FTTC download speed halved - external cause?

Hi,

FTTC download speeds have halved in recent weeks.

Normal speed is 34Mbps (on 40/10) and cabinet is between 260-300m by cable run from the property.

This graph shows the spectrum in use and was captured when connected directly to the master socket, which is an LJ 2/1A type, without anything else connected.
Normal setup is an NT5E after the master and the same results are achieved.

Only differences between my modems is the attainable speed.

Reported to ISP and awaiting response.

Any ideas what would knock-out the upper part of the spectrum?

TIA.

G.

When you say download speed do you mean speed test or sync speed from modem?

What are teh max attainable speeds? Today and if you have/remember from before the problem. I would be expecting to see a download Max Achievable in the region of 90Mbps on a line that short - the attenuation figures should also be around 12dB. Upstream will should show a max in excess of 25Mbps

From VMG3925-B10C:

VDSL Port Details Upstream Downstream
Line Rate: 7.199 Mbps 16.493 Mbps
Actual Net Data Rate: 7.200 Mbps 16.494 Mbps
Trellis Coding: ON ON
SNR Margin: 17.2 dB 23.8 dB
Actual Delay: 0 ms 0 ms
Transmit Power: 6.8 dBm 11.2 dBm
Receive Power: -10.5 dBm -2.5 dBm
Actual INP: 54.0 symbols 55.0 symbols
Total Attenuation: 17.3 dB 13.7 dB
Attainable Net Data Rate: 23.341 Mbps 53.913 Mbps

From Vigor130 (bridge mode):

---------------------- ATU-R Info (hw: annex A, f/w: annex A/B/C) -----------
Running Mode : 17A State : SHOWTIME
DS Actual Rate : 16494000 bps US Actual Rate : 7200000 bps
DS Attainable Rate : 87937712 bps US Attainable Rate : 19849000 bps
DS Path Mode : Fast US Path Mode : Fast
DS Interleave Depth : 1 US Interleave Depth : 1
NE Current Attenuation : 15 dB Cur SNR Margin : 23 dB
DS actual PSD : 6. 6 dB US actual PSD : 11. 4 dB
NE CRC Count : 0 FE CRC Count : 3815
NE ES Count : 0 FE ES Count : 1932
Xdsl Reset Times : 0 Xdsl Link Times : 3
ITU Version[0] : b5004946 ITU Version[1] : 544e0000
VDSL Firmware Version : 05-07-06-0D-01-07 [with Vectoring support]
Power Management Mode : DSL_G997_PMS_L0
Test Mode : DISABLE
-------------------------------- ATU-C Info ---------------------------------
Far Current Attenuation : 17 dB Far SNR Margin : 15 dB
CO ITU Version[0] : b5004244 CO ITU Version[1] : 434da48c
DSLAM CHIPSET VENDOR : < BDCM >

Original speed:

34Mbps was measured from ThinkBroadband speed test weeks before I registered.

Edit:

Fritz!Box summary:

Down Up
Attainable Kbps: 69881 22583
Current: 16494 7200
SNR (dB): 23 17
Line attenuation: 15 16

Edited by deleted (Mon 15-Jun-20 15:57:38)

In reply to a post by MHC:

Today and if you have/remember from before the problem.

No stats before the problem other than 34Mbps from ThinkBroadband speed test.

Do you have a clear dialling tone on the landline when tested at the master socket ?

In reply to a post by Zarjaz:

Do you have a clear dialling tone on the landline when tested at the master socket ?

Yes.

17070 only identifies line number - no option for silent test.

Downstream Attenuation 13.7 dB - not far from my guess but suggest top end or maybe a little more than your 300m. Max achievable of just under 88Mbps which ties in

The SNR at 23.8dB is the major cause of the slowdown and you are going to have to wait for that to recover - unless there is a major occasional noise source that comes and goes.

I would say, power off the modem/hub. Wait a minute and disconnect from line and all PCs. Wait for maybe an hour - power up modem and wait 5 minutes. Connect to the line and wait for another 5 minutes. Reconnect PC and access stats.

If there is not much change you can either push the ISP to get a reset on the line or wait for it to recover.

In reply to a post by MHC:

The SNR at 23.8dB is the major cause of the slowdown and you are going to have to wait for that to recover - unless there is a major occasional noise source that comes and goes.

It's persistent now for the past fortnight.

Using the SNR adjustment on the Vigor130 only gets the attainable rate to vary.

In reply to a post by MHC:

I would say, power off the modem/hub. Wait a minute and disconnect from line and all PCs. Wait for maybe an hour - power up modem and wait 5 minutes. Connect to the line and wait for another 5 minutes. Reconnect PC and access stats.

Ta.

In reply to a post by MHC:

Downstream Attenuation 13.7 dB - not far from my guess but suggest top end or maybe a little more than your 300m. Max achievable of just under 88Mbps which ties in

More than 400m see stats below (default at 6dB SNR)

500m long line (line attenuation 14.0dB on FTTC) 60/20
450m long line (line attenuation 13.3dB on FTTC) 65/20
400m long line (line attenuation 12.6dB on FTTC) 70/20
350m long line (line attenuation 11.9dB on FTTC) 75/20
300m long line (line attenuation 11.2dB on FTTC) 80/20
250m long line (line attenuation 10.5dB on FTTC) 80/20
200m long line (line attenuation 9.8dB on FTTC) 80/20

Edited by adslmax (Mon 15-Jun-20 16:39:37)

When did you invent those figures?

In reply to a post by forafewbytesmore:

In reply to a post by MHC:

The SNR at 23.8dB is the major cause of the slowdown and you are going to have to wait for that to recover - unless there is a major occasional noise source that comes and goes.

It's persistent now for the past fortnight.

Using the SNR adjustment on the Vigor130 only gets the attainable rate to vary.

Openreach expressly say adjusting the SNR, (actually it is the SNRM), should not be done. Unlike on the BT Wholesale and LLU ADSLx where it was standard practice.

Also, given that it is SNRM, which way are you adjusting it? SNR needs to be as high as possible. SNRM as low as possible so long as stability is maintained and error rates acceptable.

I'm very suspicious you have increased this yourself, as both the sets of stats show you as on Fast Path with one showing a zero "delay".
Those are remarkable given those SNRMs. You should be on heavy interleaving with a substantial "delay".

Default setting on FTTC for SNRM is 6dB. With luck DLM will then lower it automatically over time.

In reply to a post by RobertoS:

I'm very suspicious you have increased this yourself, as both the sets of stats show you as on Fast Path with one showing a zero "delay".

Nope. Adjustments took place to fight the issue after it appeared (with no effect).

In reply to a post by RobertoS:

I'm very suspicious you have increased this yourself, as both the sets of stats show you as on Fast Path with one showing a zero "delay".
Those are remarkable given those SNRMs. You should be on heavy interleaving with a substantial "delay".

I'll need to read up to understand that.

In reply to a post by RobertoS:

With luck DLM will then lower it automatically over time.

For the original 34Mbps would the graph show a usage (in blue) from 2816 onwards OR more bits in use at the lower end?

TIA.

FTTC bins/frequencies aren't used like ADSLx where the upstream ones are on the left and the downstream on the right.

On VDSL2 (which is what actually reaches your premises) they are all mixed up in a number of bands. I can't show you what mine looks like as I ditched FTTC 18 months ago.

For a longer explanation of how SNR and SNRM interact, this oldish page is still a useful explanation.

In reply to a post by forafewbytesmore:

It's persistent now for the past fortnight.

Using the SNR adjustment on the Vigor130 only gets the attainable rate to vary.

Your line looks like it has been capped (banded), on both the downstream and the upstream.

Banding is usually applied when a line is resyncing too much.

This is sometimes caused by the end user troublesome with their modem and rebooting it over and over or swapping modems around.

If the sync rate stats the exact same between resyncs and the SNRM is always above 6dB then the line is definitely banded.

It needs stability (no resyncs) for DLM to gradually remove the banding.

In reply to a post by RobertoS:

I'm very suspicious you have increased this yourself, as both the sets of stats show you as on Fast Path with one showing a zero "delay".
Those are remarkable given those SNRMs. You should be on heavy interleaving with a substantial "delay".

Huawei cabinet, G.INP active on both downstream and upstream.
Delay should be zero.

Edited by j0hn83 (Mon 15-Jun-20 19:33:00)

It would not have those stats in that situation!

The shovel would be self-cleaning, everything would slide off it so fast.

In reply to a post by RobertoS:

It would not have those stats in that situation!

It would.

Actual INP: 54.0 symbols 55.0 symbols

DSLAM CHIPSET VENDOR : < BDCM >

Huawei cabinet, G.INP on both down and upstream.

Confusion might come from...

DS Path Mode : Fast US Path Mode : Fast

I think Drayteks report Path Mode: Fast any time the delay is 0, which includes G.INP.

In reply to a post by forafewbytesmore:

Reported to ISP and awaiting response.

ISP called about 48 hours after support ticket submitted. Impressive.

Line test passed.

Awaiting network team call or text regarding site visit.