Print

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

[boost]

The first link I found for rough line lengths and speeds is here: https://www.thinkbroadband.com/guides/fibre-fttc-ftth-broadband-guide

400m   42 Mbps   16 Mbps   45%

Based on that, your line (434m) would appear to be performing above average, in terms of sync speed?
If we allow ourselves to assume that your sync speed has not appreciably changed, I suppose we must consider what changes might affect your perception of the speed?

From your stats, two things stand out to me:

Since Link time = 9 days 11 hours 23 min 18 sec
FEC:            0               6695
CRC:            9501            494
ES:             6655            489
SES:            4               0
UAS:            0               0
LOS:            0               0
LOF:            0               0
LOM:            0               0
HEC:            23414           0

You have zero downstream FECs. That strikes me as very odd
This means your modem corrects zero downstream errors. In short, it means everything is too bollocksed for it to correct. Things that can't be corrected hit the CRC/HEC/ES counter and get dropped.

In all likelihood, the slow down you have been experiencing can probably be attributed to TCP constantly having to re-request segments after your modem decided it couldn't do anything with the corrupted frames/cells and subsequently dropping them. For web traffic, this means delays loading pages and potentially having to renegotiate SSL/TLS more than is required, I suppose.

It seems quite likely to me that you would benefit from some DS INP and/or interleaving. It may be that this was on your line previously and for whatever reason, was removed, thus triggering your perception of reduced speeds?

If forum anecdotes are to be believed, simply requesting a DLM reset will see your line adopt DS INP and interleaving. If that solves it, it might be worth considering being moved to a 'less tolerant' error profile.

Standard/Stable, perhaps?

[vultura]

The first link I found for rough line lengths and speeds is here: https://www.thinkbroadband.com/guides/fibre-fttc-ftth-broadband-guide

400m   42 Mbps   16 Mbps   45%

Based on that, your line (434m) would appear to be performing above average, in terms of sync speed?
If we allow ourselves to assume that your sync speed has not appreciably changed, I suppose we must consider what changes might affect your perception of the speed?

From your stats, two things stand out to me:

You have zero downstream FECs. That strikes me as very odd
This means your modem corrects zero downstream errors. In short, it means everything is too bollocksed for it to correct. Things that can't be corrected hit the CRC/HEC/ES counter and get dropped.

In all likelihood, the slow down you have been experiencing can probably be attributed to TCP constantly having to re-request segments after your modem decided it couldn't do anything with the corrupted frames/cells and subsequently dropping them. For web traffic, this means delays loading pages and potentially having to renegotiate SSL/TLS more than is required, I suppose.

It seems quite likely to me that you would benefit from some DS INP and/or interleaving. It may be that this was on your line previously and for whatever reason, was removed, thus triggering your perception of reduced speeds?

If forum anecdotes are to be believed, simply requesting a DLM reset will see your line adopt DS INP and interleaving. If that solves it, it might be worth considering being moved to a 'less tolerant' error profile.

Standard/Stable, perhaps?

According to this link https://www.increasebroadbandspeed.co.uk/2013/chart-bt-fttc-vdsl2-speed-against-distance, at 500m 60Mbps should be possible.

There were DS FEC errors with the HG612 installed (see stats from that higher up thread).  Wonder if the new socket/faceplate combined with the Zyxel have made things worse in this respect?

I thought G.Inp had been disable on ECI cabinets.  I could well be wrong though.

[ktz392837]

>rough line speeds

Those are very worse case estimates in my opinion - I would say bordering on being wrong.  At 16db attenuation 50Mb sync is extremely poor.

If MDWS was still with us I would say at 16db attenuation (or "434m" if you prefer) the average sync would be 70Mb.  Even OR believe on an *impacted* line 76.7Mb is possible!

I am no expert and may have missed some stats saying otherwise but that number of CRC/ES is nothing in the whole scheme of things you can have 2880 ES a day before DLM does anything and you can get 9000 CRC errors probably in a few blinks of an eye.  So that is nothing over a 9 day period.

Adding interleaving to the line will just slow it down further for no reason.

Ginp would help no end but us on ECI cabs don't have that luxury ... don't get me started on that subject ...

[boost]

I always hope someone will highlight anything I've said that's not quite correct, thanks

I suppose the problem you have, is that you have no real idea what you were previously synched at

My stats, in case it helps are:

Code: [Select]

Downstream Upstream
Line attenuation (dB):  12.8 0.0
Signal attenuation (dB): Not monitored
Connection speed (kbps): 71866 20000

I can see the (BDCM) cab from my window. 75 - 100 metres as the crow flies.

[vultura]

>rough line speeds

Those are very worse case estimates in my opinion - I would say bordering on being wrong.  At 16db attenuation 50Mb sync is extremely poor.

If MDWS was still with us I would say at 16db attenuation (or "434m" if you prefer) the average sync would be 70Mb.  Even OR believe on an *impacted* line 76.7Mb is possible!
I am no expert and may have missed some stats saying otherwise but that number of CRC/ES is nothing in the whole scheme of things you can have 2880 ES a day before DLM does anything and you can get 9000 CRC errors probably in a few blinks of an eye.  So that is nothing over a 9 day period.

Adding interleaving to the line will just slow it down further for no reason.


Ginp would help no end but us on ECI cabs don't have that luxury ... don't get me started on that subject ...

From PN member centre -

Estimated line speed:73Mb (This may vary between 54Mb and 73Mb)  - Checked on 2017-10-22 17:47:15
Current line speed:53 Mb

Maybe they should rename estimates to guesstimates and drop all pretence of a guaranteed speed.  Not just PN either, but all providers.

[vultura]

I always hope someone will highlight anything I've said that's not quite correct, thanks

I suppose the problem you have, is that you have no real idea what you were previously synched at

My stats, in case it helps are:

Code: [Select]

Downstream Upstream
Line attenuation (dB):  12.8 0.0
Signal attenuation (dB): Not monitored
Connection speed (kbps): 71866 20000

I can see the (BDCM) cab from my window. 75 - 100 metres as the crow flies.

I have a random stats capture from the HG612 in August 2018.  Why on earth logging started working and then stopped again I have no idea.  However it does show a higher sync speed than I have seen in the last couple of months.

[vultura]

When I had the Hub One installed and was keeping a keener eye on things I did notice something in the logs.

Am I correct in thinking this is caused by a crosstalker's modem rebooting or is it something else?

[vultura]

Plots from around the engineers visit.

[ktz392837]

>Am I correct in thinking this is caused by a crosstalker's modem rebooting or is it something else?
Generally I would say Yes.

>Plots from around the engineers visit.
Attenuation dropping to 0db is probably the engineer running a test/reset on the line if this was around the time the engineer was working on the line.
If at any other time the attenuation drops to 0db or varies "wildly" when the phone is ringing or just generally it is a sign of a fault.

>MDWS comparison

I have dug out of an archive of some sync speeds, attenuation and cabinet type figures from when MDWS was live.  I have included Attenuation of 15db, 16db and 17db lines.  If you exclude any 40000/39999/55000 type figures that are probably on limited speed products and perhaps deduct 5-10Mbs (chosen from thin air!) from the Huawei entries to adjust for the ECI cabs not having GINP/3db you have certainly not won the line lottery.

Code: [Select]

102388 15.0 Huawei
65082 15.1 ECI
40000 15.3 Huawei
69760 15.4 ECI
74473 15.7 Huawei
79998 15.9 Huawei
76718 15.9 Huawei
75968 16.0 Huawei
73842 16.0 Huawei
61765 16.0 Huawei
65052 16.0 Huawei
79987 16.0 ECI
79547 16.1 Huawei
55000 16.2 Huawei
77226 16.2 Huawei
62218 16.2 ECI
58850 16.3 ECI
79999 16.3 Huawei
39999 16.3 Huawei
69410 16.4 ECI
65534 16.7 Huawei
64772 16.7 ECI
67087 16.9 Huawei
69896 16.9 Huawei
68214 16.9 Huawei
79999 17.0 Huawei
20939 17.0
63529 17.0 Huawei
70152 17.0 Huawei
64992 17.1 ECI
77234 17.1 Huawei
79997 17.2 Huawei
40000 17.2 Huawei
71315 17.3 Huawei
64036 17.3 ECI
74618 17.3 Huawei
65679 17.4 Huawei
79999 17.4 Huawei
72022 17.6 Huawei
64127 17.7 ECI
66997 17.7 Huawei
68573 17.8 Huawei
77053 17.8 Huawei
70910 17.8 Huawei
78912 17.8 Huawei
60000 17.9 Huawei
79997 17.9 Huawei

Please note I am no expert and there are many people on the forum which are much more qualified to give an opinion than me.

[vultura]

>Am I correct in thinking this is caused by a crosstalker's modem rebooting or is it something else?
Generally I would say Yes.

>Plots from around the engineers visit.
Attenuation dropping to 0db is probably the engineer running a test/reset on the line if this was around the time the engineer was working on the line.
If at any other time the attenuation drops to 0db or varies "wildly" when the phone is ringing or just generally it is a sign of a fault.

>MDWS comparison

I have dug out of an archive of some sync speeds, attenuation and cabinet type figures from when MDWS was live.  I have included Attenuation of 15db, 16db and 17db lines.  If you exclude any 40000/39999/55000 type figures that are probably on limited speed products and perhaps deduct 5-10Mbs (chosen from thin air!) from the Huawei entries to adjust for the ECI cabs not having GINP/3db you have certainly not won the line lottery.

Thanks for reply ktz392837.

Sorry I should have said that I posted the plots up from around the time of the engineers visit to show as much as anything else that the sync speed did drop very slightly afterwards, which is not what you might expect having had a visit.

I see a good many others with longer lines or similar attenuation figures getting higher speeds than I see.  Perhaps it is down to old vs new cabling.

[j0hn]

The first link I found for rough line lengths and speeds is here: https://www.thinkbroadband.com/guides/fibre-fttc-ftth-broadband-guide

400m   42 Mbps   16 Mbps   45%

Based on that, your line (434m) would appear to be performing above average, in terms of sync speed?

Those estimates are utter nonsense, and way way on the low side.

I get 46Mb at roughly 1020m from the cabinet.
42Mb at 400m is terrible.

You have zero downstream FECs. That strikes me as very odd

As i mentioned earlier the line is Fastpath.
There should be zero FEC.

It seems quite likely to me that you would benefit from some DS INP and/or interleaving. It may be that this was on your line previously and for whatever reason, was removed, thus triggering your perception of reduced speeds?

I disagree. That's just going to knock another 10% off the sync.

The daily ES limit on Plusnet (Speed profile) is 2880.
The OP's line is getting less than 10% of that with an HG612 in the test socket
Less than 25% with the Zyxel and an MK4
The line does not need any error protection.

It's running on the best profile possible for an ECI cabinet.

[vultura]

Might be jumping the gun here, but was just on roadworksdotorg and found the following updated today -

MJ Quinns are to install a new DSLAM cabinet and 25m of ducting to be laid in footway and grass verge.  Planned 11th to 22nd Feb.

Location right where my existing cab is.

Now oddly I did notice one of their vans and engineers working last night at the location as I drove home around 6.30, coincidence I wonder?

[Chrysalis]

The first link I found for rough line lengths and speeds is here: https://www.thinkbroadband.com/guides/fibre-fttc-ftth-broadband-guide

400m   42 Mbps   16 Mbps   45%

Based on that, your line (434m) would appear to be performing above average, in terms of sync speed?
If we allow ourselves to assume that your sync speed has not appreciably changed, I suppose we must consider what changes might affect your perception of the speed?

From your stats, two things stand out to me:

You have zero downstream FECs. That strikes me as very odd
This means your modem corrects zero downstream errors. In short, it means everything is too bollocksed for it to correct. Things that can't be corrected hit the CRC/HEC/ES counter and get dropped.

In all likelihood, the slow down you have been experiencing can probably be attributed to TCP constantly having to re-request segments after your modem decided it couldn't do anything with the corrupted frames/cells and subsequently dropping them. For web traffic, this means delays loading pages and potentially having to renegotiate SSL/TLS more than is required, I suppose.

It seems quite likely to me that you would benefit from some DS INP and/or interleaving. It may be that this was on your line previously and for whatever reason, was removed, thus triggering your perception of reduced speeds?

If forum anecdotes are to be believed, simply requesting a DLM reset will see your line adopt DS INP and interleaving. If that solves it, it might be worth considering being moved to a 'less tolerant' error profile.

Standard/Stable, perhaps?

boost thats normal for fast path which doesnt correct DS errors.  On ECI cabinets fast path is the default.  6.6K ES in 9 days, if we assume its a steady increment is comfortably below the DLM threshold and he is averaging 1.5 CRC per ES which is a good ratio.

[Ronski]

That sounds like they are installing a second fibre cabinet, which will be a Huawei. When your contract expires and presuming the cabinet is live (no easy way to tell) you could order another line, or drop back to ADSL and then upgrade to VDSL and you should hopefully end up on the new cabinet with the benefits that brings. That's something I considered doing, but then VM installed full FTTP.

Regarding line length, my line was circa 450 meters, probably contained some aluminium, when I cancelled my line in April 2018 the sync was 41/7. My speeds would often fluctuate, downstream was often around 47Mbps, upstream sometimes as low as 5 or 6 Mbps. Most my line ever achieved for any length of time was 54/14 when I originally had it installed back in August 2012. Generally my line was was the lowest of others at the time with similar lengths, fastest I think was around 70 from memory.

You can't rely solely on length as you could have alluminium, many joints, few joints, thin copper, or in some cases copper as thick as 0.9mm copper.

[boost]

I have a random stats capture from the HG612 in August 2018.  Why on earth logging started working and then stopped again I have no idea.  However it does show a higher sync speed than I have seen in the last couple of months.

48 vs 53. Is it fair to suggest that's not appreciably higher? It's not 48 vs 80, for example.