Using C band kit to monitor Inmarsat Aero H/H+

Status
Please reply by conversation.

gsmforensics

Active SatelliteGuys Member
Original poster
Sep 18, 2016
17
34
UK
Sept11.gif



Following the disappearance of MH370 back in 2014 Inmarsat offered to provide free 15 minute check ins for commercial Oceanic flights via their ADS system.
This was to encourage operators to log in with their location on a more frequent basis.
The company conducted tests last year which were deemed successful in getting operators to speed up their
installation of Sat Comm equipment.
As can be seen from the above time lapse animation from PlanePlotter , our Oceans are teeming with
a/c as they ply back and forth across the Atlantic , Indian and Pacific Oceans.

PlanePlotter has been around for well over ten years.It was the first a/c tracking system for enthusiasts and started out as a network to plot ACARS. It seems fitting that now , following years of development and the introduction of many ground breaking techniques for a/c enthusiasts , it has come full circle to encompass Satellite ACARS as provided by ADS.

Back in the early noughties I'm sure many of the forum members will recall monitoring Inmarsat L band analogue telephony.My main interest was picking up the NASA downrange comms on launch days. It could be a little frustrating as only one side of the comms could be heard. However if you hooked up your C band dish to the same satellite , then you could monitor the other side via the C band downlink.

The same principle applies to Aero H/H+
A/c contact ground stations via L band .... the signal is fired up to the satellite and then converted to C band for the earth bound downlink.

The data contains house keeping info , plus those all important position reports.
Those reports , once processed can be plotted on the map.
Late last year , a chap called Jonti created a piece of software [JAERO ] which processed the
data bursts from Inmarsat . Initially this was just for L band , but once we realised that an 8metre dish was not required for the C band work , he set about processing the C band data [1520-1545 MHz IF ]

The technique is relatively simple and should strike a chord with satellite enthusiasts.

Using a stable C Band lnb [ we recommend the Titanium PLL C band devices] the satellite is located and the signals fed into a standard sat receiver which has loopthrough. A voltage inserter can be used instead of the receiver if you prefer. From the receiver/inserter , it is fed to a basic RTL type dongle.
The dongle plugs into your PC
SDR # software then processes the data and passes it via soundcard or virtual cable to JAERO.
JAERO turns the data burts into ACARS format...the latter is then sent to PlanePlotter which takes the positional information and plots the a/c on the map.

blueprint.jpg



Tests have shown that dishes as small as 90cm can receive the data , but ideally 1.2m or more should be used.

So apart from the dish, receiver and lnb , most of which many of us already have , all that's left is the dongle , just a few dollars , and the software. Both SDR# and JAERO are free.
If you don't already have PlanePlotter , then for anyone who is prepared to commit to a 24/7 Inmarsat feed into the PP network , that too could be free.

This is a fascinating aspect of our hobby and as dish users ,we have a head start !

If interested , by all means PM me for any additional info.

Best wishes

John
 
Thanks for all the info John. Yes, this is very very cool!!
 
Something I forgot to mention.
If you are planning to give this a go , the latest version of PlanePlotter actually tracks the Inmarsats for you in real time so that you can see exactly where the birds are in relation to the arc. This helps to position your dish accurately as some of the western satellites are quite well inclined.

When combined with standard ADS-B / Mode S traffic , this is the result....the Oceanic traffic fills the gap between the land masses seamlessly.

Atlantic all.jpg


Best wishes

John
 
  • Like
Reactions: Titanium
As a dongle-holic this is a very interesting application of an RTL-SDR. As for the dongle part I would suggest using the best quality available, that is, those with a stable TXCO and R820T2 chip such as Nooelec's or the RTL-SDR.com version. They are only several dollars more than the cheapies. Note that I don't work for either of these aforementioned reputable dongle manufacturers; just a satisfied customer!
 
Good point spongella , although I use both the basic dongle and the slightly more expensive one and find little difference for this type of application.
I do however struggle a little with temp drift on the lnb. I'm not proficient enough to do the recommended modifications as explained here
titanium-satellite-pll-lnbf-osmocom-hack
 
  • Like
Reactions: hank123
...
I do however struggle a little with temp drift on the lnb. I'm not proficient enough to do the recommended modifications as explained here
titanium-satellite-pll-lnbf-osmocom-hack
I have ruined 2 Titanium C2W PLL LNBs in my modification attempts for something different last year. (I needed to remove the circuit board so I could access and shorten the rear antenna wire for better performance around 4000 MHz Vertical Polarity. It had about 1 mm more of it sticking out inside the tube than the Horizontal wire because of the way they attached it at the factory) Removing the circuit board from the case required unsoldering the RG6 center conductor. On both attempts when I tried it some of the circuit traces on the board lifted up and killed it.

While the osmocom pll lnbf modification does not require removing the circuit board from the case, the scariest part for me would be removing the crystal. After that, it should be much less risky adding the bridge wire and the other 3 components. But then it gets expensive -- the "Mux Tee" from Orbital Research says "Call for Pricing". From past experience, anything that says that is usually out of reach financially.

I'm wondering if Jonti's 'AERO' software could be tweaked to compensate for the LNB frequency drift? It already has AFC, which works very well when I tried it with L-Band. Could the AFC be widened to work with C-Band, or am I missing an important detail here?
 
  • Like
Reactions: hank123
If you are following this thread you may be interested to know that the latest version of PP is now available...free for registered users.
 
  • Like
Reactions: hank123 and KE4EST
Good point spongella , although I use both the basic dongle and the slightly more expensive one and find little difference for this type of application.
I do however struggle a little with temp drift on the lnb. I'm not proficient enough to do the recommended modifications as explained here
titanium-satellite-pll-lnbf-osmocom-hack

What frequency stability is required for this project? Isn't some of the accumulated drift associated with the DVB-T dongle? Might the project benefit from a higher stability dongle and LNBs like +/-5KHz or +/-10KHz Norsats? I understand that these are likely to cost considerably more, but they might be a better match for the extremely narrow bandwidth signal.

We spec the standard Titanium LNBFs @ +/-50KHz and the lite model @ +/-100KHz, which is excellent compared to regular DRO LNBFs +/-3MHz. The external mod that you reference makes this reference frequency exact, but even more importantly provides increased SNR. Not sure it would address minute thermal drift as this is primarily introduced component and surface mount variances not related to the reference frequency.
 
  • Like
Reactions: hank123
Wonder if SDR# can track the carrier with AFC or something similar? Maybe an add-on...Then the requirment for stability would not be such an issue.
 
  • Like
Reactions: Titanium
Hi Titanium....I should make it clear that my comment was in no way a criticism of the PLL LNB , it was more aimed at my inability to do the mod !
The drift I see is only about 30khz , well within the design spec of the product. Interestingly the signal drifts in two ways...I see a 10khz "jump" twice a day and then a slow drift over time.
So over 24 hours the signal drifts back and forth...more noticeable on days when we have higher temps.

Cham's suggestion that it could be tracked in some way by SDR# is one that has been considered and I think suggested , but I'm not sure how it could be achieved.

I have a Norsat Ka mil band LNB by the way and can confirm the difference in price is exponential !

For the price we pay the PLL C Band LNB is fantastic value , don't think we could have progressed so far with out it.
I suppose that's what is so intriguing about the project , it has involved expertise from a number of sources to solve various problems, the main components being Titanium , JAERO , SDR# and of course our own developer Bev from COAA/PP who has pulled it all together.

Best wishes

John
 
Sorry if I came off as defensive! Not my intent. I only wish to address and understand the project requirements and if the components are appropriate.

This observed drift could be with either the LNB or the dongle tuner. Both are dependent on the stablity of a reference signal.

I am interested in the twice daily 10KHz jump. Any indication of a trigger? The slow minimal drift within a spec range is the component/hardware variance that is normal. To achieve tighter stability ratings, manufactures sort the production run, label and price accordingly. :)
 
Yes , the 10khz "jump" is interesting , usually around 0700 gmt and again late afternoon in the opposite direction.
At least one other user has observed this. Personally I thought it was satellite related , possibly some sort of shift initiated by Inmarsat for some reason....but that seems unlikely.

Its something I intended to monitor to see if it has any relationship to the inclination of the bird , but never got round to it.

I see the same using both the cheap rtl dongle and the RTL-SDR.com derivative

A wider bandwidth in SDR# might help [its max 16khz ] , however I monitor two T channels on Alphasat which are only 30khz apart , with an "R" channel between them
so a wider bandwidth in SDR# would result in more chance of near channel interference.

Best wishes

John
 
Interesting. Please post any additional finding for this frequency shift. It wouldn't seem to be associated with your system hardware if two locations simultaneously observe the same 10KHz frequency shift.
 
John,

Here is that Earth Map we were chatting about which could be used in place of the wireframe. You could either use this as a texture map to "wrap" around a sphere, or offer an alternative flat map for users of your software.

I used my software (Natural Scene Designer) to add shading. The one below shows the topography on the ocean floor. If you want a flat ocean or a higher resolution map, just let me know and I would be happy to provide that for you
EarthWithShading.jpg
 
Yes , the 10khz "jump" is interesting , usually around 0700 gmt and again late afternoon in the opposite direction.
At least one other user has observed this. Personally I thought it was satellite related , possibly some sort of shift initiated by Inmarsat for some reason....but that seems unlikely. ...
My best guess is that the 10khz jump is due to the satellite switching between batteries and solar power -- twice a day somewhere around the time it enters and exits the earth's shadow. Or it could be the sudden change in temperature. I will monitor it here in California to see if a similar thing happens after I get everything setup.

My initial attempt to receive C-Band circular polarity on 97.6W and 54W failed. I didn't even see a blip in the SDR waterfall. Not sure what I did wrong. I had the LNB powered using an old receiver with "loop out" connected to the SDR. The dielectric plate was inserted. I tried both 13 and 18V on the old receiver.
 
I need to pay more attention to the drift and frequency "jumps"
Interestingly the 10khz jumps now seem to be taking place around 1000-1100 GMT , as the satellite approaches the arc from the north.

Time permitting I will try to log the movements.

Meanwhile Bev is continuing with the " work in progress" rendering project.

Atlantic render.jpg
 
I need to pay more attention to the drift and frequency "jumps"...
Some of the gradual frequency drift is caused by the doppler effect of the satellite's motion relative to your ground location. With an inclined orbit, the distance from the dish to the satellite changes over time -- about half the time its moving closer to you, and the other half its going away.

Also the speed and direction of the transmitting aircraft causes its own doppler shift.

After asking Titanium a few questions today about my new C2WPLL, I managed to receive some T-Packets from 54W using SDR# with the audio piped to JAERO! (Screen shot below) I haven't tried processing the raw data with PP or other software yet. That will wait until I can get better signal strength.
T_PacketsOn54W.JPG
 
Fantastic work Brett !
No dropped packets there , which is great.

If you decide to feed into PP [ which I hope you do ] then make sure that in JAERO you select the port as localhost 8765 [not 18765] and tick output to UDP.
In PP tick ACARS from JAERO.
 
Fantastic work Brett !
No dropped packets there , which is great.

If you decide to feed into PP [ which I hope you do ] then make sure that in JAERO you select the port as localhost 8765 [not 18765] and tick output to UDP.
In PP tick ACARS from JAERO.
Actually there were a lot of dropped packets. I just showed one of the stronger bursts of packets which managed to get through OK. :)

Getting this to work is extremely challenging. The satellite is a slowly moving target (in elevation), and the frequencies move too, as you are well aware.

I have some ideas on hardware and software changes I can do to make this easier. Will post any successes here.
 
  • Like
Reactions: hank123
Actually there were a lot of dropped packets. I just showed one of the stronger bursts of packets which managed to get through OK. :)

Getting this to work is extremely challenging. The satellite is a slowly moving target (in elevation), and the frequencies move too, as you are well aware.

I have some ideas on hardware and software changes I can do to make this easier. Will post any successes here.

Yes , 54 west is a bit of a pain as it is quite well inclined...add to that the signal drift and as you say , its a challenge.
Ideally what I would like to do is set up the network so that , in the case of the more inclined sats , one dish ground station looks at the upper part of the arc , another at the lower.Thus making it easier.
Those using smaller 1.2m dishes have a little less hassle as the beamwidth of the dish is wider so holds the signal a bit longer.

I monitor Alphasat here , which is inclined about 1.5 degs so I only need to adjust my 1.8m twice a day...although I have to adjust frequency every 3 hours or so in daylight. During the night temps are more stable so drift is not too much of an issue.
Akphasat at 25E has pretty wide coverage .....all the plots on the image below , apart from three, are from my dish , Mid Atlantic to Indian Ocean.
alphasatsept30.jpg


Best wishes

John
 
Status
Please reply by conversation.
***

Users Who Are Viewing This Thread (Total: 1, Members: 0, Guests: 1)

Who Read This Thread (Total Members: 2)