiers process thsm into analog signals as well. But the digital processing of the computers and micro processing systems require digital signals from those devices. Or each of their outputs are input ino one of the ADC channels, probably I would use A0, A1, A2, A3 and A4 and them put those out as digital words after sampling the signals. Then those can be further processed and displayed on screen or recorded. My question regards eh ADC were in respect to just how much sampling , ie sample rate is effective and noting that the built in ADC on the Arduino I will use is only a 10 bit unit; that is pretty coarse for really defined signals, but in truth for my system and what I am looking for it might be good enough. I do have a number of ADC units though co I can cobble something together if needed for very fine microvolt resolution.
As for me? I plan on finishing the design work and then building my system. I enjoy the pleasure of that as much as anything I do, lol. Heck, I am an engineer. Not an electronics one, but structural dynamics, aerodynamics, and some others I can’t and won’t admit to, and have fun. I used to be a test wienee so I know how to plan and do things…
For all of you who see mistakes in what I just wrote about above, just ignore it and if you have advice for me, then pass that along, I did and do need information, lol. Criticism, of my capabilities is moot and not needed, I know where I am on the electronics knowledge charts, and that is near the bottom.. no use beating a dead horse…
Larry
Pahrump, NV
From: JRW <wall...@bigpond.com>
Sent: Sunday, February 13, 2022 2:05 AM
To: Larry Mayfield <drm...@mayfco.com>
Subject: Re: [SARA] How to extract useful data from two-element interferometer fringes
Hi Larry,
Before spending too much on this project the diagram in the referenced link needs considerable clarification and explaination.
1. You are talking about an observatory for 1420MHz. But the document title talks about Ku band and this is centred on 15 GHz.
Question: Is the frequency (25MHz. ) of the oscillator correct? Mixing the environment with a 25MHz. source will produce sidebands up and down by only 25 MHz.
Question: What is the centre frequency and bandwidth of the band pass filter and how is related to the 25MHz. local oscillator.
Question: Is the low pass filter really needed ? The BPF should suffice.
Question: In the block diagram the ADC has 4 inputs ??? What is going on here ?
If I were to build this system I would require much more information than is presented in this document. There is too much left out.
Regards,
Jeremy (VK5WJ).
--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to a topic in the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/sara-list/4xeiP1txipc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to sara-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sara-list/000001d82049%24b593e6a0%2420bbb3e0%24%40mayfco.com.