SDR HF radioastronomy

[spectrahm]

Hello everyone,

With a little bit of spare time, I've started the construction of a jovian antenna. The usual jovian antenna is a dual half wavelength dipole (lambda=15m), which requires an unreasonable amount of space for an urban dweller. Also the array is connected to the receiver with quite a long length of unamplified coax, with the incumbent loss.

My idea to make this doable is to use a single dipole, directly attached to an LNA then to a SDR receiver.

With this in mind I've ordered an LNA4HF, built a casing and attached it to the dipole and to my Airspy. I get some signals from hams everywhere around and more (heard three different languages), although no jovian s-burst or l-burst. The only curious feature are short bursts on the noise floor, spanning all the way up to ~50MHz so I doubt this has anything to do with Jupiter.

I'm listening at 20.2MHz with an AM demodulation (gqrx).

Maybe this is not the right way to do it with an SDR receiver? Maybe is there some signal post-processing to do?
Maybe also there are very specific times to listen to the bursts. I understand they are caused by Io acting as a dynamo inside Jupiter magnetic field, and occur mostly during some specific alignments. How can I predict these bursts, for instance using Celestia?

Anything else that I should know before going further?

Thanks in advance,
Damien

[Dave Typinski]

Hi Damien,

You should be able to observe Jovian emission at around 20 MHz using the setup
you describe. A few points, though:

In general, observing in the HF band is limited not by receiver noise, but by
the galactic background noise (and band noise during the day). A 6 dB feed line
loss will usually not affect your ability to observe Jupiter unless your
receiver has an unusually large amount of internal noise or is just plain not
very sensitive -- which may very well be the case with one of those wee SDR
dongle receivers. The LNA will help in this regard, as will having lower loss
in the coax by making it as short as practical.

So, the LNA allow you to overcome the receiver's internal noise and/or lack of
sensitivity. But what of the galactic background noise?

In general, the way we overcome the galactic background is with directivity.
That is the whole point of an array of antennas: the assembly is more sensitive
to one spot in the sky at the expense of sensitivity everywhere else.
Unfortunately, you won't be able to obtain much directivity with a single
dipole. A low noise amp cannot bring Jupiter above the galactic background
since it amplifies both signals equally.

We've covered solutions for the galactic background and for insensitive
receivers. What about band noise?

The answer there is easy: observe Jupiter at night, when the ionosphere is too
thin to support terrestrial propagation at your observing frequency.

I do not know of any post-processing /required/ for Jupiter observing.

Radio Jupiter Pro is a software application that can tell you when the
Earth-Jupiter-Io geometry is a favorable configuration to produce observable
Jovian emission.

http://www.radiosky.com/rjpro3ishere.html

I also recommend reading the book "Listening to Jupiter" by Richard Flagg. It's
on the Radio Jove Reference CD.

http://radiojove.gsfc.nasa.gov/office/order_form.html

Hope that helps, Damien!
--
Dave

> --
> --
> 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
> <mailto:sara-list+...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

[spectrahm]

It sure did help, Dave, this answered all my questions. Thank you!

Apparently it's just a matter of observing during the storms. The program you mentioned will be handy for that because they are short and only occur once or twice a day.

Examining the program, it seems like I should be able to predict the storms by myself with the position of Io and Jupiter. Maybe I'll develop a Python program later to make this open and cross platform. If I do, I'll publish it on my GitHub account for those who are interested.

Being still new to the HF world, dipoles are kind of a magic science. To me, the Jovian dipole is quarter-wavelength since only half of the dipole is connected to the probe, the other half being connected to the ground. When I also connect this second half to the probe I get a stronger signal, which sounds right to me but should not if I got the dipole physics right. I understand some resonance should occur between the two poles (ref) and increase the gain, so I'm not sure why I got a better result when I transformed it into a monopole.

Do you think using a monopole could be a way to improve my chances of success, if it seems to have a better gain than a same length dipole?

I'm also still in the process of determining the impedance of the dipole before building a balun for it. So far it works great without it but the signal often “pumps” and I'm wondering if this could not be the culprit.

As always, the lower the frequency, the more challenging it gets.

Thanks again for your answer,
Damien

[Dave Typinski]

Hi Damien,

Why do you feel the monopole configuration seems to have better gain? Is there
an increase in local signals, an increase in background noise, an increase in
interference, or....?

The impedance mismatch is usually of little concern unless it is so horribly
mismatched that nothing is getting through. For example, an SWR of 4:1 only
produces a 2 dB loss. While not ideal, it probably won't stop you from seeing
Jupiter.

For a dipole, a 1:1 balun is good enough. At around 1/4 wavelength above the
ground, a dipole has a feedpoint impedance of about 80 ohms. That gives you an
SWR of 1.6:1 and a loss of a whopping whole 0.2 dB. The insertion loss of an
impedance transformer is very likely going to be more than that. But, you would
still want to convert the balanced antenna terminals to unbalanced coax. Two
easy ways to do that: slip several ferrite beads over the coax at the antenna
feedpoint end, or wind about 5 turns of coax into a 4.5" coil at the feedpoint end.

Do you have an image of the "pumping" of the signal? What does the RF
spectrogram look like? Hard to say what that might be without seeing a picture
of it.
--
Dave

On 05/30/2015 08:00, spectrahm wrote:
> It sure did help, Dave, this answered all my questions. Thank you!
>
>
> Apparently it's just a matter of observing during the storms. The program you
> mentioned will be handy for that because they are short and only occur once or
> twice a day.
>
> Examining the program, it seems like I should be able to predict the storms by
> myself with the position of Io and Jupiter. Maybe I'll develop a Python program
> later to make this open and cross platform. If I do, I'll publish it on my
> GitHub account for those who are interested.
>
>
> Being still new to the HF world, dipoles are kind of a magic science. To me, the
> Jovian dipole is quarter-wavelength since only half of the dipole is connected
> to the probe, the other half being connected to the ground. When I also connect
> this second half to the probe I get a stronger signal, which sounds right to me
> but should not if I got the dipole physics right. I understand some resonance
> should occur between the two poles (ref

> <http://www.hottconsultants.com/pdf_files/dipoles-3.PDF>) and increase the gain,

> so I'm not sure why I got a better result when I transformed it into a monopole.
>
>
> Do you think using a monopole could be a way to improve my chances of success,

> if it /seems/ to have a better gain than a same length dipole?

[Nathan Towne]

Just a quick question. Does anyone use the usual 300-75 ohm
transformer-type baluns available for VHF and UHF television? Do they
work at 20 MHz and not too lossy? They otherwise seem like a nice match
to the problem.

Nathan

[Dave Typinski]

Match to what problem, Nathan?
--
Dave

[Nathan Towne]

Matching a dipole antenna to 75-ohm coax at 20 MHz.

[spectrahm]

Dave,

Thank you for your insights. When I connect the second half of the dipole to the probe, I see an elevation of the noise floor, but the HF signal from hams get also a little louder, which made me wonder if it would do better for Jupiter. The SNR is probably about the same for both configurations.

Following your suggestion, I added a ferrite bead to the very short piece of coax between the connecting block and the F->SMA adapter before the LNA, it also seemed to help with the SNR.

For some reason, my Airspy seems to perform worse than my R820T2 SDR dongle at the same bandwidth. The Airspy gets a kind of "comb-ey" signal on the waterfall which looks like some harmonics as I only get noise when I tune to them. The noise floor with the SDR dongle is nice and flat at the same frequencies while the signal is as clear as the Airspy where there is some.

So far, this is what I get in the ham band:

You'll notice the "pumping" effect on the waterfall. However I get some signals from far away countries considering the languages I hear, so it may just be caused by the atmosphere. The FM radio is crystal clear even without the LNA, so I think my dipole is performing well.

With the same gain settings, this is what I get at 20.1MHz so far:

Yesterday I had the noise floor "pumping" but at random and short bursts. The hardware AGC is obviously off so I thought I might have spotted some bursts, but they wouldn't translate into their typical AM sound, only the same regular noise. I'm still not sure about what it was.

There should be a jovian storm in about 30mn and Jupiter is in plain sight westward, facing my dipole. I'll keep monitoring the waterfall in the meantime.

Thanks again for your advises,
Damien

[Dave Tucker]

There is a Java jar file for predicting the storms here:  http://jupiter.kochi-ct.jp/jrm/  I use it on Mac and Linux.

[spectrahm]

That is helpful, Dave. The other program is an older VB6 program, closed source and Windows-only, so I thought of writing my own.

By the way, I've just realized I can pick up the hydrogen line with this low frequency dipole. I would never have thought it possible.
With simple_ra and a bit of integration, I can even get something meaningful from it:

[Dave Typinski]

Well, a dipole 1/4 lambda above the ground is already nearly 75 ohms, so the
only matching needed is from the balanced feedpoint to the unbalanced coax. For
that, a few ferrite beads or a few loops of coax will suffice. The CATV balun
includes a 4:1 impedance transformer -- which we certainly would not want or need.
--
Dave

[Dave Typinski]

Hi Damien,

I was afraid of that. The "pumping" (as good a word for it as any) I have seen
before myself. It was due to power line noise overloading the ADC inside the
SDR. The only for-sure fix is to kill the RFI. Could be any number of
electronic gizmos or AC mains components causing it. Trial and error of turning
things off one at a time is the best way to start. Alternatively, you can hope
that the overloading is comping from a portion of the spectrum way above or
below the upper HF band -- and put a stout bandpass filter, say, 15 to 30 MHz,
ahead of the receiver.

On the ferrite beads: one bead probably does not do very much. Three to six
beads next to each other might work better.

I do not understand how you are connecting the feed line to your dipole. What
is the "probe" you mention? The coax should be connected to the dipole
similarly (at least in the electrical sense) to what is described in the Radio
Jove Antenna Manual, available here:

http://radiojove.gsfc.nasa.gov/telescope/ant_manual.pdf

--
Dave

On 05/30/2015 16:01, spectrahm wrote:
> Dave,
>
> Thank you for your insights. When I connect the second half of the dipole to the
> probe, I see an elevation of the noise floor, but the HF signal from hams get
> also a little louder, which made me wonder if it would do better for Jupiter.
> The SNR is probably about the same for both configurations.
>
> Following your suggestion, I added a ferrite bead to the very short piece of
> coax between the connecting block and the F->SMA adapter before the LNA, it also
> seemed to help with the SNR.
>
> For some reason, my Airspy seems to perform worse than my R820T2 SDR dongle at
> the same bandwidth. The Airspy gets a kind of "comb-ey" signal on the waterfall
> which looks like some harmonics as I only get noise when I tune to them. The
> noise floor with the SDR dongle is nice and flat at the same frequencies while
> the signal is as clear as the Airspy where there is some.
>
> So far, this is what I get in the ham band:
>

> <https://lh3.googleusercontent.com/-X0PFlnKtfE4/VWoVhyEGWuI/AAAAAAAAAAM/cSe2kPLj1Ic/s1600/gqrx-hf.png>

>
>
> You'll notice the "pumping" effect on the waterfall. However I get some signals
> from far away countries considering the languages I hear, so it may just be
> caused by the atmosphere. The FM radio is crystal clear even without the LNA, so
> I think my dipole is performing well.
>
> With the same gain settings, this is what I get at 20.1MHz so far:
>
>

> <https://lh3.googleusercontent.com/-OLo6UjL1oDg/VWoV_MlLseI/AAAAAAAAAAU/C9KcPtbf75A/s1600/gqrx-20mhz.png>

> > To post to this group, send email to sara...@googlegroups.com <javascript:>

> > To unsubscribe from this group, send email to

> > sara-list-...@googlegroups.com <javascript:>

> > For more options, visit this group at
> > http://groups.google.com/group/sara-list?hl=en
> <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 <javascript:>
> > <mailto:sara-list+...@googlegroups.com <javascript:>>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.
>
>

[Nathan Towne]

Thanks, that is what I need to know. Nathan

[spectrahm]

Hi Dave,

My apologies for the poor wording, my English can be awkward (not to say broken) at times. By "probe" I meant the center core of the coaxial cable which protrudes to the connexion block.

Following your recommendations I made a new connector to the dipole, as the first one was literally as short as practical and could not accommodate more ferrite beads.

The "pumping" (for lack of a better word) was indeed caused by a saturation of the ADC. I couldn't lower the gain as it would make the signal too faint, but switching to the Airspy with its 12bit ADC solved the problem.

There was an Io-A storm occurring from 22:00 to 00:00 UTC. I noticed a lot of "pops" using the AM demodulation, while the noise floor kept  bursting. Unfortunately this doesn't look nor sound like short or long jovian burst but rather some short broadband RFIs. Here is a screenshot of the waterfall. I also recorded 12 seconds of it + spectral display if you are interested.

(all the pictures are clickable).

The dipole works great but I'm starting to think it just won't work for Jupiter in my suburban area... At some point I will need to purchase some land for my antennas. Surprisingly, the hydrogen line at 21cm is quite visible with this HF dipole, which is even more of a surprise considering I'm using a HF LNA with a 50MHz lowpass filter.

Thanks again for your advises,
Damien

[Dave Typinski]

Hi Damien,

Now I understand, thank you. Nice looking coaxial feed cable.

Yes, urban environments can be difficult. Please note, however, that Jupiter is
currently a little too close to the Sun, making observation of Jupiter difficult
-- more band noise to cover up Jupiter and a thicker ionosphere to attenuate
Jupiter. The next period of Jupiter being in good position relative to the Sun
will start this November.

I wish you luck finding a quiet observing location, Damien.
--
Dave

On 05/31/2015 19:01, spectrahm wrote:
> Hi Dave,
>
> My apologies for the poor wording, my English can be awkward (not to say broken)
> at times. By "probe" I meant the center core of the coaxial cable which
> protrudes to the connexion block.
>
> Following your recommendations I made a new connector to the dipole, as the
> first one was literally as short as practical and could not accommodate more
> ferrite beads.
>

> <http://s22.postimg.org/8vgunvadd/DSC_1221.jpg>

>
>
> The "pumping" (for lack of a better word) was indeed caused by a saturation of
> the ADC. I couldn't lower the gain as it would make the signal too faint, but
> switching to the Airspy with its 12bit ADC solved the problem.
>
> There was an Io-A storm occurring from 22:00 to 00:00 UTC. I noticed a lot of
> "pops" using the AM demodulation, while the noise floor kept bursting.
> Unfortunately this doesn't look nor sound like short or long jovian burst but
> rather some short broadband RFIs. Here is a screenshot of the waterfall. I also
> recorded 12 seconds of it + spectral display if you are interested.
>

> <http://s22.postimg.org/7qi74qyip/floor.jpg>

>
> (all the pictures are clickable).
>
> The dipole works great but I'm starting to think it just won't work for Jupiter
> in my suburban area... At some point I will need to purchase some land for my
> antennas. Surprisingly, the hydrogen line at 21cm is quite visible with this HF
> dipole, which is even more of a surprise considering I'm using a HF LNA with a
> 50MHz lowpass filter.
>

> <http://s22.postimg.org/efoqkrjup/hf_dipole_h1line.png>

> <mailto:sara-list+...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.