Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[waltergmw]

@ Les,

1.  Re the 301c, I'm interested that your trace shows a significant blip at the point you suspect is the PCP.
All the ones I've  observed show very little change in the trace.
Have you confirmed your FTTC position by seeing the "Beacon" jumping up and down at that distance ?

2.  We have now recorded three instances, albeit on the same ECI FTTC, where there has been a significant increase on very poorly performing VDSL lines if the ECI modem is replaced with a currently-upgraded HG612.

Kind regards,
Walter

[les-70]

   Thanks for that Walter, the blip is about 50m sooner than my most optimistic estimate of the CAB distance and previously with adsl I had attributed it to the end of a gauge change, split pair, bad joint, or less likely a bridged tap.  Maybe I also have very little trace of the cabinets and my previous adsl based view is correct.

  I did not notice any movement on the TDR trace but I am quite ignorant of beacons and even that they were there.  I would be prepared to pay for an SFI visit if I really thought that I would get a professional and carefully check and possible fixes made.  I am not keen on paying and being told its fine. 


I am using a HG612, as noted above it seems to do better in speed or currently at 80/20 in srnm,  but also if you look at the ping plots using TBB BQM above, in packet loss.  i.e. the ECI modem seems to loose packets, I tried two one unlocked and one locked - both the same. The HG612 does not loose any.

[burakkucat]

Regarding the MSAN "beacon". It is very characteristic and once seen is never forgotten.

Granted I have only seen it on lines connected to Huawei MSANs (for that is all to which I have access) but it is an upward "blip" that appears and disappears at a frequency of approximately 0.5 Hz. ("Now you see me, now you don't.")

I was told that a decision was made to enable the "beacon" so that the most new recruit (or numpty subcontractor) could easily identify and thus determine the pair length to, the MSAN. 

[les-70]

  Had another look at the TDR but this time via the vdsl output of the face plate and not into the  test socket.  I am assuming that will make little difference.  It certainly gives the same trace and no many how long I look the only movement is what I would call an odd tiny 1 "display bit" changes.  Assuming any beacons are intended to be clear there seem to be none on or enabled in my ECI cab.   

 Had what I think is another connection join the party but the impact of each successive connection seems to be dropping.  That may however just be luck depending on where connections are located relative to me.  Successive snrm drops have been 4.2,1.4, and 0.4 so far and consistent with each 1db drop matching a 4Mb/s drop in attainable speed.  I tried setting a few max down load sync speeds ( e.g. xdslcmd configure --maxDataRate 64000 16000 80000) on the HG612- 038 and deduce that things drop like 4Mb/s per db snrm provided all tones are in use and by about half that i.e. 2Mb/s per db if about half the tones have bits.  That seems consistent with the adsl2 being about 0.5 Mb/s per db on a fair connection.  So far the DLM seems to have tolerated these tests but I won't push my luck with more mucking about for bit.

[Chrysalis]

the maxdatarate syntax what does the 3rd value do?

[burakkucat]

  Had another look at the TDR but this time via the vdsl output of the face plate and not into the  test socket.  I am assuming that will make little difference.  It certainly gives the same trace and no many how long I look the only movement is what I would call an odd tiny 1 "display bit" changes.  Assuming any beacons are intended to be clear there seem to be none on or enabled in my ECI cab.

That is interesting. I wonder if the "beacon" was turned off during the recent MSAN firmware update? Perhaps Walter might be able to "take a look" at an Ewhurstian VDSL2 service, connected via an ECI cabinet?

[les-70]

 @Chrysalis  It called the maxaggregate and I guess that is what it is.  I have always used to sum of other two values for it hence xdslcmd configure --maxDataRate 64000 16000 80000. It won't make you faster but if used in moderation may get fast path back or it may break the DLM!!!  I would not advise setting value that would upset you too much if the DLM then capped you at that value.   The DLM does not seem to have noticed my experiments with it - but!!!

[les-70]

  I think I am at the end of the training period and no DLM intervention below 80/20 fast path has occurred or probably needed to occur.  I had recently in any case capped myself to help it ignore me.

 Recently I got out a twisted pair dsl lead -- previously with adsl2 it made no difference that I could detect.  It is much the same with the VDSL but at an exactly fixed sync the SNR does increase by 0.1 to 0.2 db.  I have changed two and fro twice and each time had a delta or 0.1 or 0.2 up/down and always best with the twisted lead.  I am leaving it in place but it is a very small improvement.  I guess if the cables runs near some noisy things you might see a bigger change.

[Darren]

The details are hazy in my mind but asbo once wrote about it being important for the faceplate > modem cable characteristics to match the lead in to minimise signal loss back up the cable from when it hits the different wire. Can't remember all the tecknical ins and outs but it made sense at the time, maybe someone else also read it and has a better memory. I vowed to get hold of some cable which is the same as my lead in and make a patch cable out of it but have not got around to it yet.

[Black Sheep]

FYI ...... may be going off at a tangent, but I visited a premises recently and the guy was using an Ethernet cable (from the output port of the Hub3), and strangely enough had another Ethernet cable wrapped exactly 3 times around the 'connected' Ethernet cable near the jack-plug. Then the two lengths were loosely taped together along their length, and at the other end where the other jack-plug is, he had wrapped it around another 3 times ??.

I asked him what that was all about, and he said he'd read it on some forum where he'd been given good advice before, and that it apparently gave him a better service (Bloke was a 'Gamer' as well). He said the instructions he'd been given were pretty exact about the 3 'wrap arounds' near to the plugs on the 'used' Ethernet cable.

Never seen, or heard about this before but thought it worth a share.

[les-70]

  @Darren Thanks for that I am sure matching impedance's  between the dsl lead and the cable to master socket would be ideal.  I also had thought of that but lack the crimping tools for the connectors.  Rather like the twisted pair cable I tried I suspect a small improvement may occur if you did that depending on how bad the original cable was.  Until recently I had a favorite lead which did seem a little better than the others but just as fttc was installed the locking tag fell off!. 

[Chrysalis]

FYI ...... may be going off at a tangent, but I visited a premises recently and the guy was using an Ethernet cable (from the output port of the Hub3), and strangely enough had another Ethernet cable wrapped exactly 3 times around the 'connected' Ethernet cable near the jack-plug. Then the two lengths were loosely taped together along their length, and at the other end where the other jack-plug is, he had wrapped it around another 3 times ??.

I asked him what that was all about, and he said he'd read it on some forum where he'd been given good advice before, and that it apparently gave him a better service (Bloke was a 'Gamer' as well). He said the instructions he'd been given were pretty exact about the 3 'wrap arounds' near to the plugs on the 'used' Ethernet cable.

Never seen, or heard about this before but thought it worth a share.

gamers can often get things wrong and misinformation spreads like a plague.

Probably read about twisted cabling been good, and somehow they translated it to twisting ethernet cables on their lan

[NewtronStar]

gamers can often get things wrong and misinformation spreads like a plague.

Probably read about twisted cabling been good, and somehow they translated it to twisting ethernet cables on their lan

That sounds like the gamer is trying to make an RF Choke he would have been better of using a ferrite bead at each end of ethernet cable.

[Darren]

les, I got a crimp tool for about £4 of ebay, didn't expect it to be any good but it's been fine for the couple of dozen I've done so far.

NewtronStar, that's what I thought when reading Black Sheeps post. A choke on the power supply cable before it enters the modem/router can apparently sometimes help too.

[Black Sheep]

Sorry Darren, it may not have read as well as I would have liked. I didn't mean the 'power supply' lead, I meant the Ethernet cable between the Router/Hub and the PC Tower.