Crosstalk speculation

Over on the Plusnet forums I've suggested that the 40 -> 55 package boost by BT might (and I've tried to be tentative) cause a noticeable increase in crosstalk affects across the country. Do you think it might? I'm not talking about a sudden drop in speeds across the boards. Merely that if someone were analysing sync speeds over time they might notice a slight 'bump' in the graphs as the package is rolled out.

It's just an idle curiosity/musing thought not an attempt to warn people of the possibility

It might, but then someone moving into the area with a short wave interfering device could also. The joy of mights.

Is that a real possibility? Increasing sync speed doesn't involve increasing the line power. All it does is increase the bit loading and eat into the SNR margin. I suppose the modulated waveform will be that bit more complex too, but would that cause a measurable level of increase cross-talk?

If it does, it is probably only likely to affect and be caused by those who currently have an achievable in excess of around 35Mbps.

The higher speeds will use the higher tones and maybe slightly change the bit loading on some of the tones just below. So, if anyone is to notice any effect it could be those on an 80Mbps service.

Your post is actually quite timely as I had been thinking about this over the past few days. I suffer with a major cross talk issue from on neighbour four houses away - when their service is OFF, I have an achievable and sync about 4Mbps higher than when it is ON. Then again in the past few days I have suffered a further 1Mbps drop. I have a feeling they may have recently moved to the 55Mbps option.

I'd not expect it to drop peoples sync speed (but it might for some on the up to 76 Mbps as might see more competition in their part of the spectrum).

Without spending a few months in a lab and lots of complex mathematics that may be shown to be wrong in real world, such is the complexity of the RF world impossible to say for sure.

In reply to a post by MrSaffron:

I'd not expect it to drop peoples sync speed (but it might for some on the up to 76 Mbps as might see more competition in their part of the spectrum).

Without spending a few months in a lab and lots of complex mathematics that may be shown to be wrong in real world, such is the complexity of the RF world impossible to say for sure.

Exactly ... as I suggest in my other post.

I thought that all the VDSL2 services used the same frequency range and what actually changed was the bit loading. A 40mbps link can still be putting bits into the highest frequency slots. I suppose it might be that on some medium length lines there was no bit loading in the top frequency bins at all and going to 55mbps might mean some so far unused bins now have some bit loading.

This strikes me as all a bit marginal. A few previously unused bins in the higher range due to the 40-55 move on some lines doesn't seem to me to be likely to be more than very marginal on its impact. I would have thought the visibility of any such impact will be swamped by the general increase in FTTC take-up.

In reply to a post by TheEulerID:

This strikes me as all a bit marginal. A few previously unused bins in the higher range due to the 40-55 move on some lines doesn't seem to me to be likely to be more than very marginal on its impact. I would have thought the visibility of any such impact will be swamped by the general increase in FTTC take-up.

You're probably right. Certainly taken over the longer term the effect is irrelevant. But in the short term it just struck me that there might be a blip in affects.

To be clear where I've said there might be some effect I am not saying that it will be apocalyptic, but given crosstalk is so unpredictable, unless you run massive processing power over it (how vectoring works) then hard to say.

Unused bins mean the frequency is not used, so using more bins can mean more power and spectrum used by a specific line. In theory yes take-up is likely to swamp the small change.

Its one of those scenarios where you cannot say zero effect and quantifying effect is difficult, i.e. people love 100% type answers but you cannot always given them.

In reply to a post by TheEulerID:

I thought that all the VDSL2 services used the same frequency range and what actually changed was the bit loading. A 40mbps link can still be putting bits into the highest frequency slots. I suppose it might be that on some medium length lines there was no bit loading in the top frequency bins at all and going to 55mbps might mean some so far unused bins now have some bit loading.

For downstream:
When the modem has tones available for use on, say, a service capped at 40Mbps but the line is capable of an awful lot more, I'm pretty sure the filling algorithm will choose to load all tones to a low bit-loading, rather than highly-loading some and leaving others empty.

That means any line that gains from 40Mbps to 55Mbps will do so by extra bitloading on tones that are already transmitting. In digital terms, this gives the receiver a harder job (it has to distinguish between more/closer amplitude and phase options). But in analogue terms, the transmitted signal will still be of the same order: still a modified sine wave, still the same power.

I think the end result will be similar crosstalk.

I've just checked a couple of lines on MDWS, which are currently 40Mbps but have capabilities of 80Mbps. They all use the full set of tones, but a lot of D3 only uses 2 bits/tone.

For Upstream:
The filling algorithm gets affected by upstream power back off, so U1 can get left with unloaded tones, while tones in U2 are more heavily used.

If upstream speed caps were being increased, then you might well see more crosstalk there.