Hi,
I have made some captures on the L-band uplink channels and tried to decode the RACH messages. All four consecutive TCH channels (according to the spot beam) were recorded and a call was made during the recording.
I am using the GRC block gmr_rach_scan.grc. I have removed the IQ swap from the graph, as suggested in the mailing list. The sampling rate and samples-per-symbol values have been set to 93.6k and 4, respectively.
It ran fine without any errors, but I couldn't see any results either on the wireshark or under /tmp/rach.txt.
Maybe I overlooked something here. Any ideas or remarks?
Many thanks, Johann
Hi,
I have made some captures on the L-band uplink channels and tried to decode the RACH messages. All four consecutive TCH channels (according to the spot beam) were recorded and a call was made during the recording.
I am using the GRC block gmr_rach_scan.grc. I have removed the IQ swap from the graph, as suggested in the mailing list. The sampling rate and samples-per-symbol values have been set to 93.6k and 4, respectively.
Oops, looks like I complete missed this ... sorry.
That graph was made to scan large bandwidth capture for RACH ... It's very probable that a bunch of constants and thresholds value are just not appropriate for a narrow band capture. It kind of always required tweaking depending on the case. Basically manually finding a RACH (in frequency and time), then look at all the intermediate steps of the graphs (and possibly inside the block) to see if it detects that particular one and tweak it until it does. And then you can run it to find the other ones.
I don't have any setup any more to capture C-band uplink so I have not run that graph in years .... but that's all I remember about it.
Cheers,
Sylvain
Hi Sylvain,
Thank you for getting back on this. I will look further into the block as you suggested. It is also possible that the uplink signal of my initial capture was too strong and somehow saturated the receiver.
I plan to perform more tests in the coming weeks as the weather is getting better.
Best regards, Johann
On Tue, May 25, 2021 at 7:26 PM Sylvain Munaut 246tnt@gmail.com wrote:
Hi,
I have made some captures on the L-band uplink channels and tried to
decode the RACH messages. All four consecutive TCH channels (according to the spot beam) were recorded and a call was made during the recording.
I am using the GRC block gmr_rach_scan.grc. I have removed the IQ swap
from the graph, as suggested in the mailing list. The sampling rate and samples-per-symbol values have been set to 93.6k and 4, respectively.
Oops, looks like I complete missed this ... sorry.
That graph was made to scan large bandwidth capture for RACH ... It's very probable that a bunch of constants and thresholds value are just not appropriate for a narrow band capture. It kind of always required tweaking depending on the case. Basically manually finding a RACH (in frequency and time), then look at all the intermediate steps of the graphs (and possibly inside the block) to see if it detects that particular one and tweak it until it does. And then you can run it to find the other ones.
I don't have any setup any more to capture C-band uplink so I have not run that graph in years .... but that's all I remember about it.
Cheers,
Sylvain
Hi Sylvain,
Thanks to your suggestions, I managed to decode the RACH from the captures on L-band. I just had to tweak a few parameters in the graphs. I have decoded the Paging Response and Location Update messages.
Do you have some suggestions on decoding the uplink speech? On the downlink, the code follows the Immediate Assignment and decodes the traffic channel. Due to the limitation of the SDR, I am not able to capture both the downlink and uplink channels simultaneously.
Thank you again.
Best regards, Johann
On Tue, May 25, 2021 at 7:26 PM Sylvain Munaut 246tnt@gmail.com wrote:
Hi,
I have made some captures on the L-band uplink channels and tried to
decode the RACH messages. All four consecutive TCH channels (according to the spot beam) were recorded and a call was made during the recording.
I am using the GRC block gmr_rach_scan.grc. I have removed the IQ swap
from the graph, as suggested in the mailing list. The sampling rate and samples-per-symbol values have been set to 93.6k and 4, respectively.
Oops, looks like I complete missed this ... sorry.
That graph was made to scan large bandwidth capture for RACH ... It's very probable that a bunch of constants and thresholds value are just not appropriate for a narrow band capture. It kind of always required tweaking depending on the case. Basically manually finding a RACH (in frequency and time), then look at all the intermediate steps of the graphs (and possibly inside the block) to see if it detects that particular one and tweak it until it does. And then you can run it to find the other ones.
I don't have any setup any more to capture C-band uplink so I have not run that graph in years .... but that's all I remember about it.
Cheers,
Sylvain
Hi,
Thanks to your suggestions, I managed to decode the RACH from the captures on L-band. I just had to tweak a few parameters in the graphs. I have decoded the Paging Response and Location Update messages.
TBH I'm impressed you managed to decode those from uplink without the sync from the downlink channels.
Do you have some suggestions on decoding the uplink speech? On the downlink, the code follows the Immediate Assignment and decodes the traffic channel. Due to the limitation of the SDR, I am not able to capture both the downlink and uplink channels simultaneously.
Not really, they are already tricky to get from downlink because they are very small, use crappy sync patterns and have very little error correction applied, so on uplink that's going to make it even harder.
Cheers,
Sylvain
Hi Sylvain,
Thank you for getting back to me. I suspected it would not be easy to decode the uplink speech and you confirmed that. I will probably not spend more effort on it.
Your code and suggestions have helped me a lot. I have decoded my own downlink speech and now the RACH. Thanks again for your help.
Best regards, Johann
On Thu, Jul 1, 2021 at 10:14 AM Sylvain Munaut 246tnt@gmail.com wrote:
Hi,
Thanks to your suggestions, I managed to decode the RACH from the
captures on L-band. I just had to tweak a few parameters in the graphs. I have decoded the Paging Response and Location Update messages.
TBH I'm impressed you managed to decode those from uplink without the sync from the downlink channels.
Do you have some suggestions on decoding the uplink speech? On the
downlink, the code follows the Immediate Assignment and decodes the traffic channel. Due to the limitation of the SDR, I am not able to capture both the downlink and uplink channels simultaneously.
Not really, they are already tricky to get from downlink because they are very small, use crappy sync patterns and have very little error correction applied, so on uplink that's going to make it even harder.
Cheers,
Sylvain
Hi,
What SDR are you using btw ?
I know several of the mimo sdr (limesdr / b210 / ...) can RX on multiple channels but tuned to the same frequency. And I had the idea of using an external mixer to shift the uplink with an external mixer to make it match the downlink frequency. Never tried it in practice though.
Cheers,
Sylvain
Hi Sylvain,
I am using both the USRP B210 and B205-mini. So far I haven't attempted to receive on multiple channels. It might be an idea.
Best regards, Johann
On Fri, Jul 2, 2021 at 9:54 AM Sylvain Munaut 246tnt@gmail.com wrote:
Hi,
What SDR are you using btw ?
I know several of the mimo sdr (limesdr / b210 / ...) can RX on multiple channels but tuned to the same frequency. And I had the idea of using an external mixer to shift the uplink with an external mixer to make it match the downlink frequency. Never tried it in practice though.
Cheers,
Sylvain
Hi Johann,
I'm trying to get in touch with the GMR PHY signals, but I don't have proper equipment to record uplink signals myself. Could you please share the uplink recording you mentioned in this post? Would be very helpful.
Thanks, Jon
-
Johann T. -
jon.hofmann＠web.de -
Sylvain Munaut