FTTC performance - Initial line speeds no longer received

I recently upgraded from ADSL with Zen to their unlimited fibre product
Initial estimates for the line were 61Mbps down and 20 Mbps up, so I signed up for the upgrade. The speeds were good for the first 13 days and I was very happy
Now the speeds have dropped right back and after speaking with Zen, have been told the following by their support department:-

"I have reviewed your connection details and I can see the original advised Estimated connection speeds were 61mbps Downstream and 20mbps upstream. The connection speed range allowed for your connection, as advised by OFCOM, is between 43.04mbps and the advised 61mbps for the downstream service and throughput should not fall below 35mbps.
Therefore, as your current connection speed is 47.6mbps Zen would not be able to log a fault with Openreach for investigation, there are diagnostics that can be completed to try to increase the speed of the connection, for example ensuring the connection is made to the Openreach Modem from the Openreach installed pre-filtered master socket, with no other devices connected to the line, using the supplied cabling."

I feel that this situation is not good and feel I have been misled; as I have only had the FTTC since 14th October, the promised speeds lasted for a very short period of the one year contract. Surely, if the initial estimated speeds were reached for the first 13 days of the contract, there is no reason that they should have dropped to just above the OFCOM standards as the line was obviously capable of those speeds?

I have been a happy customer of Zen for well over 15 years for home & business ADSL and for Leased Lines for business and feel they are the best ISP, but feel that this has taken the edge off their reputation for me. I feel let down but this speed reduction is probably due to BT rather than Zen . . . .
Have any other people experienced such speed drops after initial estimated speeds?

Curious to know more about this Ofcom standard?

fttc starts at best speed and dlm lowers till it reaches a stable speed. The wholesale checker gives the range that Zen seem to have now mentioned.

They might let you downgrade to a 40/10 product at lower price.

At end of day likely to be same connection speeds no matter which ISP

We need to test something that may or may not work.

Zen may have you on a BT Wholesale system, in which case it will. On a number of exchanges they have a sort of LLU setup, and it may not. The first stage should work in either case.

Please run the BT Wholesale Performance Test. Ignore the red instructions for now, just click you have done them.

Once you get the results page, make a note of the figures, then click Further diagnostics at the bottom. This is the bit that may not work. If it doesn't, just give us the noted results, though they probably won't help much.

It will ask for your phone number. Once you get the results from that please copy and paste the full contents of the down and up results text boxes. We don't need to see the graphics.

First: Actual FTTC Speeds
Yes, unfortunately these can drop for no apparent reason.
There is of course a reason, in fact there are two, but neither are obvious to the average user.

Reason 1: Crosstalk
As take-up increases, the number of modems transmitting VDSL signals increases. Modems on neighbouring lines all interfere with your signal, degrading things to a greater or lesser extent ... This increase in noise generally reduces the speeds you can achieve, and the effect is almost entirely random - the timing can be random, and the amount of impact can be random. Worst case can suck 30, 40 or even 50% of your speed, over time.

Reason 2: DLM
BT's DLM monitors the error statistics reported by your modem 24x7, with the aim of keeping your connection stable. If the number of errors gets beyond a threshold, DLM intervenes ... which can eat anywhere between 10% and 20% of your speed.

As takeup increases, that extra noise can cause errors, which means you are more likely to trigger DLM. A double whammy.

Early adopters (within a particular cabinet) are likely to see speeds higher than the estimate at first, but can then drop considerably over time.

Solution: Employ vectoring. This technology attempts to negate crosstalk, giving people their speeds back (and likely more). It is being trialled, and isn't ready for mass deployment yet.

Second: Estimates
BT provides the estimates to ISP's like Zen, who "interpret" them for you. They originally start in the form of an A range and a B range, while Zen have obviously only passed on a single figure.

BT make estimates based on the performance of "similar lines"; the top of the range is the speed that the 80th percentile achieves, while the bottom of the range is the 20th percentile (which means that 20% of similar lines achieve lower speeds!)

Remember - an estimate is not a promise.

Third: Getting BT to investigate
BT know the effects of crosstalk & DLM, and know that subscribers will be hit by them, so expect speeds to drop. There is nothing they can do about these effects (until vectoring gets deployed), so they really want to avoid sending engineers for speed drops that are unavoidable.

Until recently, BT had set thresholds: If your speed was above the minimum of the range (which is the 20th percentile), they wouldn't investigate. If it dropped more than 25% in one go, they would investigate.

There was a story yesterday reporting that BT had changed the threshold that they would act at - from the "minimum of the range" (ie the 20th percentile) to the 10th percentile, which is rather lower. It appears (from this report of the story) that the 10th percentile also happens to be the Ofcom-recommended threshold for determining remedial action.

That probably explains why Zen refer to an Ofcom range as being 43 to 61Mbps: the 43Mbps would be the speed of the 10th percentile of similar lines (ie 10% of similar lines get less than 43, and 90% get more).

It seems that take-up of FTTC is now hitting the critical threshold - and people are being affected by this issue more and more.

It really needs ISPs to be honest; they need to tell people that whatever speed they get initially may drop, and can drop to anything within the range and still be considered acceptable.

Unfortunately, people have been conditioned by ADSL behaviour, where they achieve a certain speed, and believe it is their inalienable right to achieve that speed forever more. This is no longer true.

Edited by deleted (Fri 05-Dec-14 00:08:52)

Sorry - I should have pointed you to the BT Wholesale speed checker.

First, try this checker using a phone number; it should give you the estimate ranges (which you can copy into a reply)

If that doesn't work, you'll have to try the address checker; first insert your postcode and submit, then choose your full address from the dropdown list.

Hi RobertoS

BT Wholesale speed test carried out this am:-
Download - 41.52 Mbps
Upload - 11.59
Ping latency - 35.13

Hi WWWombat
Thanks for the reply
I used the same BT availability checker before I changed to fibre and this was happily showing the between 79.3 and 60 Mbps - Zen's website line checker showed 60 Mbps at that time. (BT's availability checker is now showing speeds of between 51.6 and 35.6)
I was initially running at between 55 and 57 Mbps for 13 days and then it suddenly fell to 44 /45 Mbps and has dropped a little since then
Interestingly, I installed a Vigor 2860 VDSL modem / router on the first day after the initial BT install / face plate change and this worked really well with stability and good speeds.
When I raised the support call with Zen, they asked if I still had my BT modem installed; so I reverted back to the BT modem and the figures have declined even further
Last night I put back the Vigor 2860 and speeds seem to have stabilised and even increased a little

Don't forget to do the "further diagnostics" page. That holds the IP profile data, from which the sync speed can be calculated.

the situation is very similar to that of ADSL as FTTC is only VDSL over phone lines so the data rates maybe higher and the distances it travels much shorter but otherwise it is the same beast.

As already mentioned, the more lines that go live via your cabinet the greater the crosstalk issue, this was the same for ADSL and is why people on long ADSL lines got slower and slower and some dropped off completely when more people closer to the exchange went live.

What is interesting with the BDUK process for state funded roll out is how the achievable sync/speed is going to be policed to ensure compliance with the requirements of the contract with the Local Authority.

The wording seems to imply that the tick box speed will be the one measured at first connection. not what happens after the DLM training period or when the rest of the lines on the cab light up.

Although to be fair the LA's i have questioned just shrug when asked what procedure they have in place for confirming service speeds achieved. most just quote they believe the projected figures from Openreach's data .

FTTC/VDSL is a very analogue system just as ADSL and is susceptible to all manner of interference and degradations as it travels twisted pair phone lines. If you have overhead lines then this situation worsens from Shortwave broadcast interference. More so than ADSL.

So the have / have not and speed disparity issue from ADSL will still exist as will those who start with a working long line FTTC connection that then fails once take up escalates.

I suspect the service is provided on a best efforts basis as with ADSL too. Shout too loud and they will remove it etc.. After all they can always put you back onto ADSL.

In reply to a post by WWWombat:

Don't forget to do the "further diagnostics" page. That holds the IP profile data, from which the sync speed can be calculated.

The sync speed should be available on the Vigor 2860 line stats page.