Troubles Receiving NOAA APT with Gqrx

[Nick Gregory]

Hi,

Im really new to SDR technology, so please forgive me if I make obvious mistakes in terminology, settings, etc. I have been trying unsuccessfully over the past few days to receive NOAA APT satellite transmissions with (RTL)SDR and Gqrx as the software and want some advice. I tried to follow the guide here but to no avail. I can hear the satellites somewhat distinctly, but the images WxtoImg produces are unrecognizable. I can post the raw audio as well as the images produced if anyone would like to hear/see them. I suspect there is one setting not mentioned in the guide that needs to be set for this to work, but I'm not positive. I am using a homebrew QFH antenna (newly remade) so I'm almost positive that the antenna is not the issue. The RTL2832 based dongle I'm using can be found here. I'm running ubuntu 12.04 desktop and have all of the required packages installed.

Any help is appreciated.

Thanks!

-Nick G.

[Paul Warren]

I've had problems as well, but I think that's more to do with the
antenna I'm using, than anything else! I'm using a discone I built (
https://plus.google.com/photos/105903369207307996868/albums/5772418318465371121
), which has terrible properties for satellite reception, but I can
hear the same noise as in Alexandru's youtube clip, just it's very
quiet, and has a lot of noise. In the polygon trace, I can't tell the
signal from the noise, in the waterfall, I can just see patterns
emerging. but it's not very clear at all.

How clear is the audio for you, I think that's the main thing needed
for wxtoimg to work well.

Going further, I think I'll be building a QHA, or a spiral + tracker
comboto get my SNR up a bit.


Cheers
--
Paul.

> --
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To post to this group, send email to gq...@googlegroups.com.
> To unsubscribe from this group, send email to
> gqrx+uns...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/gqrx/-/Z1OsTjrjcL0J.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>



--
Paul Warren
p dot a dot warren at gmail dot com
pawarren.blogspot.com

[Nick Gregory]

Well its hard to say how clear the sound is for me, but one thing that I'm pretty sure is important that the sound I'm receiving is lacking is the "tick-tock" that you can hear in Alexandru's video. I can clearly hear the high pitched signal, but (I think) the line marker (the "tick-tock") is getting caught up in the static.

Another thing, if I set the filter to "User (45k)" as it says in the guide, it looks to be encompassing about 3x the spectrum that the signal I can see on the waterfall is taking up. I have tried using the 45k filter as well as tuning it myself so it just encompasses the signal I can see (this turns out to be around 18k if I remember correctly).

Do you think a preamp of some sort would help? Im only running around 15 feet of co-ax so signal loss over the co-ax isn't really an issue, but do you think a preamp would help the overall SNR?

Lastly, does NOAA 17's APT still work? It says on the status page that it does, but nowhere I look can I find a recent image. I ask because there is a ideal (90 degree MEL) pass tomorrow, but if the APT isnt working, no reason to get up early to get it :).

Ill post a picture of my screen during the next pass as well as the captured audio.

> gqrx+unsubscribe@googlegroups.com.

[Nick Gregory]

Oh and another question I had: is there any specific frequency to tune the FCD PLL to relative to the satellite frequency? I read in the original funcube article that it should be +/- 23kHz above the desired frequency, but in the RTL video, it is set to 173kHz above the satellite frequency (with the filter set accordingly).

Would this make a difference?

Thanks

[Paul Warren]

I think with the RTL-SDRs, it's important to have your frequency of
interest somewhat offset from the tuned frequency. It's something to
do with I/Q balancing that I don't quite follow, but means that
there's a big noise peak at the tuned frequency. So setting your
filter offset to -150kHz or so allows for that noise peak to be well
outside the frequencies of interest.

I think that the funcube dongle can't receive as much bandwidth at
once, and so I don't think you can set an offset as large as 100kHz,
hence they say to use a smaller offset, but I'd settle for anything
that lets the full 45kHz of filter bandwidth fit in without covering
the centre.

In terms of filter width, I'm just going on what Alexandru said :) It
did seem to be a bit wider than what I could see of the signal, that
might change with a correctly tuned antenna though.

Glad someone else is on the same journey as me with SDR and weather
satellites :)

Cheers
--
Paul

>>> > gqrx+uns...@googlegroups.com.

>>> > To view this discussion on the web visit
>>> > https://groups.google.com/d/msg/gqrx/-/Z1OsTjrjcL0J.
>>> > For more options, visit https://groups.google.com/groups/opt_out.
>>> >
>>> >
>>>
>>>
>>>
>>> --
>>> Paul Warren
>>> p dot a dot warren at gmail dot com
>>> pawarren.blogspot.com
>

> --
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To post to this group, send email to gq...@googlegroups.com.
> To unsubscribe from this group, send email to

> gqrx+uns...@googlegroups.com.

> To view this discussion on the web visit

> https://groups.google.com/d/msg/gqrx/-/1SStuv1O_7sJ.

[Nick Gregory]

Yea, I was wondering if the spike on the tuning frequency was "real" or just some strange artifact that didnt actually hinder performance.

Ill let you know if I discover anything, or if it is just a matter of the antenna.

Do you have any images that you can post so I can do a comparison to what I'm getting (or are they all blanketed in static)?

Thanks.

-Nick

>>> > gqrx+unsubscribe@googlegroups.com.

>>> > To view this discussion on the web visit
>>> > https://groups.google.com/d/msg/gqrx/-/Z1OsTjrjcL0J.
>>> > For more options, visit https://groups.google.com/groups/opt_out.
>>> >
>>> >
>>>
>>>
>>>
>>> --
>>> Paul Warren
>>> p dot a dot warren at gmail dot com
>>> pawarren.blogspot.com
>
> --
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To post to this group, send email to gq...@googlegroups.com.
> To unsubscribe from this group, send email to

> gqrx+unsubscribe@googlegroups.com.

[Paul Warren]

Yeah, pretty much static filled images so far, with a hint of some
sort of timing/calibration stripe down one side.

Have just run the recording through wxtoimg a few different ways,
comes out a bit better than atpdec, you should see them on my album
at:

https://plus.google.com/photos/105903369207307996868/albums/5772418318465371121

Cheers
Paul

>> >>> > gqrx+uns...@googlegroups.com.

>> >>> > To view this discussion on the web visit
>> >>> > https://groups.google.com/d/msg/gqrx/-/Z1OsTjrjcL0J.
>> >>> > For more options, visit https://groups.google.com/groups/opt_out.
>> >>> >
>> >>> >
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> Paul Warren
>> >>> p dot a dot warren at gmail dot com
>> >>> pawarren.blogspot.com
>> >
>> > --
>> > You received this message because you are subscribed to the Google
>> > Groups
>> > "Gqrx SDR" group.
>> > To post to this group, send email to gq...@googlegroups.com.
>> > To unsubscribe from this group, send email to

>> > gqrx+uns...@googlegroups.com.

>> > To view this discussion on the web visit
>> > https://groups.google.com/d/msg/gqrx/-/1SStuv1O_7sJ.
>> >
>> > For more options, visit https://groups.google.com/groups/opt_out.
>> >
>> >
>>
>>
>>
>> --
>> Paul Warren
>> p dot a dot warren at gmail dot com
>> pawarren.blogspot.com
>

> --
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To post to this group, send email to gq...@googlegroups.com.
> To unsubscribe from this group, send email to

> gqrx+uns...@googlegroups.com.

> To view this discussion on the web visit

> https://groups.google.com/d/msg/gqrx/-/JfXUPKZ9sewJ.

[Nick Gregory]

Yea. About 2 years ago I was doing this with just a radioshack scanner and basically the same antenna im using now and got these results:  http://www.nickproductions.com/wxtoimg/index.html.

Im not sure if the antenna has just fallen apart (it doesnt appear to have, but the coax may have gotten messed up somehow?) or if im doing something wrong on the SDR side.

>> >>> > gqrx+unsubscribe@googlegroups.com.

>> >>> > To view this discussion on the web visit
>> >>> > https://groups.google.com/d/msg/gqrx/-/Z1OsTjrjcL0J.
>> >>> > For more options, visit https://groups.google.com/groups/opt_out.
>> >>> >
>> >>> >
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> Paul Warren
>> >>> p dot a dot warren at gmail dot com
>> >>> pawarren.blogspot.com
>> >
>> > --
>> > You received this message because you are subscribed to the Google
>> > Groups
>> > "Gqrx SDR" group.
>> > To post to this group, send email to gq...@googlegroups.com.
>> > To unsubscribe from this group, send email to

>> > gqrx+unsubscribe@googlegroups.com.

>> > To view this discussion on the web visit
>> > https://groups.google.com/d/msg/gqrx/-/1SStuv1O_7sJ.
>> >
>> > For more options, visit https://groups.google.com/groups/opt_out.
>> >
>> >
>>
>>
>>
>> --
>> Paul Warren
>> p dot a dot warren at gmail dot com
>> pawarren.blogspot.com
>
> --
> You received this message because you are subscribed to the Google Groups
> "Gqrx SDR" group.
> To post to this group, send email to gq...@googlegroups.com.
> To unsubscribe from this group, send email to

> gqrx+unsubscribe@googlegroups.com.

[Alexandru Csete]

On Mon, Aug 6, 2012 at 3:33 AM, Nick Gregory <compute...@gmail.com> wrote:

Well its hard to say how clear the sound is for me, but one thing that I'm pretty sure is important that the sound I'm receiving is lacking is the "tick-tock" that you can hear in Alexandru's video. I can clearly hear the high pitched signal, but (I think) the line marker (the "tick-tock") is getting caught up in the static.

Another thing, if I set the filter to "User (45k)" as it says in the guide, it looks to be encompassing about 3x the spectrum that the signal I can see on the waterfall is taking up. I have tried using the 45k filter as well as tuning it myself so it just encompasses the signal I can see (this turns out to be around 18k if I remember correctly).

The bandwidth of an FM modulated signal depends on both the deviation and the modulating audio. As I explained in the Funcube Dongle article the total bandwidth for APT is

BW = 2 × (17 + 2.4) kHz = 38.8 kHz

Then you need to add a few kHz for Doppler shift.

As you can see on the screenshots there, the APT signal pretty much fills out the full filter width. If your signal does not do so, then either something is wrong with the signal (e.g. satellite broken) or the signal is too weak. what is the difference in dB between the peak of the APT signal and the noise floor? I think you will need at least 20 dB for good reception, probably more.

 

Do you think a preamp of some sort would help? Im only running around 15 feet of co-ax so signal loss over the co-ax isn't really an issue, but do you think a preamp would help the overall SNR?

Since the RTL dongle has no preamp in front of the tuner, a very low noise preamp mounted close to the antenna might reduce the overall noise figure of your receiver. I only tested with Arrow antenna connected to the receiver with very short coax.
 

Alex

[Paul Warren]

Hrm, I didn't get a screenshot, but my signal was not detectable above
the noise in the top graph (What's that called by the way?) In the
bottom one, I could see repetitive structures, and the 45KHz bandwidth
looked a lot skinnier than in that screenshot.

I've figured out where to set the deviation now. Because I don't
really know what I'm doing yet, I assumed that was taken care of by
setting the mode to FM-N, but for reference, Click the Settings
(Spanner and Hammer) button next to Mode, and you can set the
deviation and Tau. Currently, I've no idea what that means, but off I
go to do some reading :)

/contemplates building a cross-yagi from coathangers in the next 3
hours before NOAA 19 is overhead...

Thanks Alex :)

[Alexandru Csete]

Tau is the time constant for the de-emphasis filter, see
http://en.wikipedia.org/wiki/FM_radio#Pre-emphasis_and_de-emphasis
As suggested by the tooltip, it is best to switch it off when sending
data over FM channel.

Alex

[Nick Gregory]

Hmm...

My signal was pretty obvious from the noise in the top view, and in the waterfall I could clearly make out the signal difference ("lines" of the signal). Ill check the difference between noise and the signal on the next pass.

Paul: One thing I noticed was that if you changed a setting (say in the settings for the demod.) the filter bandwidth would still say 45k, but it would in reality be 10k (if you hold CTRL and scroll as you would to change it, the bandwidth jumps down to 10k then starts going back up). That may be what was happening to you.

Alexandru: Is there a minimum sampling rate that you need for this? Im running it at the lowest because I am running Gqrx inside a ubuntu VM on my MacBook Pro, but I should be able to push it higher.

Another thing I may try is using a amplified rabbit ear antenna for the TV just to get a comparison.

One last thing, would turning AGC on somehow mess up the signal? I dont remember, but I think AGC may be on for me...

Thanks.

[Alexandru Csete]

On Mon, Aug 6, 2012 at 7:33 PM, Nick Gregory <compute...@gmail.com> wrote:
> Hmm...
>
> My signal was pretty obvious from the noise in the top view, and in the
> waterfall I could clearly make out the signal difference ("lines" of the
> signal). Ill check the difference between noise and the signal on the next
> pass.
>
> Paul: One thing I noticed was that if you changed a setting (say in the
> settings for the demod.) the filter bandwidth would still say 45k, but it
> would in reality be 10k (if you hold CTRL and scroll as you would to change
> it, the bandwidth jumps down to 10k then starts going back up). That may be
> what was happening to you.
>
> Alexandru: Is there a minimum sampling rate that you need for this? Im
> running it at the lowest because I am running Gqrx inside a ubuntu VM on my
> MacBook Pro, but I should be able to push it higher.

The minimum sample rate is what the desired bandwidth is, i.e. ~ 50
ksps, so anything for rtlsdr (900+ ksps) should work fine. You should
however check that your device works properly at the rate you are
using, e.g. by trying to listen to WFM and check that it "sounds
right".

> Another thing I may try is using a amplified rabbit ear antenna for the TV
> just to get a comparison.
>
> One last thing, would turning AGC on somehow mess up the signal? I dont
> remember, but I think AGC may be on for me...

AGC has very little impact on FM-type signals because frequency
deviation/shift is translated to audio level. However, it is a good
idea to leave the AGC on to prevent the signal path to be saturated in
case of very strong signals.

Alex

[Nick Gregory]

Yea, listening to FM radio works fine.

Unfortunately there are not very many good passes where I am over the next few days, but ill report back with more detailed info once there is a descent pass.

[Nick Gregory]

Alright, good news and bad news.

Good news: I got a partial image. Got up at 3:30AM today to get a 90 degree pass to see what it sounded like and it paid off.

Bad news: Partial image is completely unrecognizable. This is the decoded image from wxtoimg after using sox to resample it ("sox orig_file new_file rate 11025" is the command used.)

Suggestions?

Here is the original audio file and here is the resampled file.

[Robin Gape]

Nick,

looks to this pair of eyes as if there is slant on a very featureless image, an IR channel, perhaps or a lot of cloud cover. Have you tried the suggestions in the Image Slant section of http://www.wxtoimg.com/support/wxfaq.html?

According to the screenshots, WX to IMG would seem to require audio sampled at 11025.49 Hz. This may be an artefact of the calibration on the example machine used for the FAQ. The sox command quoted will (presumably) precisely resample to 11025.00 Hz, implying one slip every 2 seconds, which is believable for your decoded image. (There may be other issues, of course!)

Good luck

Robin

--

You received this message because you are subscribed to the Google Groups "Gqrx SDR" group.
To post to this group, send email to gq...@googlegroups.com.

To unsubscribe from this group, send email to gqrx+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msg/gqrx/-/wsGvzAzDyHcJ.

[Nick Gregory]

OK, well I tried 2 things:

1) Using sox to convert it to 11025.49Hz. No difference (I didnt even bother to look if sox could resample to fractions of a Hz so I didnt really expect this to work).

2) Using WxtoImg's slant correction feature to correct it. Some success. The image is MUCH more recognizable (it appears to be an IR channel and the heat channel (hence the near white on one of the channels). That being said, the image still is far from perfect. There are still misalignment issues as you can see in the image here. Also, when I do this, the slant correction window comes up and says "NOAA sampling frequency should be 10973.64" which leads me to believe two things:

    a) The 11025.49Hz shown in the example photo is the slight deviation from the wanted 11025Hz that is causing the slant in the demo photo (on the wxtoimg website)

    b) Something else is really wrong here.

Oppinions? Suggestions?

There is a pass of the meteor 3-5 satellite tonight at around 11pm which I will probably catch (its only a 56 degree, but ill be able to see if there is a difference between the reception of the NOAA satellites and others.) I know there shouldn't be a difference, but who knows.

Thanks!

[Nick Gregory]

Well... it would appear virtualbox has managed to screw this up. After installing ubuntu (not in a VM) on my laptop, what do you know, no weird slant artifacts. Here is the image from the most recent pass (just a few minutes ago). It was an 89 degree pass, so more is expected out of this setup than what I am getting, but that is on the antenna side I am almost certain. I'll probably try and rig a amplified TV antenna using my antenna instead of the rabbit ears and see if that can act as a preamp of sorts and improve the signal.

[Alexandru Csete]

Hi Nick,

Thanks for confirming that trying to run a real time task (an SDR) in
a non-real time operating system, running inside a VM, running in
another non-real time operating system doesn't really work :-)

Jokes aside, if you open your original wav file in audacity you will
notice that the audio level is way too high and saturates very often.
The full scale level is +/- 1.0 and you should stay below that.
Personally I prefer -10 to -15 dB below full scale, which is is
sufficiently loud and leaves some margin. Since APT uses the amplitude
of the 2.4 kHz tone to encode the pixel values, you will loose
contrast if the amplitude saturates. The decoder may also have trouble
tracking the tone.

If you are ging to use a preamp, it will only make sense if it has
significantly lower noise figure than the rtl dongle (something like 1
dB) and/or you have very long cable with very high loss. I doubt any
TV antenna amplifier will give you any improvement.

Alex

> https://groups.google.com/d/msg/gqrx/-/QhB19ayYhV8J.

[Nick Gregory]

Yea yea yea, the only reason I was running a VM is because Ubuntu has never liked my laptop (graphical glitches, lack of compatibility with hardware, etc.) , but it seems to have been resolved with 12.04.

I also noticed that the wav was clipping (after the fact of course.) Now im setting the audio gain to around -12dB and that seems to work well.

Lastly, for the most recent pass, I tried using just the TV antenna (not using it as a preamp, but using the "rabbit ears" and the built in amp) and got some amazing results. I was able to get a very good picture all the way down to 8 degrees. My QFH with no preamp can only pick up the signal when the satellite is > 75 degrees(ish) elevation. Im thinking for now im just going to set up the TV antenna (probably on the roof) and see how that goes.

Today has been a good day for testing as there have been 2 >80 degree passes today and one more in about 45 minutes that Ill be trying to get with the TV antenna.

Thanks for the advise with the preamp. Now that I know that TV antenna works as well as it does, I need to figure out if it works so well because it has a built in amplifier, of if it works well compared to the QFH because something is seriously wrong with the QFH antenna.

[Nick Gregory]

One last post for good measure.

Here is the most recent image I got using the TV antenna.

Thanks everyone!