Sharp drop in SNR during TCH

Bhaskar11 niceguy108 at
Wed Mar 20 15:37:40 UTC 2013

Dear Sylvain,

I'm studying traffic data using ccch_scan in burstind branch.

In prim_sniff.c I track the SNR levels in l1s_sniff_resp() using code from
print_tch.c to read snr values from: dsp_api.db_r->a_serv_demod[D_SNR];

I find that the level changes sharply from ccch mode to tch mode.

In ccch mode most of the snr values are 16384. After switching to tch mode
they range from 200 to 600 typically!

Is this normal? Am I doing something wrong when changing mode?

Below is the dump of snr values and freq_err during the run.


=============== data dump
fnx = fn % 26;
freq_err = ANGLE_TO_FREQ(dsp_api.db_r->a_serv_demod[D_ANGLE]);
snr      = dsp_api.db_r->a_serv_demod[D_SNR];

fnx=6, bidx=0, freq_err=-24, snr=16384. SACCH
fnx=7, bidx=0, freq_err=-6, snr=16384. SACCH
fnx=8, bidx=0, freq_err=9, snr=16384. SACCH
fnx=9, bidx=0, freq_err=-15, snr=16384. SACCH
fnx=25, bidx=0, freq_err=-26, snr=16384.
fnx=0, bidx=0, freq_err=10, snr=16384.
fnx=1, bidx=0, freq_err=38, snr=16384.
fnx=2, bidx=0, freq_err=-14, snr=*16384*.
fnx=24, bidx=0, freq_err=46, snr=*16384*.
fnx=25, bidx=0, freq_err=23, snr=*16384*.
fnx=0, bidx=0, freq_err=-1, snr=*16384*.
afc_correct(error=-1Hz, arfcn=24): delta=0, afc_dac(old=544,new=544)
fnx=1, bidx=0, freq_err=22, snr=16384.
*(Immediate Assignment starts here)*
L1CTL_FBSB_REQ (arfcn=4, flags=0x7)
Starting FCCH RecognitionFB0 (1266881:8): TOA= 8784, Power= -61dBm,
afc_correct(error=10831Hz, arfcn=4): delta=1283, afc_dac(old=-700,new=583)
FB0 (1266892:9): TOA=10032, Power= -61dBm, Angle= -340Hz
afc_correct(error=-340Hz, arfcn=4): delta=-40, afc_dac(old=583,new=543)
FB1 (1266902:8): TOA= 8755, Power= -61dBm, Angle=    0Hz
afc_correct(error=0Hz, arfcn=4): delta=0, afc_dac(old=543,new=543)
  fn_offset=1266900 (fn=1266902 + attempt=8 + ntdma = 6)
  delay=8 (fn_offset=1266900 + 11 - fn=1266902 - 1
  scheduling next FB/SB detection task with delay 8
=>FB @ FNR 1266900 fn_offset=1266900 qbits=4836
LOST 3717!
SB1 (2533807:1): TOA=   26, Power= -61dBm, Angle=  176Hz
=> SB 0x00a9dd65: BSIC=25 fn=1266912(955/10/21)=> DSP reports SB in bit
that is 1127709799 bits in the future?!?
=> DSP reports FB in bit that is 1127709776 bits in the future?!?
LOST 1907!
afc_correct(error=4Hz, arfcn=4): delta=0, afc_dac(old=543,new=543)
L1CTL_DM_EST_REQ (arfcn=12545, chan_nr=0x15, tsc=1)
LOST 3047!
*(tch mode is now active)*
fnx=4, bidx=0, freq_err=-92, snr=*238*.
fnx=6, bidx=0, freq_err=-132, snr=*82*.
fnx=8, bidx=0, freq_err=103, snr=*145*.
fnx=10, bidx=0, freq_err=-234, snr=*316*.
fnx=12, bidx=0, freq_err=-89, snr=*292*. SACCH
fnx=13, bidx=0, freq_err=46, snr=*270*.
fnx=15, bidx=0, freq_err=35, snr=*387*.
fnx=17, bidx=0, freq_err=201, snr=*1055*.
fnx=19, bidx=0, freq_err=130, snr=549.
fnx=21, bidx=0, freq_err=-94, snr=51.
fnx=23, bidx=0, freq_err=-52, snr=173.
fnx=0, bidx=0, freq_err=157, snr=322.
fnx=2, bidx=0, freq_err=107, snr=95.
afc_correct(error=12Hz, arfcn=24): delta=1, afc_dac(old=543,new=544)
fnx=4, bidx=0, freq_err=86, snr=135.
fnx=6, bidx=0, freq_err=-126, snr=207.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the baseband-devel mailing list