Impact of FTTC on downstream users

[#1]

Our exchange went FTTC (and FTTP for some) some months ago. I am too far downstream (2.6 km) to benefit from this, so remain on a BT ADSL Wholesale Max copper line. My download speed has been steadily dropping from 5 mbps a yr ago to 2.5 now. I've had a couple of BT engineer visits recently to look at this. The first could not find any issues, but apparently turned interleaving on without mentioning it to me. The second visit a few days later could also find no faults and ascribed the reduced speed to the fact that users in the village upstream from me were piling into FTTC and their usage was impacting on my speed. Something about the increased amount of HF traffic made it harder (slower) for the broadband signal to get to me. Is this possible/correct?

[#2]

Hi there, 

Uptake for FTTC on the same exchange should have no impact on an ADSL service. 

I've just been running checks on your line but don't see any obvious issues. Have you tried using a wired connection to see if that performs any differently? 

[#3]

Hi Adam. See the attached screen shot of results from my wired Samknows whitebox of download speed over the last year. The drop in speed coincides with FTTC and FTTP going live in the exchange that I share with my village 2.4km away - Stottesdon to be precise. The uptick in speed in Feb 2018 was unsustainable, resulting in frequent dropouts, hence the BT engineer callout and the events I describe in my original post

[#4]

I assume you are not on an EO (exchange-only) line? If you have an ADSL Max connection that connects via an FTTC enabled cabinet you may be seeing increased crosstalk - the HF (high frequency) that the engineer mentioned. If so, there's probably nothing anyone can do.

[#5]

The transmitted power levels of the VDSL2 FTTC signals are cut back on the frequencies shared with ADSL, so that the crosstalk from them should be no different than that from other ADSL lines. The rest of the frequency range VDSL2 uses, higher than the ADSL signals, shouldn't make much difference.

[#6]

The Samknows variable results are more likely to due to traffic on the network, either your LAN, or possibly in the exchange. In my humble opinion, their tests and results are poorly thought out, and seem to cause more confusion than help. 

[#7]

@Mustrum the SamKnows tests may be poorly thought out, in your opinion, but the graph does quite clearly show a drop in the average level of download speed, which according to @randomcastle coincides with the introduction of FTTC at his exchange. It could be that as a result of work being done to convert other people to FTTC, the connection has been 'damaged' - hard to tell without more information, particularly on noise levels. Full router stats might be useful, but unless a record has been kept of SNRM, we won't know if the line has worsened 

[#8]

I'm confused as to whether crosstalk is a possible issue. I read this https://www.increasebroadbandspeed.co.uk/vdsl-vectoring-arrives-on-a-rural-fibre-broadband-cabinet
but wasn't sure if it applied because I'm 2.4 km from the exchange.
There's quite a few houses potentially affected if this is the case. And the conclusion is that Openreach have worsened the service to a few people to improve it to many, when they have a solution (vectoring) available.

[#9]

It's my understanding that crosstalk doesn't affect ADSL, just VDSL. However, if Openreach have been doing a lot of work in your can, they may have disturbed something which in turn has increased the noise interference on your line which will have reduced your speed.

[#10]

Vectoring only works for VDSL2 lines and is used to improve speeds on longer lines. It will do nothing for ADSL connections.

[#11]

It appears to me that everyone has a theory so I thought I would throw mine in.  

Whether the cause is crosstalk caused by the uptake of VDSL traffic depends on the amount of broadband traffic  on any mulicore cable that carries both ADSL and VDSL data, so this is downstream of the fibre cabinet before the multicore is split out to properties. I take the point that within the ADSL frequency band that the VDSL levels are compatible but high frequency VDSL 'mush' will still be injected onto the ADSL lines. So if you looked at this on an oscilloscope you could imagine the ADSL data being swamped with high frequency VDSL noise.

The question then has to be:  how good is the modem at rejecting this out of band noise? If the answer is: 'not very', then surely this will be presented as a reduced SNR?

Others thoughts?

[#12]

My current router (Linksys xac1900) won't allow me to monitor snr, so I've just bought a billion 7800n to do this and maybe some tweaking. Thanks

[#13]

---

@Baldrick1 wrote:
The question then has to be:  how good is the modem at rejecting this out of band noise? If the answer is: 'not very', then surely this will be presented as a reduced SNR?

Others thoughts?

---

I don't think this theory is likely to be correct, or at least not a significant factor. Higher frequencies are attenuated more, which is why FTTC is no good 2.6km from the cabinet. If the theory were true, then the ADSL signals on shorter lines using higher frequencies than longer lines would probably be just as much of a problem.

I've heard a similar theory about why ADSL2+ works worse on long lines than ADSL1, based on the idea that the higher frequencies of ADSL2+, which aren't even being used on the long lines, but ignoring that fact, are somehow becoming noise on the lower frequencies. I think this theory is wrong, and the actual reason is nothing to do with any of the theoretical aspects of ADSL2+, but due to the earliest generation of BT ADSL2+ exchange equipment being not very good at it.

Also, I don't think two sudden drops in the speed support an ADSL-VDSL2 crosstalk hypothesis. I think it's unlikely that everyone switching to FTTC did it on one or other of the two days, and then the number of FTTC users remained constant for about a month. If it were crosstalk I would have expected a more gradual decline in speed as FTTC uptake increases. I suppose it's not impossible that everyone who wanted FTTC was watching and waiting for it to become available and managed to order it as soon as it became available and then Openreach did all the work at the cabinet for all those lines all at once on the same day, I just don't think it's likely.

When actually did FTTC become available in this village? From looking at that graph of the download speeds, I wouldn't say it has been steadily dropping. There are two slow periods, near Jul 17 and Nov 17, but apart from those and ignoring the few short blips, it's been fairly constant from Apr 17 until it suddenly plummeted in Feb 18.

[#14]

@ejs

Thank you for taking the time to explain why my theory is incorrect. It is much appreciated. The fact that short lines would be similarly affected totally passed me by.

[#15]

I keep reading "download speed" but we seem to be focusing on sync speed. Could @randomcastle please confirm if they're saying the sync speed has gradually halved, or has the sync stayed the same and throughput has halved?