Print

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

[les-70]

 Have a look at stats for iMiMx, Tbailey2, g3uiss and  JasonKruys.  from very quick look on MDWS at interleaved folk I think they and yourself may be nearest to a fast path return.

  Some may be on standard profile and need half the error rate of the others.  It really would be big help if we knew the limit and with whole day or two to spare it might be possible to look for past transition on MDWS and see the largest preceding error counts. My guess would be less than ~28 on fast profile and less than ~14 on standard profile.  One day with ES over such a  value and the count starts again. As I have said J0hn 's advice on applying a really large speed cap to be absolutely sure that DLM removes interleaving is probably best to get back as quickly as possible into known DLM ES territory on fast path.   In your case it looks like that may be cap to 25 or 30Mb/s.

 

[N0STIE]

I wouldn't mine capping sync to 30 MB/s but perhaps it is worth waiting few more days as j0hn said it might take up to 10 days to remove interleaving. If nothing happens after 10 days I'll cap the sync to 30 MB/s and still be monitoring the error rate. The most ES seem to appear between 17-20 then there is just 1 or 0 ES for the rest of the night till afternoon. Currently the average amount of ES is under 20 per 24 hours. I am with BT so pressumably they use Speed profile which is more tolerable when it comes to errors?

Looking at their stats and their uptime it seems really strange to me that they are still interleaved with similar amount of ES to me. Does it mean I might be waiting that long time as they are?

[les-70]

  I have always had to wait 15 days.  I think how long may depend on past line history and as I mentioned it is not the average error rate per day but the max number on an any one of the 10 or 15 days that counts. You had 27 one day, that is the only number that counts so far no matter how small the ES are on other days.  You need low ES on all of 10 or 15 consecutive days

 If you follow one of the speed capping sequences given above that includes the 30 mins wait and avoids a double resync it should not make you have to starting counting from day one again. 

[N0STIE]

  I have always had to wait 15 days.  I think how long may depend on past line history and as I mentioned it is not the average error rate per day but the max number on an any one of the 10 or 15 days that counts. You had 27 one day, that is the only number that counts so far no matter how small the ES are on other days.  You need low ES on all of 10 or 15 consecutive days

 If you follow one of the speed capping sequences given above that includes the 30 mins wait and avoids a double resync it should not make you have to starting counting from day one again.

Thanks for making it clear now.

I understand that 27 ES per day isn't as low as it should be considering I am interleaved but it may be low enought to bring back fast path unless the errors will get higher.

If ES exceed 27 or higher in the next days, I'll cap the speed to 30MB/s as I think it doesn't matter now, because the highest number is 27 already and then it never exceeded 20 per day.

[ejs]

Do we know for certain that's it's only ES that the DLM is considering? What if the FEC+interleaving was correcting all the errors, shouldn't the DLM consider the FEC count or FECS (FEC seconds) before switching it off?

[j0hn]

Do we know for certain that's it's only ES that the DLM is considering? What if the FEC+interleaving was correcting all the errors, shouldn't the DLM consider the FEC count or FECS (FEC seconds) before switching it off?

I've never seen anything written saying it looks at FEC numbers but in my experience it does. I was ILQ green for months without interleaving dropping off. It's only when I'm in the low thousands/hundreds FEC's per hour that interleaving drops.

[les-70]

    I really don't know but I have looked at MDWS a few times over the last couple years checking for interleaved folks ES rate.  I don't think anyone failed to go back to fast path with a sustained ES less than about 20-30/day.  That may reflect FEC but my guess is that the DLM keeps things simple and just checks ES and resyncs.  I think FEC just mean that interleaving works and ES means it does not quite do the job.

[j0hn]

I don't think anyone failed to go back to fast path with a sustained ES less than about 20-30/day.

I have about 6 months of interleaving stats on MDWS that says otherwise, username Minted.
Most of those 6 months my ES rate was half the ES limit for my Standard profile. It was only when I capped my line and FEC's reduced from 10000+ a minute to under 10000 an hour that I returned to fastpath. My line uncapped, with interleaving+FEC, never returns fastpath despite always being under the ES limit. Uncapped I get anything between 10000-50000 FEC a minute.

[N0STIE]

There was a power outage in my area. I didn't let HG612 resync on its own after the power came back, I capped sync to 30000 as suggested. Hopefully it will return to fastpath quicker.

[N0STIE]

Only 5 ES in 24 hours after capping sync to 30000. I guess not that bad.

[les-70]

I see that you have fast path back  .  I would wait a few days to see the daily error count and then increase the speed in steps until it looks like you have 0.3 or 0.5 of the daily max of 2880ES.  If you go too high in prevailing errors a bad day may get interleaving back.

[N0STIE]

I see that you have fast path back  .  I would wait a few days to see the daily error count and then increase the speed in steps until it looks like you have 0.3 or 0.5 of the daily max of 2880ES.  If you go too high is prevailing errors a bad day may get interleaving back.

Yes, that is right. I understand by 0.3 or 0.5 you mean 30%-50% of 2880?

I just posted about this in this topic early today.

http://forum.kitz.co.uk/index.php/topic,18791.15.html

[N0STIE]

I am wondering if cordless phone (the receiver) which is next to my master socket connected using short 0.50m cable can cause negative impact on errors? Phone with handset is downstairs. Can anyone confirm?

[roseway]

The phones themselves shouldn't interfere with DSL, but the power supplies in their chargers can.

[N0STIE]

The phones themselves shouldn't interfere with DSL, but the power supplies in their chargers can.

Thanks Eric for reply.

The charger with the phone itself is downstairs. Only the receiver is very close to NTE5. Would it make any difference if the charger was here close to NTE5?