@UKTechHub
.
.

RT-AC86U 384.16 Hi,   I updated to the latest Merlin firmware this morning (about 9 am) and since doing so my NE CRC count has gone up to 326598 over 5 hours. …

UKTH forums 🛜 Wireless Routers & Modems ASUS & Wireless RT-AC86U 384.16 Hi,   I updated to the latest Merlin firmware this morning (about 9 am) and since doing so my NE CRC count has gone up to 326598 over 5 hours. ...

  • This topic has 717 replies, 4 voices, and was last updated 6 years ago by Bully.
Viewing 15 posts - 196 through 210 (of 718 total)
  • Author
    Posts
  • #6750
    Bully
    • Replies 386
    • Forum Regular

    Is the Below WAN setup correct?

     

    #6757
    UK SentinelUK Sentinel
    Keymaster
    • Replies 8,641
    • The Skipper

    Is the Below WAN setup correct?

    I have just plugged in my HG612 running Howling Wolf’s firmware, so will have a proper look at mine in the morning when house is a little quieter (Only 1 Ethernet cable plugged into Router WAN port from Modem) but settings do look correct/workable

    I was interested to see from your graphs the Inteleave depth was still high and I am sure @Mrk26 will be interested in how busy your bitswapp-ing actually is

    In a completely sane world, madness is the only freedom (J.G.Ballard).

    #6759
    AvatarMrk26
    • Replies 465
    • Forum Addict

    On mine bitswaps barely swapping, however I think I will need to stop this exercise even today, for some reason I’m getting bunch of CRCs what didn’t happened on vigor.

    #6760
    Bully
    • Replies 386
    • Forum Regular

    In the WAN settings what does the port binding do?

    #6761
    AvatarMrk26
    • Replies 465
    • Forum Addict

    Allocating port to the socket, so wan to lan 1(socket )  and lan to lan2(socket), at least how I understand this, maybe worong.

    #6762
    Bully
    • Replies 386
    • Forum Regular

    @Mrk26 – Did you see my last 2 posts on page 13, stats, FEC and bitswap graphs?

    #6763
    AvatarMrk26
    • Replies 465
    • Forum Addict

    Today’s stats

    Something must be wrong there, you getting now much more FECs 567k and on top of that you getting a lot more CRC errors. I’m not sure if first suggestion from @Uk Sentinel about noise on the line is not correct. Also wonder if you not getting higher amount of CRCs and FECs because you line sync faster now. When line was banded these errors was under control (kind of), but snr was 8db now you are on standard 6db. There is actually option to put cap on the line to make it more stable. Let’s see how stats will looks like tomorrow.

    #6765
    AvatarMrk26
    • Replies 465
    • Forum Addict

    @Mrk26 – Did you see my last 2 posts on page 13, stats, FEC and bitswap graphs?

    Yes, I see. When I compared to mime yours it’s like constantly swapping, mine was up to 8, usually 2-4 on scale plus some single spikes which line up with CRC errors (I had two bursts of CRC errors today). So looks like modem still working hard on your line. Did you setup QoS in asus router? Also could you post your QLN and Hlog graphs.

    #6766
    AvatarMrk26
    • Replies 465
    • Forum Addict

    @Bully

    Could you please change scale in bitswap per tone graph, it doesn’t show all tones, just wonder on which one is the bigger bitswap.

    #6767
    Bully
    • Replies 386
    • Forum Regular

    xdslcmd info –Hlog
    xdslcmd: ADSL driver and PHY status
    Status: Showtime
    Retrain Reason: 6
    Last initialization procedure status: 0
    Max: Upstream rate = 12136 Kbps, Downstream rate = 49604 Kbps
    Bearer: 0, Upstream rate = 12097 Kbps, Downstream rate = 39709 Kbps

    Tone number Hlog
    0 -96.0000
    1 -96.0000
    2 -96.0000
    3 -96.0000
    4 -96.0000
    5 -96.0000
    6 -96.2500
    7 -96.2500
    8 -10.3750
    9 -10.3750
    10 -10.3750
    11 -10.3750
    12 -10.3750
    13 -10.3750
    14 -10.3750
    15 -10.3750
    16 -5.3750
    17 -5.3750
    18 -5.3750
    19 -5.3750
    20 -5.3750
    21 -5.3750
    22 -5.3750
    23 -5.3750
    24 -5.0000
    25 -5.0000
    26 -5.0000
    27 -5.0000
    28 -5.0000
    29 -5.0000
    30 -5.0000
    31 -5.0000
    32 -96.0000
    33 -96.2500
    34 -96.2500
    35 -96.2500
    36 -96.2500
    37 -96.2500
    38 -96.2500
    39 -96.2500
    40 -6.6875
    41 -6.6875
    42 -6.6875
    43 -6.6875
    44 -6.6875
    45 -6.6875
    46 -6.6875
    47 -6.6875
    48 -6.3750
    49 -6.3750
    50 -6.3750
    51 -6.3750
    52 -6.3750
    53 -6.3750
    54 -6.3750
    55 -6.3750
    56 -6.3750
    57 -6.3750
    58 -6.3750
    59 -6.3750
    60 -6.3750
    61 -6.3750
    62 -6.3750
    63 -6.3750
    64 -6.6875
    65 -6.6875
    66 -6.6875
    67 -6.6875
    68 -6.6875
    69 -6.6875
    70 -6.6875
    71 -6.6875
    72 -7.1875
    73 -7.1875
    74 -7.1875
    75 -7.1875
    76 -7.1875
    77 -7.1875
    78 -7.1875
    79 -7.1875
    80 -7.6875
    81 -7.6875
    82 -7.6875
    83 -7.6875
    84 -7.6875
    85 -7.6875
    86 -7.6875
    87 -7.6875
    88 -8.0000
    89 -8.0000
    90 -8.0000
    91 -8.0000
    92 -8.0000
    93 -8.0000
    94 -8.0000
    95 -8.0000
    96 -8.1875
    97 -8.1875
    98 -8.1875
    99 -8.1875
    100 -8.1875

    xdslcmd info –QLN
    xdslcmd: ADSL driver and PHY status
    Status: Showtime
    Retrain Reason: 6
    Last initialization procedure status: 0
    Max: Upstream rate = 12136 Kbps, Downstream rate = 49604 Kbps
    Bearer: 0, Upstream rate = 12097 Kbps, Downstream rate = 39709 Kbps

    Tone number QLN
    0 -160.0000
    1 -160.0000
    2 -160.0000
    3 -160.0000
    4 -160.0000
    5 -160.0000
    6 -160.0000
    7 -160.0000
    8 -121.0000
    9 -121.0000
    10 -121.0000
    11 -121.0000
    12 -121.0000
    13 -121.0000
    14 -121.0000
    15 -121.0000
    16 -120.0000
    17 -120.0000
    18 -120.0000
    19 -120.0000
    20 -120.0000
    21 -120.0000
    22 -120.0000
    23 -120.0000
    24 -118.5000
    25 -118.5000
    26 -118.5000
    27 -118.5000
    28 -118.5000
    29 -118.5000
    30 -118.5000
    31 -118.5000
    32 -118.5000
    33 -123.1875
    34 -115.7500
    35 -116.1875
    36 -116.7500
    37 -116.6250
    38 -118.3750
    39 -117.7500
    40 -118.5625
    41 -118.3125
    42 -118.2500
    43 -119.2500
    44 -118.8750
    45 -119.6875
    46 -119.5000
    47 -120.1250
    48 -120.6875
    49 -120.5000
    50 -120.5625
    51 -120.3125
    52 -120.8125
    53 -121.1875
    54 -121.0625
    55 -121.6250
    56 -121.4375
    57 -121.0625
    58 -122.3750
    59 -121.4375
    60 -122.0625
    61 -121.3750
    62 -122.6250
    63 -122.3125
    64 -122.0000
    65 -122.5625
    66 -122.8125
    67 -122.5625
    68 -122.6875
    69 -122.2500
    70 -122.1250
    71 -122.5625
    72 -123.4375
    73 -123.1250
    74 -123.0625
    75 -123.4375
    76 -123.5000
    77 -122.8750
    78 -124.3750
    79 -123.9375
    80 -123.6875
    81 -124.0000
    82 -124.4375
    83 -124.3125
    84 -124.3750
    85 -124.4375
    86 -124.4375
    87 -124.3750
    88 -125.5000
    89 -124.9375
    90 -125.1250
    91 -124.0000
    92 -124.0625
    93 -125.5000
    94 -124.7500
    95 -125.2500
    96 -124.8125
    97 -125.8125
    98 -125.0625
    99 -125.2500
    100 -125.0000

    *EDIT* – QOS on my Asus is set to game mode.

    #6772
    AvatarMrk26
    • Replies 465
    • Forum Addict

    Hlog looks OK, maybe bit short “tail” at the end, but looking at Qln I will say that there is some issues, it’s quite spiky, and spikes usually indicates some kind of RFI or EMI noise, if its smooth bump then is crosstalk (good examples on green part).

    On mine Qln looks very similar (shape way), but without these big spikes.

    #6774
    UK SentinelUK Sentinel
    Keymaster
    • Replies 8,641
    • The Skipper

    The other option is to maybe consider disabling Bitswap in both directions as some do think CRC are made worse by having bitswap enabled ?

    This was true for the DSL-AC68U but that did use a MediaTeK DSL chipset which was hyper sensitive etc

    FYI – Bit swap could be seen as a first line of defence against brief burst type noise spikes.  Without bit swapping most lines would see a lot more errors and loss of syncs.   The tones where the most bit swap occurs should mirror the tones where SNR is lost from.  For example when I had the line fault which affected my U0 channel, thats where all my bitswap was happening, and none in the higher frequencies.  You can also use a bitloading graphs to see where noise has been problematic over say 24 hours

    Issue is underlying line fault (or compatibility with IFTN tones) and/or noise (my view) not much @bully can do regarding old firmware on cards so we are just left with tweaking modem to get the best we can

    In a completely sane world, madness is the only freedom (J.G.Ballard).

    #6776
    AvatarMrk26
    • Replies 465
    • Forum Addict

    FYI – Bit swap could be seen as a first line of defence against brief burst type noise spikes. Without bit swapping most lines would see a lot more errors and loss of syncs. The tones where the most bit swap occurs should mirror the tones where SNR is lost from.

    Agree on that, bitswap is needed. Speaking about modem tweaking, @Bully could try to cap his line, the only issue maybe that he won’t be able to remove cap, even on fastpath because as soon he remove it he will be banded again. And with higher snr he should be able to achieve better line stability, unless he will be able to get engineer to check his line.

    @UkSentinel

    Yesterday exercise wasn’t good idea, I had two bursts of CRC errors (one over 6k and one over 11k errors) today my line is interleaved on downstream (thanks God only on downstream). So looks like for me vigor is doing a very good job with FEC on downstream 😉. I swapped modems yesterday evening but obviously was too late, scored CRCs during the day judged on DLM decision.

    #6778
    UK SentinelUK Sentinel
    Keymaster
    • Replies 8,641
    • The Skipper

    Agree on that, bitswap is needed. Speaking about modem tweaking, @Bully could try to cap his line, the only issue maybe that he won’t be able to remove cap, even on fastpath because as soon he remove it he will be banded again. And with higher snr he should be able to achieve better line stability, unless he will be able to get engineer to check his line.

    Higher SNR is good for stability, alas @Bully has always had issues with synch rates and 39Mb currently is a great improvement,

    @UkSentinel Yesterday exercise wasn’t good idea, I had two bursts of CRC errors (one over 6k and one over 11k errors) today my line is interleaved on downstream (thanks God only on downstream). So looks like for me vigor is doing a very good job with FEC on downstream 😉. I swapped modems yesterday evening but obviously was too late, scored CRCs during the day judged on DLM decision.

    Vigor (Lantiq) are quality chipsets usually at the cost of slightly lower synch. I wonder what caused your CRC bursts ?

    I wonder if : (from a stability/reliability perspective)

    • Lantiq Modem chipsets work well on Broadcom exchanges/cabs
    • Lantiq Modem chipsets work well on IFTN (ECI) exchanges/cabs
    • Broadcom Modem chipsets work well on Broadcom exchanges/cabs
    • Broadcom Modem chipsets work poorly on IFTN (ECI) exchanges/cabs

    (how long is that piece of string) 

    In a completely sane world, madness is the only freedom (J.G.Ballard).

    #6780
    AvatarMrk26
    • Replies 465
    • Forum Addict

    <p style=”text-align: left;”>For me modem with lantiq chipset always sync faster than broadcom. I think that these CRC bursts may be result of overloaded network, can’t explain this any different, as this never happened before. About two weeks ago I have one day where modem working hard as I have 20k FEC errors, where in normal conditions is about 3k, since lock down 6k. Usually broadcom modem is not that bad on my line especially zyxel with built-in dsl filter… wish I could get FTTP 😔</p>

Viewing 15 posts - 196 through 210 (of 718 total)
  • The topic ‘RT-AC86U 384.16 Hi,   I updated to the latest Merlin firmware this morning (about 9 am) and since doing so my NE CRC count has gone up to 326598 over 5 hours. ...’ is closed to new replies.
Latest Posts