Repeated FACH/SACH

hi vang

repeated facch – better outgoing ho efficiency rate
repeated sacch – less call drop rate “amr”

that would be very visible in cells where usage of lower codecs is high. This usage can be assessed through RMS indicators.

could you tell me what is your subset of codecs AMR FR ?

The repeated FACCH or SACCH should not have any impact on TFO.

TFO with AMR… there should be an increase on the number of TFO requests in your network. I don’t know what are the names of the indicators, but all the ones which contain the string ‘TFO’ will definitely reflect the change.

There should be more AMR codec changes, as a result. So here again, you might see such effect in the RMS.

regards
pix

Hi Pix,
After these changes, there are some cells with inceased CHO_DL/UL_quality. Did you have any idea what can be the reasons?
Best regards,

Hi Pix,
this value for cells with inrceased CHO_DL/UL_quality:

AMR_FR_THR_1(dB)=5
AMR_FR_THR_2(dB)=10
AMR_FR_THR_3(dB)=13
KEEP_CODEC_HO=free choice of the codec
L_RXQUAL_DL_H_AMR=4.5
L_RXQUAL_UL_H_AMR=4.5
L_RXQUAL_DL_P_AMR_RXACCH=(2(SFH); 3(No SFH))
L_RXQUAL_UL_P_AMR_RXACCH=(2(SFH); 3(No SFH))
A_LEV_HO(SACCH)=6
A_QUAL_HO(SACCH)=6
EN_RXQUAL_DL=TRUE
EN_RXQUAL_UL=TRUE
HO_INTRACELL=FALSE
L_RXQUALDLP=2
L_RXQUALULP=2

Best regards,

hi vanG,

In those cells, could you check the indicators related to repeated SACCH ? do you see a high amount of calls with RSACCH activated ?

In such cells, AMR calls are more protected and are less prone to interference. In other words, even if RXQUAL is bad, the voice and the SACCH are still holding. So I would consider increasing the HO QUAL thresholds (L_RXQUAL_DL_H_AMR and UL_H_AMR) to 5.5 and the PC thresholds “L_QUAL_*_AMR_RxACCH” to 3.5.

I’m surprised that your SFH PC Thresholds are higher than non-SFH. That should be the opposite.

All my advices are based on the inputs you gave me, which are extremely little. So if that doesn’t work, I can’t be held responsible 🙂

regards
pix

hi vang,

the tfo should not influence the rxqual of the call, so that’s not where we should be looking at first.

i would love to have more inputs :
– what is the increase in the ratio and numbers of HO QUAL UL & DL (separately) ?
– what is the evolution of the RTCH Drop Rate in one of those cells.
– have you tried increasing the L_RXQUAL_DL_H_AMR and L_RXQUAL_UL_H_AMR (= not the PC but the HO thresholds), I suggested that in my previous post.

Regards
pix

hi,

ok, it is a very small increase. maybe the calls are not dropping anymore, but instead they do a HO. Did you notice a decrease of TCH Drop Rate ?
In your BSC indicators table below, could you add the “Call_Drop_Rate” ? instead of the average 10 days before / after, could you do the average of the 10 busy hours of the 10 days before and after ?
Frankly, your changes are so small that it might just be a statistical effect. Have you checked those 10 days before and after ? I guess in 10 days you might have only one week-end during the “before” period and two week-ends during the “after” period. So that would definitely bias the average values… That’s just a thought 🙂

Regards
pix

vanG,

your problem with the UL Quality HO is very interesting ad would like to help you more, but unfortunately I would need to truobleshoot it by myself: do drive-tests, check NPO and run some tests on parameters.
I have no clear idea on what should be checked, there are many things that pop to my mind.
First I would do a drivetest and check that the codec mode changes are taking place, and measure the UL RxQual for that call (if you can trace it ?). Try to understand for which codec the UL RxQual degrades. In RNO, the RMS measurements about AMR could give you a good overview too. Maybe you can share those vectors & matrices here ?

RMS_AMR_UL_FR_BadSpeechFrames
RMS_AMR_UL_FR_Codec_sample 
RMS_AMR_UL_FR_RxLevel_sample 

Just drag n drop those indicators one by one to the bottom panel, for a day for which AMR FR was activated.

I’m don’t know how useful will be the results, but well…

Regards
pix