Generic selectors
Exact matches only
Search in title
Search in content

SFH 1*1 with ALU B10

  • This topic is empty.
Viewing 15 posts - 1 through 15 (of 31 total)
  • Author
    Posts
  • #67303 Reply
    Ian
    Guest

    Dear Experts,

    I noticed there are not so many Networks using SFH 1*1 with ALU B10, i’m looking for a feedback from Engineers who had tested this configuration, maybe we could share our findings.

    BR
    Ian

    #67304 Reply
    blackhole
    Guest

    Hi Ian,
    Our network has used this configuration for a long time and as far as I am concerned, everything is OK. So is there any prob with this?

    #67305 Reply
    pix
    Guest

    hi,

    no problem with SFH, it is applied on several networks ๐Ÿ™‚
    it’s quite tricky because you have to work hard on your coverage overlaps between adjacent cells. But this fact is not specific to ALU, it is the general idea behind SFH : frequency planning is easier, but antenna design is much more complex.

    But you still have to plan your BCCH frequency though.

    blackhole > how many BCCH freq, and how large is your MA list ? how many trx per cell, in average ?

    regards
    pix

    #67306 Reply
    Blackhole
    Guest

    Hi Pix/Ian.
    For 900: BCCH: 21 for 3 sectors (7 for each sector). MA list: 18 freqs
    For 1800: BCCH: 24 for 3 sector (8 for each). MA list: 24 freqs.

    For average, we have nearly 4 trx/cell. The largest configuration per cell may reach 12 when have 1 900 and 8 trx for 1800 (we have 2 MA list for 1800)-concentric configuration.

    Yes, you are right, planning is very easy. HSN can even be the same for nearby site without KPI affected (if the sites are not synchronize). Of course for safety we plan them different. But the prob is we don’t know what to do when we have TCH interference ๐Ÿ˜€

    #67307 Reply
    pix
    Guest

    blackhole

    thanks a lot for sharing !

    when you say that you have 4 TRX/cell, you mean 4 TRX 900 and/or 4 TRX 1800 ? Does it include the BCCH TRX ?

    regarding interference… well, you have to reduce neighbour cell coverage ๐Ÿ™‚ (but i know you are just pretending not to know the solution…)

    cheers
    pix

    #67308 Reply
    Ian
    Guest

    Hi Pix and Blackhole,

    Thank you for sharing experience.

    We have here almost everywhere 4 TRX/sector, we had used 28 hopping frequencies. BCCH is very clean with 30 BCCH.

    The bigger problem we encountered was some high HO Uplink Quality (25 to 30%) when DL_Quality is(around 15%).

    We also had to Disable DTX_DL as this is a reported Bug with ALU.

    We also activated PWCRL_DL.

    Our parameters for HO and Pwrcrtl are pretty in the standard, so i dont really understand this sudden increase.

    Thanks for your comments
    Ian

    #67309 Reply
    pix
    Guest

    Hello ian !

    what do you mean “DTX DL” is a reported bug?
    “DTX DL + AMR + SFH” is not a good combination, but “DTX DL + Legacy codecs + SFH” is not a problem. Or at least, nobody reported a bug about this… since it exists for the past 10yrs, I would assume it is stable and not faulty ๐Ÿ™‚

    Regarding your increase of UL QUAL and DL QUAL HO. It can be due to so many things…

    – 1st : SFH settings. Are your MAIO, MA List and HSN well configured within cells from same BTS and between cells from different BTS ?

    – 2nd : antenna design. If you have too much overlap between neighbour cells, the Quality will sky-rocket !

    – 3rd : HO settings. With activation of hopping, the thresholds for Qual HO are offset by Offset_Hopping_HO (def 1) and QuaL PC is offset by Offset_Hopping_PC (def 0). Ensure those parameters are at default values.

    Then you have to go drive test, to see where are the interferences, and what’s hapenning exactly.

    Regards
    pix

    #67310 Reply
    Ian
    Guest

    Hi Pix,

    First of all we are in B10ED02

    You are right, it is the DL_DTX for AMR but TAC had told us to deactivate ALL DTX_DL ( the Downlink_DTX_HR as well ), have you tested the simple DTX_DL_FR ?

    As per the codecs, which one are you using ? we are using 4 in FR (12.2;7.95;5.9;4.75)and 4 in HR(7.4;5.9;5.15;4.75), and we had found many cells using 5 codecs !? stange no ? now TAC is telling us there is a problem with the config : TFO+TwinTRX+UL_DTX..

    For the settings of SFH, for a site with 4 TRX :

    Same HSN for all sectors; for hopping TRX, we have :
    sector1 : MAIO 0,6,12
    Sectror2 : MAIO 2,8,14
    Sector3 : MAIO 4,10,16

    What do you mean by MA list ? Neigbours List ?

    2) Indeed the terrain is hilly and the city is down , some sites are well high but we manage to set up E-GSM to those, as per the antenna, i know that k742264 is not the best antenna for SFH 1*1

    3) for HO settings :

    We kept L_rxqual_UL_DL_H to 29 and set Offset_Hopping_HO to 10

    as well L_Rxqual_UL_DL_P=29 and Offset_Hopping_HO=10

    as per HO, we have
    L_rxlev_UL_H=-100
    L_rxlev_DL_H=-91;

    and for PwrCtrl :
    L_rxlev_UL_P=-87
    L_rxlev_DL_P=-80
    U_rxlev_UL_P=-77
    U_rxlev_DL_P=-70

    If you see anything wrong, do not hesitate.

    #67311 Reply
    IAN
    Guest

    Hi Pix,

    For MA List if you are talking about hopping List, we are using 29 frequencies, same list for 3 sectors, channels 35 to 62; Hopping list is 35,36,37,…62

    BR
    Ian

    #67312 Reply
    pix
    Guest

    hi ian,

    i’ve no idea where is your TAC getting their info, on my side i didn’t see anything wrong with DTX, etc.
    I will double-check again…
    Could you send me by email who is your tac contact? This guy seems to know stuff that I don’t ๐Ÿ™‚ (it drives me crazy ๐Ÿ™‚ )

    Regarding your settings, what do you think about this:

    3) for HO settings :

    We kept L_rxqual_UL_DL_H to 29 and set Offset_Hopping_HO to 10
    >
    > Default value is 39. Putting 29
    > will increase your % HO Qual !
    > Especially since your PC setting
    > is at 29 too !
    >
    as well L_Rxqual_UL_DL_P=29 and Offset_Hopping_HO=10
    >
    > OK, value is good but it should
    > be lower than your _H setting
    > above. So if your *_H is 29,
    > then your *_P should be 19.
    >

    and for PwrCtrl :

    L_rxlev_UL_P=-87
    > that’s higher than default. You
    > should get better results with
    > -90 or -93dBm

    U_rxlev_UL_P=-77
    > based on my value above, that
    > should be -80 or -83dBm

    That will decrease your UL power, therefore it will decrease your UL interference.

    — caution — apply all proposals slowly, step by step. Maybe what I’m telling will not work in your network.

    Have fun ๐Ÿ˜‰

    pix

    #67313 Reply
    Ian
    Guest

    Hi Pix,

    The problem is with DTX_DL, it has been corrected with B10 Ed07, for the previous version, they told me not to use DL_DTX, i can send you a contatc but within your personal email, dont want to advertise anything here.

    As per L_rxqual_UL_DL_H , the value i indicated in the previous post is wrong, it is indeed 39 for HO and 29 for PC.

    I’ll try your values for UL Power control but i don’t want to end up in a situation where the Mobile is going to perform some Hand_over before Power control ends. I know with some vendors, algorithm of PC is very slow to react.

    We noticed as well with our parameters that there is not much Level HO, mostly is Power Budget or Quality.

    BR
    Ian

    #67314 Reply
    Daoud
    Guest

    Hi ,

    by using the BCCH frequency in in a cell and while we have this ARFCN in our MA list what would be the most negative impact on the quality ?

    #67315 Reply
    pix
    Guest

    ian,
    ok, i’ll wait for your email, thanks !

    regarding PC versus HO :
    the UL LEV handover is triggered at -100dBm
    while the UL PC “stops” at -87dBm.

    I got good results with a delta of 6dB (UL HO = -100dBm, UL PC = -94dBm).
    You are using a delta of 13dB, it’s way too conservative (IMHO). That could very well be the main reason for high UL Qual when using SFH.
    Anyway, ajust the values by steps of 2dB only, you should see an impact very quickly (I hope it’ll be a positive impact !), without taking too much risk.
    I would recommend adjusting the DL PC as well (decrease both L-rxlev-dl-p and U-rxlev-dl-p by steps of 2dB), especially if you didn’t ativate the DL DTX (i don’t know if you are in ed7 or not)

    Cheers,
    pix

    #67316 Reply
    Ian
    Guest

    Hello Pix,

    Thanks for information, we are in B10 Ed02 and sorry for the misinformation but the DTX_DL will be corrected in B11, i had sent you the contact via personal email, keep it confidential please.

    I’m going to test your proposed values, what would be your recommandations for DL Power control ?

    And as well, what do you think of Averaging windows below :

    A_Lev_HO=6 ; W_Lev_HO=1
    A_qual_HO=6; W_qual_Ho=1
    A_pbgt_Ho=8; W_pbgt_ho=1
    A_lev_PC=4; W_lev_PC=1
    A_qual_PC=4; W_qual_PC=1
    A_qual_ca_hr_fr=4
    A_qual_ca_FR_HR=61

    BR
    Ian

    #67317 Reply
    pix
    Guest

    ian,

    your avg windows look ok, i think you can tune them step by step if you want, but the effects are so important that I would only change one value at a time…

    Regarding PC, try this

    L_rxlev_UL_P=-90
    U_rxlev_UL_P=-80

    and then, 2nd step:

    L_rxlev_UL_P=-92
    U_rxlev_UL_P=-82

    and then, 3rd step (for DL PC):

    L_rxlev_DL_P=-82
    U_rxlev_DL_P=-72

    if things are looking better then let’s work on ‘upper’ thresholds:
    U_rxlev_UL_P=-84
    U_rxlev_DL_P=-74

    (the effect is to decrease the “target” rxlev, which is the avg btw upper and lower thresholds, without getting to close to HO thresholds)

    If still ok, then decrease symmetrically:
    L_rxlev_UL_P=-94
    U_rxlev_UL_P=-86

    L_rxlev_DL_P=-84
    U_rxlev_DL_P=-76

    etc.

    That would probably the lowest you can go though, but you can try to go further down ๐Ÿ™‚

    regards
    pix

Viewing 15 posts - 1 through 15 (of 31 total)
Reply To: Reply #67311 in SFH 1*1 with ALU B10
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="">