Jump to content

Narrowband

SID detection with SpectrumLab


Recommended Posts

Hi,

A while back I bought a copy of 'The radio sky and how to observe it' and finally got round to building the aerial for detecting SIDs.

I have had it running this afternoon and have been logging the results.

58f4fc6383054_VLFaerial1.thumb.jpg.e95837baf4314556ae018b444827948b.jpg

I think it is all working but one thing I have noticed is that the signal to noise ratio (vertical axis) is pretty poor.  

Does anyone have an idea about the point at which the ratio becomes too low to yield meaningful results? 

Thanks,

Dave

Link to comment
Share on other sites

  • 2 weeks later...
  • Replies 38
  • Created
  • Last Reply

I think mains hum is causing some trouble...

Frequency.thumb.jpg.6e0e2b605c1cdb9fce7da0e9fade9505.jpg

have left it running since 17:00 last night logging every second, so hopefully I will have a good stack of data to analyse.

Link to comment
Share on other sites

On 4/17/2017 at 18:39, HN50 said:

I think it is all working but one thing I have noticed is that the signal to noise ratio (vertical axis) is pretty poor.  

Does anyone have an idea about the point at which the ratio becomes too low to yield meaningful results?

Hi Dave,

What is your antenna setup as i'm presuming you are using a soundcard?

-115 to -120dB is too low I think, without an amplifier you should be able to get -80 to -100dB with a good sized loop antenna. The issue is the variance in signals is not large, so you really want to pull the signal out of the noise floor to start off with, and other than 19k8, there rest are very close to the point of not being able to tell for sure if they are there or not from the graph. 

Can you take your logged data to excel and plot and share on here, or put the data here. It should be obvious if you have enough signal or not as you will either clearly see the expected pattern or not.

Also looking at the screenprint there is not much above 21k - were you using 44k or 48k sampling. My experience is a lot of standard and even "HD" sound cards max out around 20-21k or get noisy at that point.

Mains hum should not be an issue if you can build the right antenna and locate/direct it right, if it is you can try put a preamp on your antenna (assuming it is outside away from mains lines/wires) but this may be too much to start off with.

Link to comment
Share on other sites

Hi Ian,

Sorry for the delay, have been a bit pre-occupied with work.

Basically the methodology entails making a wooden cross and winding a diamond of enamelled copper wire round it, 125 turns of 24swg copper wire.  

IMG_873.thumb.jpg.54e538cfcc6ff2967a7da91c26186280.jpg

This then gets connected to a 3.5mm headphone jack that gets plugged into a laptop running Spectrum Lab, which should anable you to detect various VLF transmitters and log fluctuations in their signal strength.

However I think at the moment the antenna is not working as I found there is no difference to the signal when I have it plugged in or unplugged to the extension cable.  I think I might need to use the continuity function on my multimeter as I think something has gone wrong.

If you don't want to buy the book there was an article about this in Sky at Night magazine back in September 2015.  I would like to say that there is a copy for sale on eBay but I bought it myself :)

However... if you google 'RAG_SIDv1.USR' one of the links is to a PDF of that edition with the details.  

Link to comment
Share on other sites

On 03/05/2017 at 13:54, 7170 said:

Hi Dave,

What is your antenna setup as i'm presuming you are using a soundcard?

-115 to -120dB is too low I think, without an amplifier you should be able to get -80 to -100dB with a good sized loop antenna. The issue is the variance in signals is not large, so you really want to pull the signal out of the noise floor to start off with, and other than 19k8, there rest are very close to the point of not being able to tell for sure if they are there or not from the graph. 

Can you take your logged data to excel and plot and share on here, or put the data here. It should be obvious if you have enough signal or not as you will either clearly see the expected pattern or not.

Also looking at the screenprint there is not much above 21k - were you using 44k or 48k sampling. My experience is a lot of standard and even "HD" sound cards max out around 20-21k or get noisy at that point.

Mains hum should not be an issue if you can build the right antenna and locate/direct it right, if it is you can try put a preamp on your antenna (assuming it is outside away from mains lines/wires) but this may be too much to start off with.

Hi 7170,

Thanks for the post, I have been doing some testing before replying. 

Yes, I am using a loop plugged directly into the sound card on a laptop.  Since my prior post I have;

- changed the socket and retested

- tried several different cables and retested

Just to try and eliminate them as a possible source of problems.  However the signal is still what it was, but at least I am happy I can  discount them.

I have ordered one of those external sound cards that S@N article mentions, as for £6 I figured it was worth a go.  The sound card in my laptop is set to 48kHz, but for some reason it was logging a station at about 26kHz which is not what I was expecting.

Though I think ultimately the issue is with the antenna :) 

I will post that first set of data later.  The logger is currently running and I am going to leave it going over the weekend.  

One thing I do want to try is taking it out to a spot in the South Downs where I go with my VLF whistler radio as it is that bit further from the mains.  There are telephone lines outside my house which impart 'a bit' of a hum when I listen.  I will modify the base so it will attach to my Horizon tripod without a g-clamp.

Thinking about other avenues;

I could try building a larger diameter antenna.  The S@N article has theirs at 1m across, and I have seen others at 1.5.  That makes placing it a bit more tricky, it would need to go in my roof.

Or I guess I need to think about a simple-pre amp...  I had wondered about modifying my VLF radio and using that, but I have to admit I am more of a tinkerer with electronics. Do you have any ideas about these?

Thanks,

Dave

Link to comment
Share on other sites

What a difference spending £6 can make.

I have had it pointed out to me that in one of the earlier posts I talked about mains hum and included a screenshot showing a significant drop off in all signals at about 7kHz.  Looking back through screenshots for my whistler radio I had seen this there but not really appreciated it before; I just wrote it off as how the radio worked.

However just to try I plugged this little fellow into the laptop.

IMG_884.thumb.jpg.91dc9ae434abc9e233d8137d93731459.jpg

I the laptop detected it automatically and I set it as the default microphone input.  I then set it as the input for Spectrum Lab and the plot changed to something like this;

IMG_886.jpg.7258607e605f2454e161f79b9cebe877.jpg

Aaargh! The noise!  I think though there is something wrong with the built in audio input on the laptop as it was dropping off above 7kHz.  But when I looked at the plot something had happened at 10:30 when I made the switch;

590db5ba3d1ae_VLFaerialnewaudioadapter1.thumb.jpg.460d880c15994b2e7f4f1972a9fbac09.jpg

19.8kHz had jumped to about -50dB.  Unfortunately the noise level is now stronger than several of the other transmitters, but I was at least getting a better signal.  I also noticed I was not displaying all the transmitters, so tinkering with the plot further I got this;

590db67c12f92_VLFaerialnewaudioadapter2.thumb.jpg.5910989c0e1e82194081afcfd0c3c92e.jpg

I am guessing that some of the transmitters are now weaker than the background noise and won't show much, but it does leave me with 4 that are a bit above.  So I won't post my previous data, I will leave it running for a few days while I think about a permanent location for it to run from.

Not a bad morning. :)

Link to comment
Share on other sites

8 hours ago, HN50 said:

Not a bad morning. :)

Indeed - it is looking good. The signal strength is much better now and I would say there is no need for a preamp.

Seeing 4 transmitters at reasonable strength is pretty going going for a small-ish single loop. If you want more you can find yourself needing to build a setup that has two loops at 90 degree angles. 

If mains hum is still an issue you have two options, first is to use an active USB extender (you can get them cheap on eBay), and put the USB soundcard as close to the antenna as possible in some sort of waterproof box, or build a passive high pass RC filter (given the frequency you can make one and test using a free PC sound card based oscilloscope etc). Mains hum can also become an "issue" if you have gain turned up to much and you may see this as strong harmonics across the whole range due to the soundcard, when in fact there is nothing at the higher harmonics, this is also common with RTL Dongles being used as SDRs and strong local FM stations.

For location, atics seem to be the next most popular place after the garden and they do have one benefit (in addition to being dry), still air, so you don't need to worry about your antenna moving slightly in the rain or wind which will add noise to the signal. I would not advise a size under 1m x 1m myself.

If you are wanting something outside that will stand up to the elements and is a little more aesthetically pleasing than a wood frame, you can get a large kids hula-hoop and cut a slit all around the outside parameter using a dremel or table saw and wind your wire through this slit so it sits inside the hula-hoop, then seal it up with self-amalgamating tape and clamp it, top and bottom to a wood post fixed in the ground. Some people bury them in the ground but i've never tried this.

In terms of the S@N article - I have not read it but there is one thing I disagree strongly with, having automatic gain control on. The idea is to monitor the strength of a signal which is being impacted by solar activity affecting the ionosphere etc, all other things being consistent. If you have AGC on, then your sound card will be trying to make the signals more consistent in amplitude, which is the opposite of what you want to be seeing. Personally I would have AGC off, and if possible set gain manually. 

Please do share your raw data, it will be very interesting to see your results.

Glad to hear you like tinkering - as this is an area where you can certainly tinker for a long time!

Link to comment
Share on other sites

On 06/05/2017 at 22:21, 7170 said:

 

Please do share your raw data, it will be very interesting to see your results.

 

 
Hi 7170,
 
Please find attached my raw data. 
 
 
I left on AGC for consistency across the dataset, but in the future will leave it off.  I have had a quick look in excel but not had the time to do more.  As I have recorded by the second interval I am interested to see how the results change when I simulate logging at say 10 or 20 seconds. 
 
I would be interested to see what you think, I think maybe there are only really two signals strong enough to yield useful data?  Based on this data do you think an antenna in the 1m to 1.5m size might be an improvement?
 
The next thing I need to do is get it set up and running up in my loft, all I need is a little hole in the ceiling and I can feed a cable up.
 
I had fitted a 10kHz high pass filter, however I think I have got the values wrong as that screenshot from Saturday is with it in, so I don't think I have it right.  My values are 150 ohms and 100nF .  Would I only need it on one lead on the antenna?  I think it is doing something though as when I connect the earth to a radiator pipe the signal gets slightly stronger.
 
With regards having two loops at 90 degrees, electrically would it just need be a case of soldering them in parallel or would I also have to add in diodes?
 
Sorry, lots of questions!
 
Thanks again,
 
Dave
Link to comment
Share on other sites

16 hours ago, HN50 said:
 
Hi 7170,
 
Please find attached my raw data. 
 
 
I left on AGC for consistency across the dataset, but in the future will leave it off.  I have had a quick look in excel but not had the time to do more.  As I have recorded by the second interval I am interested to see how the results change when I simulate logging at say 10 or 20 seconds. 
 
I would be interested to see what you think, I think maybe there are only really two signals strong enough to yield useful data?  Based on this data do you think an antenna in the 1m to 1.5m size might be an improvement?
 
The next thing I need to do is get it set up and running up in my loft, all I need is a little hole in the ceiling and I can feed a cable up.
 
I had fitted a 10kHz high pass filter, however I think I have got the values wrong as that screenshot from Saturday is with it in, so I don't think I have it right.  My values are 150 ohms and 100nF .  Would I only need it on one lead on the antenna?  I think it is doing something though as when I connect the earth to a radiator pipe the signal gets slightly stronger.
 
With regards having two loops at 90 degrees, electrically would it just need be a case of soldering them in parallel or would I also have to add in diodes?
 
Sorry, lots of questions!
 
Thanks again,
 
Dave

Hi Dave,

I've done some analysis of your numbers and you have two strong signals as you say (18k3 and 19k8) but also have a possible signal on 20k9 as there is a difference between day and night (blue squares in my graphs), but a longer data span would be needed to confirm.

To generate the larger graphs I just disseminated the data to only keep readings every 30s (with the unused readings just binned). To be honest I only did this as Excel isn't great with a large number of points on a graph. In the past I have either used 30s or 60s logging period. I also added a moving average trend line to try and smooth out the noise to see the real trend.

Before changing the loop size, I would run it in the loft for a week and see how it looks, and perhaps play with the loop direction to see what works best before securing more permanently. Without comparing to official data and getting an understanding of the trend over time, it is hard to know if the peeks are SID related or noise/AGC etc from the graphs. A good example of a clean graph to aspire to is:

dbenham.1j.jpg).

In spectrumlab what frequency range are you monitoring the amplitude, for each "frequency". I have found +-100hz from the centre point is a good starting point and then you can narrow down depending on your soundcard accuracy etc.

With regard to the 90 degrees idea, the ones I have seen have been wired in series, but I've not yet tried it properly other than just "dabbling" with a flimsy test rig. It is something I hope to try this year again so will post on here when I finally get a chance. A loop antenna is picking up the magnetic component so I don't think you should take either end to earth. It maybe the case that your high pass filter is tuning the circuit? What happens if you don't have it?

James

 

all signals.jpg

18_3.jpg

19_8.jpg

20_9.jpg

Link to comment
Share on other sites

Attached are two examples of some data I have captured over the last few days with my portable rig as this thread inspired.

This was captured using a 60cm x 60cm loop antenna, with 20 turns of (c0.75mm I think) enameled wire. It was at ground level in the shed, pointing approx E-W.

It clearly shows the expected sunrise, sunset, daytime and nighttime pattern (as exampled by my post above), though it is a little noisey, but with some careful analysis and comparisons to official records you could probably identify some SIDS.

A larger antenna should get a cleaner signal. Making SID identification possible.

I logged data for the following stations (ones with a * yielded data that was usable )

RSDN-20 Alpha 

GBZ (Anthorn UK) *

GQD (Skelton UK) *

DHO38 (Rhauderf) *2

NAA (Cutler USA) *1

NPM (Pearl Harb)

 

*1 - this was the most usable and provided the cleanest plot.

*2 - some data but i'm not sure it was transmitting the whole time

The NAA - Cutler USA appears to show a SID (red box) at 20.08 on the 10th, but online sources do not seem to show anything, so the fun of trying to work out what it was begins (it is possibly the strip light in the shed being turned on, but then it is not on the GQD plot....).

 

May_2017.png

May_2017_2_NAA_C.png

Link to comment
Share on other sites

Hi James,

sorry, been a bit busy with work but your last two posts are really interesting, looking to reply properly later as I have a follow-up question or two.

Cheers,

David

Link to comment
Share on other sites

On 03/05/2017 at 20:23, HN50 said:

 'RAG_SIDv1.USR' one of the links is to a PDF of that edition with the details.  

Hi, very interesting topic :thumbsup:

I have found the article, but the link in it to the .USR is broken, do you know of an alternative location for it ? (my Googling only turned up two hits, the other was to digitalrust website with the same broken link :( )

thanks.

Link to comment
Share on other sites

On ‎10‎/‎05‎/‎2017 at 14:21, 7170 said:

Hi Dave,

I've done some analysis of your numbers and you have two strong signals as you say (18k3 and 19k8) but also have a possible signal on 20k9 as there is a difference between day and night (blue squares in my graphs), but a longer data span would be needed to confirm.

To generate the larger graphs I just disseminated the data to only keep readings every 30s (with the unused readings just binned). To be honest I only did this as Excel isn't great with a large number of points on a graph. In the past I have either used 30s or 60s logging period. I also added a moving average trend line to try and smooth out the noise to see the real trend.

Before changing the loop size, I would run it in the loft for a week and see how it looks, and perhaps play with the loop direction to see what works best before securing more permanently. Without comparing to official data and getting an understanding of the trend over time, it is hard to know if the peeks are SID related or noise/AGC etc from the graphs. A good example of a clean graph to aspire to is:

 

In spectrumlab what frequency range are you monitoring the amplitude, for each "frequency". I have found +-100hz from the centre point is a good starting point and then you can narrow down depending on your soundcard accuracy etc.

With regard to the 90 degrees idea, the ones I have seen have been wired in series, but I've not yet tried it properly other than just "dabbling" with a flimsy test rig. It is something I hope to try this year again so will post on here when I finally get a chance. A loop antenna is picking up the magnetic component so I don't think you should take either end to earth. It maybe the case that your high pass filter is tuning the circuit? What happens if you don't have it?

James

 

 

 

 

Hi James,

Thanks for the analysis, I have been looking myself and it is quite noisy.  I have been busy and got the antenna and pc set up and it is now running in the roof, logging every 30 seconds, AGC off.  I have put it on a base with several different directions marked on it; I will leave it a running a few days facing 180 degrees and then will re-orientate it as you suggested.

With regards the earth connection I think I should have added that that is out the bottom of the high pass filter I designed;

59186ee006325_Highpassfilter10kHz.png.21621072089ee561c2ed271205f6af9b.png

I find that without the earth the signal is weaker than with it connected, though I can't see an obvious cropping of signal below 10kHz;

591873c0a4152_Withearth.thumb.jpg.29f4ac3b63bdbc6c577375cce7cedef9.jpg

The darker strip ending 10:38:15 is where I had bypassed the filter momentarily just to show the difference. 

I have also gone back over the watch list in the logging screen and set everything to 100hz either side of the target frequency, it had been set to 50Hz.

So now it is just a case of waiting to see what gets written to that log file...

Dave

 

 

Link to comment
Share on other sites

On ‎12‎/‎05‎/‎2017 at 21:57, 7170 said:

Attached are two examples of some data I have captured over the last few days with my portable rig as this thread inspired.

This was captured using a 60cm x 60cm loop antenna, with 20 turns of (c0.75mm I think) enameled wire. It was at ground level in the shed, pointing approx E-W.

It clearly shows the expected sunrise, sunset, daytime and nighttime pattern (as exampled by my post above), though it is a little noisey, but with some careful analysis and comparisons to official records you could probably identify some SIDS.

A larger antenna should get a cleaner signal. Making SID identification possible.

I logged data for the following stations (ones with a * yielded data that was usable )

 

 

 

Hi James,

Your data looks interesting, I can see the daily variation that was only really hinted at in my first data run.  Do you think though that ultimately a larger loop will just yield better results?

Thanks,

David

Link to comment
Share on other sites

20 hours ago, SilverAstro said:

Hi, very interesting topic :thumbsup:

I have found the article, but the link in it to the .USR is broken, do you know of an alternative location for it ? (my Googling only turned up two hits, the other was to digitalrust website with the same broken link :( )

thanks.

Hi,

Unfortunately I have not been able to find it either, and the version of Spectrum Lab now downloadable does not contain a .usr file of the same name.  However it does have one called VLF_Stations_Plotter that comes with the Spectrum Lab download and amended the frequencies a little in the watch list.

David

5918807d661d8_Watchlist.thumb.jpg.1b105ab25905c8399f0f85601831e796.jpg

Link to comment
Share on other sites

2 hours ago, HN50 said:

Unfortunately I have not been able to find it either,

it does have one called VLF_Stations_Plotter that comes with the Spectrum Lab download

Thanks for checking and the pointer.

I have a way to go yet, I thought I had a reel of bell-wire here with which to construct a quick loop !, but lost now :( so will need to order anew, meanwhile I am reading the manual and have a steep slope to climb. Considering we are in to the downside of the solar cycle we need all the aids to alerts that we can muster I think.

I'll be watching your exploits with interest, , , :)

 

Link to comment
Share on other sites

17 hours ago, SilverAstro said:

Thanks for checking and the pointer.

I'll be watching your exploits with interest, , , :)

 

Hi,

No trouble, hope it is of some help.  I have not sat down and gone through the entire spectrum lab manual myself, I have kind of just dipped in and out where I think I need something, so it is very likely I have missed something.  If you spot something, feel free to point it out. :)

This will be a quite a slow burn I think, I am planning to leave the antenna running about a week before I reorient it, and there are about 4 orientations I want to try initially.  

Dave

Link to comment
Share on other sites

On 5/14/2017 at 16:24, HN50 said:

I can see the daily variation that was only really hinted at in my first data run.  Do you think though that ultimately a larger loop will just yield better results?

Hi David, how large is your loop as it is hard to tell from the photo? The portable rig i'm currently using is just 60x60cm so small, I am using fairly thick enameled wire which will make up for it a bit though.

My second thought was what frequency range you are looking at and I can see you are doing +-100hz around each center too which is a good starting point.

I wonder if your issue is noise/broad interference. Compare your spectrum capture to mine (attached - sorry about the lossy JPG!). In mine other than a few obvious signals there is nothing much there above 1kHz, whereas yours has straight lines all over it and you can't pick out many signals.

capture1.thumb.JPG.00ee2b7dfd1e4881951a220b107ea292.JPG

Do you have a computer monitor/laptop monitor near it, as I get a lot of interference like this when the laptop screen is on (see second screenshot on the right centered around 9.5 kHz), and it has harmonics all the way up the frequency range.

capture2.thumb.JPG.b988df00bf0dabd96455f97f636a2308.JPG

This is despite the USB sound card stick being on a 10m active USB extension cable, located about 3m away from the laptop in the shed. It could also be poor shielding of the USB audio stick and the mic preamp being overloaded etc.

Agreed there is no obvious clipping of the signal below 10kHz, what happens if you have the resistor across the loop rather than across one end and ground? Most PC sound cards have a positive DC bias voltage which allows you to power electret microphones (often turned on when you slide up mic boost or select mic boost or similar). The impedance will be low in the loop so I wonder if current is possibly flowing to earth via your filter and ground wire, and creating noise that way in the sound card mic input stage? What does a 24h plot look like without the filter? It would be interesting to compare to rule this possibility out.

Attached is also a summary of NAA and GQD for the last few days - nothing major to report, expect that NAA was very low between 12 and 2pm and high between 2pm and 4pm on 15th, though GQD was normal.

capture_15_may_2017.thumb.png.844e727120067fefc51c882b5c467d66.png

Again more things to investigate and think about vs official records!

 

Link to comment
Share on other sites

  • 4 weeks later...
On ‎15‎/‎05‎/‎2017 at 22:40, 7170 said:

Hi David, how large is your loop as it is hard to tell from the photo? The portable rig i'm currently using is just 60x60cm so small, I am using fairly thick enameled wire which will make up for it a bit though.

My second thought was what frequency range you are looking at and I can see you are doing +-100hz around each center too which is a good starting point.

I wonder if your issue is noise/broad interference. Compare your spectrum capture to mine (attached - sorry about the lossy JPG!). In mine other than a few obvious signals there is nothing much there above 1kHz, whereas yours has straight lines all over it and you can't pick out many signals.

 

Do you have a computer monitor/laptop monitor near it, as I get a lot of interference like this when the laptop screen is on (see second screenshot on the right centered around 9.5 kHz), and it has harmonics all the way up the frequency range.

 

This is despite the USB sound card stick being on a 10m active USB extension cable, located about 3m away from the laptop in the shed. It could also be poor shielding of the USB audio stick and the mic preamp being overloaded etc.

Agreed there is no obvious clipping of the signal below 10kHz, what happens if you have the resistor across the loop rather than across one end and ground? Most PC sound cards have a positive DC bias voltage which allows you to power electret microphones (often turned on when you slide up mic boost or select mic boost or similar). The impedance will be low in the loop so I wonder if current is possibly flowing to earth via your filter and ground wire, and creating noise that way in the sound card mic input stage? What does a 24h plot look like without the filter? It would be interesting to compare to rule this possibility out.

Attached is also a summary of NAA and GQD for the last few days - nothing major to report, expect that NAA was very low between 12 and 2pm and high between 2pm and 4pm on 15th, though GQD was normal.

 

Again more things to investigate and think about vs official records!

 

Hi James,

A bit of radio silence for the last few weeks on my part due to work.

My loop is 50cm diameter, which looking online seems to make it a bit of a tiddler(!).  I am guessing that makes it less sensitive, however I think I will defer making something larger until I have completed further directional tests.  That also gives me a bit of time to work out the largest antenna I can realistically fit up in my roof.

I have my laptop nearby the loop, however I do have quite an issue with noise; there are phone lines right outside my house which when listening with my VLF radio do impart a very loud hum so when standing beneath them sferics become difficult to hear.  I have left the filter as-is for the moment while I change the antenna orientation, so by the end of next week I will have a set of data every 45 degrees.  Then I plan to re-run the direction tests again without the filter to gauge changes.  My laptop and the power cable to it are within a metre and a half of the antenna, so I think I will need to look at moving it as well when I have some more time.

I am interested in your plots though as your night time signals are significantly stronger than the daytime - on my plots the night is characterised by low signal instead.  I take it that I am reading the graphs right in that the signal strength for any given transmitter increases at night then? 

My logging is also running every 30 seconds which I think does give less erratic readings when compared to a measurement every second.  I read the SpectrumLab manual and realised that it is the average signal over the interval time, so I might drop it further to every minute to try and help smooth some of the noise.

Thanks,

David

Link to comment
Share on other sites

Over the last 3 weeks as well as work I have also been undertaking a little bit of development work.  I have put together a C# script that reformats the data so that all the signals are in a single column, and built an additional package that loads it into a database for longer term storage and analysis.

Whilst this has been going on I have been running logs for about a week at a time with the following orientations;

MinLogDateTime       MaxLogDateTime       DaysRun  Notes
2017-06-05 06:02:00 2017-06-07 19:56:00 2.5         00-180 orientation
2017-05-27 04:34:00 2017-06-05 05:56:00 9.0         45-225 orientation
2017-05-19 19:44:00 2017-05-27 04:33:00 7.4         70-250 orientation
2017-05-14 11:26:00 2017-05-19 19:35:00 5.3         90-270 orientation


The 00-180 orientation has begun running and at the weekend I will get the rest of that data loaded too, and after that it will be 135-315. 

The first analysis I have done is to average by time of day for each transmitter and for each run (e.g. RDL will have an average 24 hour plot for each of three orientations I have run it for so far).  My idea is that this would hopefully show the difference in signal strength a change in antenna direction would show.   

Looking at my plots I think the only two transmitters I can see a good signal for are the Russian RDL at 18.1kHz and the French HWU at 18.3kHz.  I can see some of the others show a slight diurnal signature but realistically I think their signatures are too weak and noisy with my current setup.  These ones are below.

(Just spotted typo.  45-135 should read 45-225.  Ugh.  I wont regenerate them tonight as LibreOffice takes forever).

59386260de15c_DHO38signalsbytime.thumb.jpg.010e9e950c992b0f08e18e14faee1257.jpg

593862633ab5f_FTAsignalsbytime.thumb.jpg.8f2f451760d5c096fd7cdd4caef0da8f.jpg

593862658fc3f_GBZsignalsbytime.thumb.jpg.e95e3d1a893b4a2cdac758e3ad1cc628.jpg

5938626a2b079_HWU22k6signalsbytime.thumb.jpg.3950bf38eddd6b8531fafbf3938e818a.jpg

5938626c7edf6_NWCsignalsbytime.thumb.jpg.24904eab3211e4d674ef5e7eb783af78.jpg

593862711f6e0_VTX2signalsbytime.thumb.jpg.3043caabaf676f9169eea9b3a0524497.jpg

Interesting but I don't think I can glean much from them.

 

Russian RDL at 18.1kHz

593864b7706a3_RDLsignalsbytime.thumb.jpg.02cdd28a379174f4b009b02da672ebca.jpg

Looking at this one I think it is telling me that the best orientation is 70-250.  I appreciate smooth is a relative term with the data I am generating (!!!), but I can't help but feel 00-180 and 90-270 seem to fluctuate far more.  90-270 also lags somewhat in terms of the average time the signal starts to increase in the morning - I was wondering if this is because as the antenna is not in an ideal orientation it requires a stronger signal to generate a current in the first place for SpectrumLab (happy to be corrected).  I gather RDL is near Moscow, so I am going to stick my neck out and suggest  135-315 would be weaker than 70-250, on account that the bearing off to Moscow is closer to 70 digress from where I am. 

HWU 18.3kHz. 

593866674009c_HWU18k3signalsbytime.thumb.jpg.0161a2d924307c0c7ae55e8de336146b.jpg

Similarly for Rosnay I think something similar is happening, but can't quite decide which represents the better signal.  Looking on a map though Rosnay is off on a bearing of 135 from where I live, so I would expect that the strongest signal will be 135-315.  I will get a handle of this on the weekend of 17th/18th June when the data is logged. 

 

I have also plotted the data for three days for both HWU and RDL, one line has the data as is, another with a 5 minute moving average, another with a 10 minute moving average.

59386816b181b_HWU18k3signalsbytime-movingaverages.thumb.jpg.4ea94a1a10b8c277bb6f187f6aca09aa.jpg5938681b6d1c1_RDLsignalsbytime-movingaverages.thumb.jpg.acfa91d79eee0d14ac0407aa12149991.jpg

The moving averages help smooth out the variations, but never the less they are still very noisy.  Whilst I can see a common diurnal change I think the noise fluctuations will still mask any SID activity. 

 

So to sum up; 

1) I can see some influence of antenna orientation on my results

2) Some of the transmitters are too weak with my current set up to give meaningful results

3) My longer term project is noise reduction, which may involve me not having the antenna at my house... :)

 

But anyway, the data is gradually accumulating so I will see what I get.

Hope you enjoyed the update.

Dave

 

 

 

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.