Kitz Forum

Broadband Related => ADSL Issues => Topic started by: Weaver on November 23, 2021, 12:09:13 PM

Title: Multiple lines down/up
Post by: Weaver on November 23, 2021, 12:09:13 PM
All three lines flapped this morning, starting at exactly 9:00:00 to the second, up down many times, back up again now but with two lines very sick, below FTR, 1 Mbps lower (downstream) than they should be. The time and the synchronisation are both very very suspicious. Talked to AA on IRC, they’re investigating so I’m leaving them to it.

What does that mean?


(https://i.postimg.cc/TPtVhGgN/2-EEF0-AA3-9-CDA-48-D3-9-D25-D9444-BDC328-F.jpg)
Title: Re: Multiple lines down/up
Post by: Weaver on November 28, 2021, 07:03:42 AM
Have switched modems 1 and four off and then back on following AA instructions. Modems 1 and 4 are now extremely fast and healthy 2.9 / 0.673 Mbps for line 1 and 3.173 / 0.396 Mbps for line 4. Notice how outstandingly good the line 1 upstream is and how, as ever, line 4 upstream is weak. Line 4’s amazingly high 3.173 Mbps downstream is still reliable: zero recent errors (in last 25 mins) in either direction.

Line 2 is slow downstream but reliable. Downstream it is just under the FTR (2.278 Mbps) at 2.158 Mbps @ 3.7dB d/s so just around 700kbps slow; upstream is normal : 0.582 Mbps @ 5.7 dB u/s. The line 2 bits-vs-tones curve is really horrible, that’s why it’s so slow downstream. Just the usual story then.

Really must get around to asking my wife to swap out modem 2 today as a sanity check.
Title: Re: Multiple lines down/up
Post by: burakkucat on November 28, 2021, 04:34:48 PM
I am beginning to think that if you can automate the power-cycling of the modems, all of the recurring xDSL problems may be dissipated.

Something like: modem-A on a Monday, modem-B on a Wednesday and modem-C on a Friday . . .  :-\
Title: Re: Multiple lines down/up
Post by: g3uiss on November 28, 2021, 06:32:11 PM
Perhaps some smart plugs that can be programmed or trigger from you iPad ?
Title: Re: Multiple lines down/up
Post by: Weaver on November 28, 2021, 07:25:13 PM
An extremely interesting thought. abut before we do that, I should swap out all the modems; I have plenty of spares. Will pre-label the old and the new ones as "2018-202" and "2021/2022-present".

There’s a snag. Janet cannot get into the office. While I was fast asleep this afternoon, a super king size mattress was delivered, and was hauled upstairs to the office by my kindly neighbours. It’s for our bed. Ever since her surgery, Janet has been sleeping across the way, in the guest bedroom, for fear of being trodden on by cats. This week my neighbours will shift it into my bedroom, while I’m in the spare room. In the meantime, it’s parked in the office, silently expanding, and Janet literally cannot get into the room, not without hurting herself. So all modem operations will be postponed for a few days.
Title: Re: Multiple lines down/up
Post by: burakkucat on November 28, 2021, 08:44:03 PM
There’s a snag. Janet cannot get into the office.
 . . .
So all modem operations will be postponed for a few days.

Understood.  :)
Title: Re: Multiple lines down/up
Post by: Weaver on November 29, 2021, 04:49:26 AM
Speculation: In Jan/Feb 2020, whenever it was, there was an atmospheric electrical event which slightly cooked all four of my modems that were in use. This caused a peculiar sporadic front end fault whereby each modem reports ‘hollow curve’ bits-per-bins disease every so often. The disease gets worse and then much much worse over a time period of a week or two. Either unplugging the line or power-cycling the modem fixes the bug (unclear) by resetting the modem’s state. End of speculation.

That is just a story at the moment. If that fairy story is true, then the usual method of swapping out modems would possibly not work as all modems are slightly cooked, possibly to an equal extent. I would have to put in a totally fresh modem for this to be a useful test. It’s interesting to speculate though; an event that causes mild damage to all connected devices equally, is that asking way too much?

Speculation 2: If we unplug a modem from the wall are we doing something beneficial to the DSLAM ? In that case, have we discovered a DSLAM fault rather than (or as well as) damage to my modems ? When we unplug a line, it occurs to me that we influence two systems not just one, one being the way we normally think about it. Is it feasible that just unplugging a line and plugging it back in can fix a problem in the CPE (be it either temporarily or long-term)?

How could AA check out a potential DSLAM-cooking fault? Doesn’t sound very likely to me I have to say, but if we first prove the modems are good and the line tests ok, then there’s nothing left, is there? DSLAMs will be very well protected; that’s why I don’t think speculation 2 is likely.
Title: Re: Multiple lines down/up
Post by: jelv on November 29, 2021, 11:01:31 AM
Speculation 3: There's a logic issue with the DSLAM SW/FW that produces this phenomenon in the very rare/extreme lines conditions such as yours.
Title: Re: Multiple lines down/up
Post by: burakkucat on November 29, 2021, 03:45:08 PM
How could AA check out a potential DSLAM-cooking fault?

I don't think they can. At best they might be able to arrange for a port-swap, for each circuit, by using the logic that the DSLAM line-card has some zapped ports.  :-\
Title: Re: Multiple lines down/up
Post by: Weaver on November 30, 2021, 04:37:50 AM
Jelv, the problems started after the lightning strike and were unknown before it - had never ever seen anything like it before. Make sense? What do you think?

Thanks to Jelv and to anyone else who wants to contribute brainpower as that’s sorely needed right now, will take as much additional GFLOPS as I can get.
Title: Re: Multiple lines down/up
Post by: kitz on December 01, 2021, 09:08:16 PM
Quote
each modem reports ‘hollow curve’ bits-per-bins disease every so often. The disease gets worse and then much much worse over a time period of a week or two. Either unplugging the line or power-cycling the modem fixes the bug (unclear) by resetting the modem’s state. End of speculation.

Not well and not entirely thinking 100% straight due to fairly high doses of morphine + pregabalin....  but my 2p is that hollow curves can occur and build up over a period of time purely down to bitswap.  Rebooting the modem will automatically generate a new BAT with subsequently higher bit rate and thus sync speed.     If you have a line that has a range of tones operating with little margin, then it's quite possible for bitswap to eventually cause these tones to be unviable and cause a decrease in the max attainable sync.    Forcing a power cycle causes a new clean bit allocation table to be generated.   

Even some exceedingly good lines can experience this type of "seepage" where it displays by causing the SNRm to slowly decrease over a period of time.  Rebooting the modem should rectify... until the next time.     It wouldn't harm in cases such as this to routinely power cycle the modem every week or so, to ensure you have a fresh clean BAT.

[Moderator edited to fix a typo.] 
Title: Re: Multiple lines down/up
Post by: Weaver on December 02, 2021, 02:02:01 AM
@Kitz: So do I need monitored tones then if I am using bitswap ? This is a VMG 1312-B10A. The bits per bin in tones 40-90 have a minimum at tone 60 approx of 2 bits instead of the max of 11! My SNR doesn’t go down over time, wouldn’t that be a necessary phenomenon? Is my modem very similar to your own?

 I do think that this was a known phenomenon with the DLink DSL 320B-Z1 and I’m assuming that that didn’t have monitored tones. Is that correct? Cheap was it MediaTek chipset?

— And Pregabalin and morphine - welcome to my world too. Plus fentanyl patches in my case.



Update, same OR engineer came out, tested line and hey presto, back to 2.97 Mbps downstream @ 3dB
Title: Re: Multiple lines down/up
Post by: Weaver on December 02, 2021, 05:08:08 AM
Questions for Kitz and Burakkucat then:

1. does a ZyXEL VMG 1312-B10A have the good kind of bitswap technology whereby it won’t just run out of bits over time ?
2. Will that model be ok because it has monitored tones ? (if indeed it does have monitored tones )
3. Should one turn off bitswap if there is no monitored tones facility?

I used to moan about Andrews and Arnold not enabling bitswap, in my ignorance and arrogance. That was the default setting for a DLink DSL-320B-Z1 and I just assumed that AA had given it no fault, had left it at the defaults, and that it was a bug on the part of DLink and AA.

Coming back to the ZyXEL modems, if we assume temporarily this fault is a natural phenomenon due to the design of the VMG 1312-B10A then you would expect that the fault would be reported before Jan 2020. I should look back in old threads but as I recall, the ‘hollow curve’ fault was something new to me and first began after Jan 2020 so leading me to think it’s lightning storm related. Is that correct ?
Title: Re: Multiple lines down/up
Post by: Chrysalis on December 02, 2021, 09:51:14 AM
I wasnt requested to reply but will, am sorry if you dont appreciate my reply.

Firstly I do not suggest turning off bitswapping, bitswapping on noisy lines is very effective, without it, if a snr for a tone goes down, the bits wont be adjusted, you could have negative snr for that tone, and either high errors or loss of sync.  a line as long as yours to have no variance I think would be unusual.

In terms of what you call the running out of bits, at first I thought you meant does bitswapping eventually stop functioning (I had these problems in the adsl max days when bitswapping was new and was buggy on various firmwares, but never recently), I actually however think you mean where it turns off specific bits/tones?  Bitswapping I believe will disable a tone if the bit loading gets too low, in my experience if it ever hits 0, then the tone gets turned off for the remainder of the sync uptime, in my opinion this is a good thing.  But the consequence is a lower SNRM for the duration of the sync.  (or lower sync if using SRA). ADSL2 has an advantage here over ADSL1 as ADSL1 has a min bitloading of 2, so would go from 2>0 whilst ADSL2 can support a bitloading of 1, so a tone can survive with 1 bit.

I would expect all Zyxel modems to have behaviours that match up with the Broadcom chip and firmware version.

I didnt know what you meant by monitored tones, but found your older post from a search.

https://forum.kitz.co.uk/index.php?topic=19326.0

I have never used a modem that doesnt do that, as far as I know thats expected behaviour within the DSL spec.  However I cant verify on any of the modern zyxel modems as my VDSL line doesnt have any disabled tones, the only one's that get low are in the power cut back range due to openreach applying extremely aggressive power cutback.  They dont get disabled, but doesnt mean thats due to lack of monitoring it could just be they never get weak enough.  of course also i am running VDSL instead of ADSL so the behaviour may have changed anyway.  But i can tell you from my ADSL days when I had a very poor line, every modem I used would disable tones when they hit 0 bitloading.

I noticed ejs replied indicating he thinks the behaviour changed on newer modems vs old ones. 

According to ejs you can use this command to check if the feature is enabled.

Code: [Select]
xdslcmd info –cfg
but on my zyxel its actually

Code: [Select]
xdslctl info –cfg
There is no mention of tone monitoring in the output however, there is an ability to turn it on and off.

Code: [Select]
xdslctl configure/configure1 [--mod <a|d|l|t|2|p|e|m|M3|M5|v>] [--lpair <(i)nner|(o)uter>]
           [--trellis <on|off>] [--snr <snrQ4>] [--bitswap <on|off>] [--sesdrop <on|off>]
           [--sra <on|off>] [--CoMinMgn <on|off>] [--minINP <sym>] [--maxDelay <ms>] [--i24k <on|off>] [--phyReXmt <0xBitMap-UsDs>]
           [--Ginp <0xBitMap-UsDs>] [--TpsTc <0xBitMap-AvPvAaPa>] [[b]--monitorTone <on|off[/b]>]

I be very surprised if this is not available on your zyxel.
Title: Re: Multiple lines down/up
Post by: Weaver on December 02, 2021, 11:38:36 AM
Thanks to Chrys for contributing, very much welcomed. I didn’t know about that command to set the monitored tones functioning status.
Title: Re: Multiple lines down/up
Post by: Chrysalis on December 02, 2021, 12:36:42 PM
I will set aside my comments on the older thread, my experience on ADSL is outdated as back then this feature didn't exist, hopefully it works in a positive way for you. :)
Title: Re: Multiple lines down/up
Post by: burakkucat on December 02, 2021, 05:11:02 PM
The following is from a ZyXEL VMG1312-B10A, used on a VDSL2 service which, in turn, is provisioned on the Openreach 40/10 Mbps product.

The "help" output is (obviously) xDSL service type agnostic --

Code: [Select]
> xdslctl --help
Usage: xdslctl start [--up] <configure command options>
       xdslctl stop
       xdslctl connection [--up] [--down] [--loopback] [--reverb]
   [--medley] [--noretrain] [--L3] [--diagmode] [--L0]
   [--tones <r1-r2,r3-r4,...>] [--normal] [--freezeReverb] [--freezeMedley]
       xdslctl configure/configure1 [--mod <a|d|l|t|2|p|e|m|M3|M5|v>] [--lpair <(i)nner|(o)uter>]
   [--trellis <on|off>] [--snr <snrQ4>] [--bitswap <on|off>] [--sesdrop <on|off>]
   [--sra <on|off>] [--CoMinMgn <on|off>] [--i24k <on|off>] [--phyReXmt <0xBitMap-UsDs>]
   [--Ginp <0xBitMap-UsDs>] [--TpsTc <0xBitMap-AvPvAaPa>] [--monitorTone <on|off>]
   [--profile <0x00 - 0xFF>|<"8a |8b |8c |8d |12a |12b |17a |30a">] [--us0 <on|off>]
   [--dynamicD <on|off>] [--dynamicF <on|off>] [--SOS <on|off>] [--maxDataRate <maxDsDataRateKbps maxUsDataRateKbps maxAggrDataRateKbps>]
   [--forceJ43 <on|off>] [--toggleJ43B43 <on|off>]
       xdslctl bert [--start <#seconds>] [--stop] [--show]
       xdslctl afelb [--time <sec>] [--tones] [--signal <1/2/8>]
       xdslctl qlnmntr [--time <sec>] [--freq <msec>]
       xdslctl inm [--start <INMIATO> <INMIATS><INMCC><INM_INPEQ_MODE><INM_INPEQ_FORMAT>] [--show]
       xdslctl snrclamp [--shape <shapeId>] [--bpshape [bpIndex-bpLevel,]]
       xdslctl nlnm [--show ] [--setThld <Thld_Num_Tones>]
       xdslctl diag [--logstart <nBytes>] [--logpause] [--logstop] [--loguntilbufferfull <nBytes>] [--loguntilretrain <nBytes>] [--dumpBuf <sizeKb>]
       xdslctl ntr [--start [output freq(default is 8000)]] [--stop]
       xdslctl info [--state] [--show] [--stats] [--SNR] [--QLN] [--Hlog] [--Hlin] [--HlinS] [--Bits]
    [--pbParams] [--linediag] [--linediag1] [--reset] [--vendor] [--cfg] [--webstats]
       xdslctl profile [--show] [--save] [--restore]
       xdslctl --version
       xdslctl --help
 >

This output is for a VDSL2 service --

Code: [Select]
> xdslctl info --cfg
xdslctl: ADSL driver and PHY status
Status: Showtime
Last Retrain Reason:    0
Last initialization procedure status:   0
Max:    Upstream rate = 18534 Kbps, Downstream rate = 44755 Kbps
Bearer: 0, Upstream rate = 10000 Kbps, Downstream rate = 40000 Kbps
Bearer: 1, Upstream rate = 0 Kbps, Downstream rate = 0 Kbps

adslAnnexAParam:        01d04980
adslTrainingMarginQ4:   -1
adslShowtimeMarginQ4:   -1
adslLOMTimeThldSec:     -1
adslDemodCapMask:       0090447a
adslDemodCapValue:      0090447a
adsl2Param:             00000000
adslPwmSyncClockFreq:   0
adslHsModeSwitchTime:   0
adslDemodCap2Mask:      00d00000
adslDemodCap2Value:     00d00000
vdslParam:              007f0040
vdslParam1:             00000000
xdslAuxFeaturesMask:    00000003
xdslAuxFeaturesValue:   00000003
vdslCfgFlagsMask:       00000000
vdslCfgFlagsValue:      00000000
xdslCfg1Mask:   00000000
xdslCfg1Value:  00000000
xdslCfg2Mask:   00000000
xdslCfg2Value:  00000000
xdslCfg3Mask:   00000000
xdslCfg3Value:  00000000
xdslCfg4Mask:   00000000
xdslCfg4Value:  00000000
maxDsDataRateKbps:      0
maxUsDataRateKbps:      0
maxAggrDataRateKbps:    0
xdslMiscCfgParam:       00000000
AFE_ID:                 10608700 00000000
 >
Title: Re: Multiple lines down/up
Post by: kitz on December 02, 2021, 09:15:19 PM
I think chrys has already covered most of it.  Definitely do not turn off bitswap... but that doesnt negate the fact when operating at low SNR most lines will experience some seepage over a period of time. Can't be helped if tones get disabled.   More important is that bits are shifted to areas where there the SNR is higher...  and that the process can also increase the power (gain).

>>> Coming back to the ZyXEL modems, if we assume temporarily this fault is a natural phenomenon due to the design of the VMG 1312-B10A   

As regards to that particular modem, I didnt have it on mine on the line long enough to notice anything untoward, but based on the chipset performance, then you could do far worse. I see hardly any seepage using my VMG8324, yet on a tuplink modem I was testing years ago,  it visibly deteriorated on a daily basis.

>> the ‘hollow curve’ fault

If the hollow curve effect reflects in QLN graph, then it could possibly be crosstalk at the MSAN.  It may not be an actual fault but just the effects of a new subscriber(s) getting adsl and being put on the same line card as you.
I mentioned it because unfortunately it may be circumstances that you can do little about rather than an actual fixable fault :/