- This topic is empty.
14th February 2007 at 08:38 #46504TomarGuest
I am not sure, that I understood what your problem is. You were written that in yours network is a lot of HO due to Uplink. From your settings: HO THRESHOLDS LEV UL RX LEVEL ………… -095 dBm. We are using lower level. From statistic, measurement and calculation path loss is UL much more critical for communication in GSM. My suggestion is to set HO Threshold for UL lower than DL (5dB).
We have same handover priority according HO criteria. My note to increasing number HO due to Quality is from my experiences.
Tomar14th February 2007 at 18:12 #46505pixGuest
could you write down some QoS Stats about your cells ? (only relevant qos stats, not all indicators)
if you’re having UL HO on one cell (not on all the network), you might want to check your BTS hardware (TRX faulty, probably)16th February 2007 at 15:05 #46506ValerijsGuest
first of all, my suppose below was about this case :2.ALSO WE HAVE A PROBLEM IN THE BSC BORDERS.THE STATS SHOWS THE 100% FAILURES EVEN THOUGH U DIFINE IN BOTHWAYS.
concerning priority. With the parameters you can tune the therhold where the ho case can happen.But priority as it is only used when more then 1 ho case are true. For exm. if you have bad (higher then threshould) ul quality and bad (lower then threshold) uplink signal the BSC will use parameters from UL quality algorithm.Usualy you have lower marging for quality then for level, and that is all. If you think that you have to much ho because UL level, just try to play with threshold.(check about ping-pong)But remember ho on quality – subscribers already experienced a quality problem, on level no.20th February 2007 at 09:14 #46507PromodGuest
2.ALSO WE HAVE A PROBLEM IN THE BSC BORDERS.THE STATS SHOWS THE 100% FAILURES EVEN THOUGH U DIFINE IN BOTHWAYS.
As i was having problem with INTER MSC and INTER BSC problem has been resolved after difining the LAC in the MSC level. So i thank for everybody giving the solution.
I would like to know what changes to be done in the below mentioned power control parameters due to the level HO problem.
POWER CTRL ENABLED ……… Y POWER CONTROL INTERVAL ….. 02 s
BS TX PWR MIN ….PEAK PWR – 30 dB BS TX PWR MAX ….PEAK PWR – 00 dB
BS TX PWR MAX1X00 PEAK PWR – 00 dB
POWER INCR STEP SIZE ……. 4 dB POWER RED STEP SIZE …….. 2 dB
POWER DECR LIMIT BAND 0 …. 38 dB POWER DECR LIMIT BAND 1 …. 38 dB
POWER DECR LIMIT BAND 2 …. 38 dB POWER DECR QUAL FACTOR ….. 1
ALA ENABLED ……………. N MIN INT BETWEEN ALA …….. 10 s
POWER LIMIT ALA ………… 6 dB
PC AVERAGING LEV DL WINDOW SIZE ………… 04 WEIGHTING …. 1
PC AVERAGING LEV UL WINDOW SIZE ………… 04 WEIGHTING …. 1
PC AVERAGING QUAL DL WINDOW SIZE ………… 01 WEIGHTING …. 1
PC AVERAGING QUAL UL WINDOW SIZE ………… 01 WEIGHTING …. 1
PC UPPER THRESHOLDS QUAL DL RX QUAL .. < 0.2% PX .. 16 NX .. 16 PC UPPER THRESHOLDS QUAL UL RX QUAL .. < 0.2% PX .. 16 NX .. 16 PC LOWER THRESHOLDS QUAL DL RX QUAL .. 0.8% - 1.6% PX .. 03 NX .. 04 PC LOWER THRESHOLDS QUAL UL RX QUAL .. 0.8% - 1.6% PX .. 03 NX .. 04 PC LOWER THRESHOLDS QUAL144 RX QUAL .. 0.8% - 1.6% PX .. 03 NX .. 04 PC UPPER THRESHOLDS LEV DL RX LEVEL .. - 68 dBm PX .. 01 NX .. 01 PC UPPER THRESHOLDS LEV UL RX LEVEL .. - 72 dBm PX .. 01 NX .. 01 PC LOWER THRESHOLDS LEV DL RX LEVEL .. - 80 dBm PX .. 01 NX .. 01 PC LOWER THRESHOLDS LEV UL RX LEVEL .. - 84 dBm PX .. 01 NX .. 01 PC UPPER THRESHOLDS DL RX QUAL AMR HR .. < 0.2% PC UPPER THRESHOLDS UL RX QUAL AMR HR .. < 0.2% PC LOWER THRESHOLDS DL RX QUAL AMR HR .. 0.8% - 1.6% PC LOWER THRESHOLDS UL RX QUAL AMR HR .. 0.8% - 1.6% Regards, Promod20th February 2007 at 12:36 #46508pixGuest
PC AVERAGING QUAL DL WINDOW SIZE ………… 02
PC AVERAGING QUAL UL WINDOW SIZE ………… 02
but that changes won’t affect your HO due to low level.
your settings for PC are fine otherwise, i don’t understand why you’re mixing HO level and PC in your case. Actually i don’t really understand your problem, because you still didn’t provide us with you QUALIY OF SERVICE STATISTICS / INDICATORS.23rd February 2007 at 10:09 #46509PromodGuest
Please find the below mentioned QOS report for your reference.
Date Segment ID 2G Network Availability / SDCCH Availability (ava_4) Availability / TCH Availability (ava_1d) Blocking / TCH Blocking (blck_8d) CSR Call Congestion / SDCCH Blocking (blck_5a) DCR Intensity / Average TCH Traffic, erlang (trf_12a) Successful Ratio / Incoming HO Success (hsr_18) Successful Ratio / Outgoing HO Success (hsr_19) TCH Sucess Assignment
4-Jan-07 58 B24_16231 100.00 100.00 0.00 99.44 949 0.00 1.42 0.46 63.80 90.74 97.59
5-Jan-07 58 B24_16231 100.00 100.00 0.00 99.40 1041 0.00 1.82 0.49 59.79 85.64 98.41
7-Jan-07 58 B24_16231 100 99.99836 0 99.54709 1269 0 1.667825 0.798752 59.72696 89.5 98.02905
8-Jan-07 58 B24_16231 100 100 0 99.54445 1061 0 1.982554 0.576971 64.51613 88.74459 97.86382
9-Jan-07 58 B24_16231 100 100 0 99.7032 1197 0 1.212553 0.777284 66.66667 88.83929 99.09502
10-Jan-07 58 B24_16231 100 100 0 99.45479 1300 0 1.51715 0.612849 60.10782 76.84887 98.91186
11-Jan-07 58 B24_16231 100 100 0 98.41598 1373 0 1.344743 0.959566 60.3211 91.41914 97.6394823rd February 2007 at 13:38 #46510promodGuest
In the below mentioned case i have problem with HO’s. I have been defined all the the neighbours in proper way.Still the problem is persisting to me. So that is why i provided MML report of HO parameter and PC parameters to for the reference.
In these case there are lot of UPLINK HO instead of PBGT HO.So i required ur favour to resolve this ASAP to meet the KPI.
Promod.24th February 2007 at 16:49 #46511pixGuest
Thanks for the qos report.
Unfortunately i didn’t manage to match the headers (indicators names) with the columns (indicators values).
Ok, what you could try to do :
1) HO THRESHOLDS LEV UL RX LEVEL ………… -102 dBm
2a) PC UPPER THRESHOLDS LEV UL RX LEVEL .. – 80 dBm PX .. 01 NX .. 01
2b) PC LOWER THRESHOLDS LEV UL RX LEVEL .. – 90 dBm PX .. 01 NX .. 01
1) Your UL threshold is way too high, sorry i didn’t notice that earlier. Those thresholds will force to trigger a UL HO eventhough the level is still within an acceptable range.
2a & 2b ) Those settings I’m proposing are not mandatory, but currently the power control is forcing the mobiles to emit with a strong power : it decreases the battery life (= less potential traffic !!) and it is always better to decrease the MS power (potential health concerns…)
Check out the call drop rate after changes, and in case of higher CDR, increase those thresholds by 2dB step increments.
In my opinion, you might also be in an area which is not covered well enough, which is the reason why there is no “power budget” handover, but only handovers based on level thresholds
Pix26th February 2007 at 13:09 #46512PromodGuest
I am going to implement ur idea as u mentioned in the below discussion and i will moniter for the one week.
So i have some more doubts to be cleared properly as i have little bit confusion.
1.Define the window size?
2.What is the Nx. and Px.?
3.What is UPPER and LOWER thresholds.
I need the detail explaination for the above mentioned queries in the short form?
Pls do the needful to me………
Promod26th February 2007 at 21:37 #46513pixGuest
1- the window size is the number of measures averaged in order to get a “smoothed-out” value. This “averaged” value is compared to the threshold.
For instance, window size for PC LEVEL = 4 means :
RX LEV (t=1) = -60
RX LEV (t=2) = -62
RX LEV (t=3) = -61
RX LEV (t=4) = -63
At t=4, the AV-RX-LEV = -61.5 dBm. This value will be cmpared to the threshold UPPER PC LEV and LOWER PC LEV.
2- i learnt about Px and Nx long time ago, but since they are not modifiable, i forgot about them. I didn’t ask you to change these values though… just keep the current values.
if av_rx_lev > upper pc lev threshold, then the Tx Power of the MS will be increased.
if av_rx_lev < lower pc lev threshold, then the Tx Power of the MS will be decreased. i hope it helps.. i suggest you have a look through 3GPP specifications regarding GSM algorithms if you need additional details.28th February 2007 at 07:49 #46514PromodGuest
Thnks for the information..
So i would like know same example of PC for the uplink and downlink level.Pls do the needful..
I have checked the Today’s DCR count of the network after implementing your idea, it got reduced to 4-5%. It is very nice
to see today improvement in the network.
Promod.1st March 2007 at 06:50 #46515PromodGuest
Yes i got the some more improvement in the network on DCR after implementing the your parameters.So thanks for your input sir.
1.I would like know how the HO margin of LMRG,PBGT and QMRG works in the network?
2.You have given the ex: of PC level,same like this i would like to know the window side size uplink level and downlink level works?
3.I would like to know about RLT?
Promod1st March 2007 at 19:45 #46516ParaHOGuest
Is rx lev -100dBm too high? If server and neighs all below threhold -100dbm where MS indoors.1st March 2007 at 20:37 #46517PixGuest
yes, if a mobile receives a signal (rxlev) of less than -100dBm, it is way too low. Most of the time it will lead to a dropped call.
i’m glad it works better now.
i think it is now time for you to learn about parameters and radio optimization 🙂 i can’t explain you everything via internet.. you’d rather ask your vendor some documents about radio optimization, radio parameters and handover procedures.
good luck !2nd March 2007 at 06:11 #46518PromodGuest
I am going through the documents of Parameters and optimisation so still i have little bit confusion on below mentioned parameters.Due to this problem i want to clear the doubts from your side, as your expert in parameters and optimisation.
Sir i am not asking you to briefly explain, you can explai in short form also. So pls do the needful….
I am very keen in your explaination…