Meteor scatter: Initial results with RTL-SDR

[BiggarDigger 227]

Hi everyone,

over the seemingly endless summer nights I've been looking at projects to do that wouldn't be affected by the sun close to the northern horizon.  In the past, I used to operate meteor scatter on amateur radio, and spurred on by the recent article in The Sky at Night, I thought I would put together a small meteor scatter reception setup.

The system comprises a HB9CV (a very simple 2 element phased array) a few metres of RG58u and a RTL-SDR connected to the PC in the study.  I rigged the antenna up in place of an old, now defunct, Sky satellite dish and pointed it south.  It's mounted on the side of the house, approximately 4 metres above ground with a clear view to the horizon.  This is ideal because the cable run is fairly short and no additional holes are required in the wall.

Over the weekend, I played with a few software packages, looking for options, not just to view waterfall plots, but also to be able to log, analyse and tabulate the results.  I wasn't expecting much as GRAVES is radiating to the South, meaning it would need a significant backscatter to be able to reach Scotland and the HB9CV has only 3-4dB forward gain.  Nonetheless, I setup HDSDR and added the M4 screen capture and logging tool from Andy, M0CYP.  Tuned to GRAVES, I was hearing a number of pings during the day and had a couple of big pings too.  So I left it running  overnight and I thought I would share some initial results from last night.

The first two images are the same ping, wrapping around the waterfall plot. 

 

The second two are part of a 61 second burst starting at 10:51:48 UTC this morning.  Lots of other small pings and bursts too.

There's much more to do to optomise the setup, particularly on the software front.   I played with a couple of other packages, notably "Echoes" trying to setup to log the echoes in a usable format to analyse duration and intensity, but it failed ot detect any pings.  Switching back to HDSDR almost immediately detected backscatter.

One thing I haven't figured out yet, is the correct settings display the monitored frequency correctly on the logging tool.  The frequency scale at the bottom of the logged plots in the detection software is incorrect, while in the tool itself and HDSDR they appear correctly tuned, see attachment 5. 

 

Note HDSDR and the detection too are tuned to 143.050, but the logged pings are at 143.0525.  Perhaps, I need to read the help file some more...

Looking forward to next weekend and the Perseids now!

[Stub Mandrel 12,943]

Hi, I've managed to get an SDR working with HDSDR in the Midlands (and from there to Spectrum Lab).

Here are some meteor traces in HDSDR, note that the frequency 'sweeps' for most of them. You might find some fo my settings helpful too:

 

 

 

[The Admiral 1,892]

1 hour ago, BiggarDigger said:

Note HDSDR and the detection too are tuned to 143.050, but the logged pings are at 143.0525.  Perhaps, I need to read the help file some more

Hi Biggardigger, welcome.

Hmm, I'm not wholly convinced that they are meteor scatters, somehow they just don't look the ticket. One does occasionally see interference close to the Graves frequency. It's normal practice to offset tune the receiver, say to 143.048MHz, so that the scatter signals are audible, in this case at 2kHz. But if you are tuned to 143.050, you shouldn't be getting a constant signal at 143.0525MHz. Echoes as long as 61s would be pretty rare, I've never seen one that long, but in any event they wouldn't have that cyclic character.

Your set up should give you results, if they are possible at all, but you are a long way north. Most observers use a 2 or 3 element Yagi, or bigger in some cases, I don't know how that compares with your HB9CV.

Have you taken a look at this thread?

Ian

 

 

[BiggarDigger 227]

Thanks both, very helpful.

Yes, that thread was one that I reviewed here.  Though it is long and multi faceted, it gave some useful tips.  Coupled with my experience in the past, I'm fairly sure the short pings and burst are meteor reflections:  I'm too far away from Dijon for them to be aircraft or other means.  The long burst may have been meteor scatter, but it's not 100% clear yet.  They might have been meteor assisted propagation such as Sporadic-E.

To clarify, the receiver is (was) tuned to 143.048MHz using narrow band CW rather than USB.  The traces in the first 4 screen shots were taken from the logging tool that appears to interpret them at 143.0525MHz.  However, the logging software is (was) set to monitor and trigger on the pings as they came in on a centre frequency at 144.050 on CW.  I think  the problem is my settings in the logging software that I need ot look at more carefully.

In the past, I've seen quite a few occasions where a single meteor scatter burst can last several tens of seconds, particularly in the run up to and over the peak of the Perseids. Nonetheless, but yes I was very surprised, given the modesty of the setup.

Other than the long bursts, the pings have the characteristic appearance and sound of short burst meteor scatter that I've worked with before.  The long bursts appeared to have  modulation of the received signal strength which may have been related to the azimuth switching of the transmitter array in Dijon.

The HB9CV is a phased 2-element antenna, roughly equivalent to a 3 element yagi and is very popular for portable amateur radio use, yielding a forward gain of around 3.5 to 4dBd, broad frontal lobes and deep side notches.  Given my range to Dijon, I mounted it horizontally polarised to take advantage of the frontal lobe as I would not be looking for reflections from high elevations that could be delivered by mounting it vertically polarised.

 

I've just adjusted the RTL-SDR to use USB and tweaked the frequencies and bandwidths as required.  I noticed I had an incorrect set sampling rate, which may have affected the "look" of the reflections.  While typing this, two further pings came in that I captured manually:

The logging software correctly recorded these at 143.0497, as per HDSDR, yet I think there is something not right in the logging start and end frequencies, perhaps they are just labels that I need to adjust:

The strong heterodyne just out of band disappears when the PC monitor is switched off overnight.

Let's see what this captures overnight.

 

Richard

(sorry, ignore the incorrect screen capture below, it's from the earlier post and I can't seem to remove it now it's been attache to this post too!)

Edited August 6, 2018 by BiggarDigger
Remove incorrect screen capture

[Stub Mandrel 12,943]

Those two look more convincing.

[BiggarDigger 227]

Sorry for replying to my own message, but as I was trying to remove the erroneously posted screen capture earlier, this was logged:

It started at 20:19:09 UTC, was still going strong off the bottom of the water fall, and finally faded at 20:19:30UTC.

 

[The Admiral 1,892]

OK Richard, I understand. I'm not familiar with your logging software.

I agree with Neil, the short pings certainly look like meteor scatters. Your last post, it does look like a meteor scatter, but to be of so long a duration outside of a shower shower surprises me. We are in the run up to the Perseids though. Still, given that I generally only ran my meteor set-up during showers perhaps I shouldn't really comment. I'm not running at the moment so I am unable to correlate any event. Perhaps others can. What is your strike rate at the moment?

Ian

[BiggarDigger 227]

Unfortunately Ian, I didn't log anything overnight.  In my haste to get a system functioning on Sunday, I set HDSDR to schedule recordings on a test configuration.  What I didn't realise is that the schedule took precedence over the current settings and also forgot the schedule was there, so at midnight last night the SDR re-tuned itself off frequency and I have no data.   That schedule has now been deleted and a lesson learned, I think.  I'm seeing a couple of very short pings just now, but can't sit with it as time is pressing this morning.

The logging software is MD4-Beta7 from Andy M0CYP https://www.rtl-sdr.com/forum/viewtopic.php?f=3&t=1734 It identifies pings based on the waterfall pot rather than audio spectrum.  It seems to be functional, but I'm not convinced that I have it configured to correctly log the frequency of the pings.  I think tonight, if I get time, I may have a play with the Spectrum Lab tool  having found some guidance and scripts on the Britastro site.

Richard

Edited August 7, 2018 by BiggarDigger

[Stub Mandrel 12,943]

12 hours ago, BiggarDigger said:

: 





I think these two you posted look much more convincing.

If you use USB (or LSB!) you can hear the meteors very distinctly as little 'dooo' noises that usually drop in frequency.

 

[BiggarDigger 227]

The seem very much like reflections I've used in the past for radio communications.

Several more short pings and a couple of  longer burst in the order of 5 or 6 seconds during the day.

This evening, I'm starting to play with Spectrum Lab, but seem to be encountering the usual newbie setup woes described in several places.  I've looked through the documentation on the BritAstro site and loaded the starter configuration.  Much of the documentation I can find relates to setting it up with a Fun Cube or similar: I haven't yet seen anything specific to allow it to feed the audio in from the RTL-SDR.  I saw a couple of descriptions in the thread Ian mentioned above, but I'm not sure what version on Windows that relates to: they don't appear to be Windows 7, 8 or 10.  Is there a driver or specific configuration that allows audio from the RTL-SDR to be fed into Spectrum Lab?

I can turn down the gain on the left channel to prevent feedback, but other than that not much progress on spectrum Lab.  Meantime, HDSDR continues to ping away nicely.

Richard

[Stub Mandrel 12,943]

Check out this thread on my adventures:

As I don't have a funcube, I followed a tip found on the forum to use Virtual Audi Cable from this site: https://www.vb-audio.com/Cable/

You set the virtual cable as output in HDSDR and as input in SpectumLab.

 

[BiggarDigger 227]

That's Brilliant Neil, how on earth did you find that?

After much playing with settings and configurations, I have a functioning Spectrum Lab (sort of) connected to HDSDR and capturing meteors:

 

There's plenty to play with in Spectrum Lab and I think I should  be able to work through most of it: One thing I'll be looking into are the triggers to record both the .wav files and spectrum plots too.  I need to adjust some gains and scaling but it's mostly functional now.

One thing that bugs me though: earlier while playing, I fed the audio out from Spectrum Lab to the system speakers, but now can't find the setting for that.  I select the output device/pipe/driver for my system speakers and apply, but nothing happens. It may have been connected before I added a virtual cable.  So do I need a second virtual cable to connect the output of Spectrum Lab to the system speakers?  It would be good to be able to hear the pings on demand.

Richard

[Stub Mandrel 12,943]

3 minutes ago, BiggarDigger said:

I fed the audio out from Spectrum Lab to the system speakers, but now can't find the setting for that.

Same here... easy to set the spectrumlab output to the sound card, but not easy to figure out how to get SL to pass on the sound...

[BiggarDigger 227]

I found how to enable the audio output, and it's obvious when you see the fix.......

Turn on the Bypass on the filter between L3 and L4:

 

I now have meteor sounds too!  

 

Richard

 

[Stub Mandrel 12,943]

Whee!

I'll try that

[Tomatobro 515]

Well done!

I use SL for my display and SDRSharp for my Dongle setup. I have built an interface to give me real time graphing of the meteor count.

I also make videos of the screen using Bandicam. Bandicam gives me a delayed start and continuous half hour recordings for later analysis.

At full screen resolution 12 hours will generate 15G of data. Use Films and TV to playback as this gives a smooth high speed scroll through.

That long trace in your earlier results would look something like this....and is space junk according to Meteor.org

 

sample.txt

[Tomatobro 515]

I find with SL that there are often duplicated commands. I think that sound output can also be turned on from the start/stop menu. Its the bottom command on the list as I recall.

[Stub Mandrel 12,943]

Not duplicated, you have to do three things to turn on sound the first time:

1. Select sound card for output
2. Remove filter bypass via circuit window
3. Select send audio option (can't remember wording, it's confusing)

But once that's done you can toggle sound on/off by undoing one of those.

A clear example of software that keep getting stuffed with more and more features without consideration for keeping basic functionality simple. It's an enthusiast's program. What it really needs is the ability to set up simplified front ends for different purposes that hide all the confusing bits and supplies obvious things like a 'sound on/off' button.

Don't you love intuitive software

[Tomatobro 515]

Its probably true to say that the only person the truly understands SL is the person who wrote it.....

[BiggarDigger 227]

That's great info on what's possible with Spectrum Lab.  The documentation does seem to be targeted at the audio enthusiast, which is fine, because that's who the target audience is I guess.

Next up will be to record and automatically log/tabulate recorded echos.

Interesting to see that meteor.org thought the long bursts were space junk.  Space junk should be more massive than most meteors and leave a more intense plasma trail for longer.  However, we are heading towards the Perseids and the classic fireballs that shower provides can leave long trail bursts as I've seen and used in amateur radio before.

What will be interesting to see is the effect on the long bursts after the shower, noting however that there are several other lower intensity showers occurring just now too - summer in the northern hemisphere is a good time for meteors. If the frequency of the long bursts out of shower activity does not reduce significantly then perhaps they are, at least partially, space junk (assuming a random distribution of junk).

The distribution of long bursts through the day will be very interesting too.  As far as I'm aware, most meteor showers peak overnight: space junk should be random(ish).  My experience on 144Mhz radio tells me that there can be long bursts out of shower activity and during the daytime, but that they are (anecdotally) less frequent.

The lack of initial Doppler shift might be an indicator that they are junk, but that can occur on meteors too, depending on the geometry of the strike.

Whooo...science!

[Stub Mandrel 12,943]

1 hour ago, BiggarDigger said:

Interesting to see that meteor.org thought the long bursts were space junk.  Space junk should be more massive than most meteors and leave a more intense plasma trail for longer. 

The space junk is probably orbiting, not burning up. The point of the Graves radar isn't to help us amateur meteor detectors, it's for France to track satellites.

[BiggarDigger 227]

Aha, yes, I forgot that Graves is not there just to please us!

Orbiting junk or live satellites/ISS would show Doppler shift I think?  Stuff vaporizing may only show an initial shift due the rapid deceleration.

Very interested to see any changes in the distribution of these long bursts over the coming weeks.

Richard

Edited August 8, 2018 by BiggarDigger

[Stub Mandrel 12,943]

My view, and I found some work to back this up, is that the 'perfect' meteor trace combines a gradually sloping line that's the doppler shifted signal from the moving ball of ionisation at the meteor's 'head' (which we usually see). This can run from above to below the centre frequency if the meteor is moving side to side across the beam. On top of this you get a static trace that represents the relatively static trail left by the meteor (which is only rarely seen).

[BiggarDigger 227]

Yes, thats my understanding too.  From previous work, the audible response sounds very approximately like "peeiiiooooooooooooooooooo". With the initial strike exhibiting a sharp attack, then being Doppler shifted and finally a tail of fairly constant frequency, breaking up towards the end as the ionization dies away.

I've been recording the pings and bursts from HDSDR, so I'll see if I can dig some out.  It's constantly recording, splitting into 10MB wav files.  I may be able to extract and save sections to post later.

[The Admiral 1,892]

1 hour ago, BiggarDigger said:

Orbiting junk or live satellites/ISS would show Doppler shift I think?

Indeed it does, as it's often that one picks up the backscatter from the ISS. The doppler shift goes from +6kHz to -6kHz for that. Mind you, that is a 'large' piece of orbiting hardware, whether we amateurs would be able to detect space junk I don't know. In theory, perhaps if it's large enough I suppose.

Ian