search

ptrkrysik / gr-gsm

Gnuradio blocks and tools for receiving GSM transmissions
Other
1.33k stars 428 forks source link

grgsm_scanner duplicated and wrong results #421

Open mazen-bassiouny opened 6 years ago

mazen-bassiouny commented 6 years ago

Dears, I am getting duplicated results with the scanner application , same cell Id on different frequencies which cannot be correct.,and tested with mobile application. For example: these are the results of scanning GSM900 band root@mazen-ThinkPad-T440p:/usr/local/src/gr-gsm/apps# grgsm_scanner -b GSM900

ARFCN: 995, Freq: 929.2M, CID: 26821, LAC: 36150, MCC: 262, MNC: 3, Pwr: -68 ARFCN: 1006, Freq: 931.4M, CID: 9058, LAC: 36150, MCC: 262, MNC: 3, Pwr: -62 ARFCN: 1010, Freq: 932.2M, CID: 6822, LAC: 36150, MCC: 262, MNC: 3, Pwr: -59 ARFCN: 1017, Freq: 933.6M, CID: 46821, LAC: 36150, MCC: 262, MNC: 3, Pwr: -57 ARFCN: 19, Freq: 938.8M, CID: 19401, LAC: 4291, MCC: 262, MNC: 2, Pwr: -59 ARFCN: 29, Freq: 940.8M, CID: 19401, LAC: 4291, MCC: 262, MNC: 2, Pwr: -56 ARFCN: 53, Freq: 945.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -52 ARFCN: 57, Freq: 946.4M, CID: 3322, LAC: 38722, MCC: 262, MNC: 1, Pwr: -44 ARFCN: 63, Freq: 947.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -25 ARFCN: 67, Freq: 948.4M, CID: 3322, LAC: 38722, MCC: 262, MNC: 1, Pwr: -53 ARFCN: 73, Freq: 949.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -57 ARFCN: 110, Freq: 957.0M, CID: 0, LAC: 0, MCC: 0, MNC: 0, Pwr: -44 ARFCN: 120, Freq: 959.0M, CID: 13622, LAC: 38722, MCC: 262, MNC: 1, Pwr: -38

Noting for CID :13522 ,,, it comes on different ARFCNs although the correct one is only (63) also CID: 19401 ,,,, it comes on 3 ARFCNs although the correct one is only (29)

can you please check ?

mazen-bassiouny commented 6 years ago

Any help here please ?:)

velichkov commented 6 years ago

Hi @mazen-bassiouny,

Thanks for reporting this and thanks for the reminder. :wink: Have you tried to debug it? Based on our previous discussions I think you have all the knowledge and skills to resolve this.

ARFCN: 19, Freq: 938.8M, CID: 19401, LAC: 4291, MCC: 262, MNC: 2, Pwr: -59 ARFCN: 29, Freq: 940.8M, CID: 19401, LAC: 4291, MCC: 262, MNC: 2, Pwr: -56

ARFCN: 53, Freq: 945.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -52 ARFCN: 63, Freq: 947.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -25 ARFCN: 73, Freq: 949.6M, CID: 13522, LAC: 38722, MCC: 262, MNC: 1, Pwr: -57

ARFCN: 57, Freq: 946.4M, CID: 3322, LAC: 38722, MCC: 262, MNC: 1, Pwr: -44 ARFCN: 67, Freq: 948.4M, CID: 3322, LAC: 38722, MCC: 262, MNC: 1, Pwr: -53

There is 2MHz between each duplicated CID and knowing that the default sampling rate is 2e6 (=2MHz) my first guess is that you are experiencing some kind of aliasing here.

Try with 1.6e6, 2.4e6, 3.2e6, 4e6, and 8e6 sample rates and post the full output.

ptrkrysik commented 6 years ago

I was looking into this problem and it seems that it is result of aliasing in analog fron-end of RTL-SDR. As grgsm_scanner channelizes whole available spectrum aliases can be seen as false-positives. There are probably many solutions to this problem, i.e.:

I'm quite reluctant to do it myself at the moment due to other tasks.