Generic selectors
Exact matches only
Search in title
Search in content
Search in posts
Search in pages

To Pix PLZ your reply

  • This topic has 11 replies, 1 voice, and was last updated 7 years ago by Manmayur.
Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • #69198 Reply
    GPRS
    Guest

    Hi Pix,
    Hi everyone,

    I’m GPRS optimization enginner and a new member in this forum.

    Well, I am facing a major problem, i have a BSC suffering from high UL_TBF_Drop_Rate (around 10%) cause T3169 Timer expiry (around 96%), and high DL_TBF_Drop_Rate (around 15%) cause T3195 Timer expiry (around 95%).

    Please advice me how to proceed and what are different parameters i must check and change to fix this problem

    your reply is appreciated

    #69199 Reply
    pix
    Guest

    hi,

    Could you please post the definition of those 2 timers ?
    10% is not so high, so it might actually be a radio issue. Reduce your initial coding scheme, enable link adaptation, and check the (speech) HO QUALITY % on the same frequency as the one carrying the GPRS.

    Regards,
    pix

    #69200 Reply
    Rex
    Guest

    Hi GPRS,
    What is your SDCCH DROP? If it is high and GPRS is configured on BCCH TRX than you should check BCCH planning. How many cells are suffering on that BSC? What is your HO Uplink quality rate?
    Best Regards,
    Rex

    #69201 Reply
    GPRS
    Guest

    Hi Pix,
    Hi Rex,

    Thanks for replying me,

    Timer definition:

    T3195 : Linked to TBF and TFI, it starts when N3105 reaches to maximum value, and when it expired the wetwork releases TFI resource

    T3169: Response to USF tolerance timer, On expiring, the USF is released (which means it can be reassigned to another mobile station)

    But it’s not really clear for me this definition

    FYI:
    – NW is Siemens
    – All freq of the cell is carrying GPRS
    – Avg of 10 % of TBF drop rate at Cell level (apprx all cells)
    – HO_Quality % at BSC level is around 7%

    PLZ ur comment

    #69202 Reply
    pix
    Guest

    hi,

    ok, sounds like real radio issue on GPRS. You should focus on decreasing the Coding Scheme (CS), in order to make it more robust.

    1/ CS-INIT = CS-1 rather than CS-2
    2/ LINK ADAPTATION (LA) = ON
    3/ Is your EGPRS activated (or GPRS only) ?
    4/ Experiment increasing OR decreasing the PDCH INACTIVITY TIMERS. I’m not sure which one will actually decrease your TBF DROP% the most.
    5/Ensure you allow 4 PDCH or 5 PDCH per TBF
    6/ Maybe decrease the MAXIMUM CODING SCHEME (MAX_CS) to CS-3 rather than CS-4
    7/ Prefer the TBF on BCCH FREQ (you should find such settings somewhere, which will favour tbf establishment on BCCH freq until it’s full, then the next estab will be on non-BCCH freq)

    That’s about it, on top of my head.

    Please let me know how it goes. Don’t forget to tell me if EGPRS (EDGE) is activated or not (MCS-1 to MCS-9)

    Regards
    pix

    #69203 Reply
    Manmayur
    Guest

    Hi to all ,
    What exactly is LAPm connection that mobile establishes with BSC during call setup ?
    Will anyone explain please ?
    With thanks in advance..

    #69204 Reply
    pix
    Guest

    hi mike,

    LapDm is the Layer-2 of the Air interface. LapDm is used to carry signalling channels towards the right “ports” in the BTS.
    It’s all very transparent to the user, and to us, telecom engineers.
    LapDm can trigger timers based on T200 timers, and that can be used to detect dropped calls.

    #69205 Reply
    Manmayur
    Guest

    Hi Pix ,
    Thank you so much for the answer . It was like some sort of an eye opener for me when you did say “port” .However I am a starter hence did not get when you said “It’s all very transparent to the user, and to us, telecom engineers.” and
    Will you explain something about T200 timer and the way we can monitor the timer to detect dropped calls please ..
    With thanks and regards ..

    #69206 Reply
    pix
    Guest

    Hi,

    It’s transparent because it’s just a “transport protocol”, meaning there’s not much room for improvement,IMO.
    Some ppl here have been experimenting with longer T200 duration, and it seems to improve drop rate.

    There is a mecanism in the LapDm : if the SACCH is not decoded succesfully in one occurence, then the timer T200 is started. At expiry, if SACCH are still not decoded, then a counter “n” is incremented by 1.
    When n reaches the value set by N200, then the lapDm consider the session as “terminated”, which is counted as a “call drop cause radio” by the BSC.
    This mechanism works independantly in UL and in DL, as far as I remember.

    It works in parallel to the “typical” Radio Link Failure mechanism.

    Regards,
    pix

    #69207 Reply
    GPRS
    Guest

    Hi Pix,
    Hi experts,
    I was on vacation

    Well, for the cells suffering from high DL_TBF_drop (%), i changed below parameters as follows:

    1/ CS-INIT = CS-2
    2/ IMCSUL8PSK = MCS2 (Initial MCS uplink 8PSK)
    3/ IMCSULGMSK = MCS2 (Initial MCS uplink redundancy GMSK)
    3/ INIMCSDL = MCS3(Initial MCS downlink)
    4/ LINK ADAPTATION (LA) = TRUE
    5/ Both EGPRS & GPRS are activated
    6/ Uplink Incremental Redundancy = TRUE

    Unfortunately, all these changes did’nt bring any results, drop rate still the same, knowing that all cells frequency are clean (no interference)

    PLZ, ur advice
    thx

    #69208 Reply
    pix
    Guest

    hi,

    welcome back:)

    1/are the drops occuring in GPRS or in EGDGE ?

    2/ are they occuring at any time of the day (proportionally to traffic), or are they especially located around the busiest hours of the day?

    3/ reduce the maximum MCS possible from MCS9 to MCS7 *and* from CS4 to CS3.

    4/ try pushing the GPRS allocation to another TRX (right now, either your GPRS is going on BCCH trx or it is not.. whatever you have chosen until now –> do the opposite)

    You may also suspect a BSC issue ? In such case, you might consider doing a BSC “gprs module” restart during the night. Have you already contacted your technical support, check if there isn’t any bug in the software release you are using ?

    If nothing worked, then it’s gonna be complicated to help you remotely. There are several things one can do, but that requires heavy analysis.

    #69209 Reply
    Manmayur
    Guest

    Thanks so much for your answer Mr. Pix..
    I am a lot clear now..to be fully aware ,will have to study a lot.
    With thanks again.

Viewing 12 posts - 1 through 12 (of 12 total)
Reply To: To Pix PLZ your reply
Your information:




<a href="" title="" rel="" target=""> <blockquote cite=""> <code> <pre class=""> <em> <strong> <del datetime="" cite=""> <ins datetime="" cite=""> <ul> <ol start=""> <li> <img src="" border="" alt="" height="" width="">