BT’s Trevor Linney Reveals G.fast Broadband UK Trial Results and Speed

The Head of Access Network Research at BT, Trevor Linney, has revealed some results from the recent trials of its 300-500Mbps capable G.fast broadband technology in Huntingdon, Swansea and Gosforth. Overall 75% of lines were able to deliver more than 300Mbps download (30-50Mbps upload).

All of the three trials took place over a period of around 6-9 months between 2015 and 2016, which included two larger scale trials (around 2,000 premises covered) in England’s Huntingdon (Cambridgeshire) and Gosforth (Newcastle) areas. The third trial in Swansea (Wales) represented a technical trial that only covered a small number of premises around Multiple Dwelling Units (e.g. blocks of flats) and business centres.

Broadly speaking G.fast technology, under BT’s current cabinet dominated approach, works in a roughly similar way to the current ‘up to’ 80Mbps Fibre-to-the-Cabinet (FTTC / VDSL2) service. This means that a fibre optic cable is run to your local Street Cabinet and then the remaining copper run to your home uses G.fast to deliver the service.

However the trials also made use of a more expensive Fibre-to-the-distribution-point (FTTdp) style deployment, where the fibre optic cable is instead run to a smaller node that could be installed inside a big building, underground or on top of a pole. The remaining line would often then go over a much shorter run of existing copper cable to reach your home, which could produce faster speeds.

Most of the PR pictures from the trial showed the FTTdp based deployment, although improvements to the technology mean that BT can now achieve much the same performance by rolling out from Street Cabinets instead (cheaper) and this is the approach that will now dominate.

Sadly Trevor Linney, who was speaking at the recent G.fast Summit 2016 in Paris, doesn’t provide a detailed breakdown of the data so that we can see how each approach fairs. Never the less we do get some useful stats, courtesy of Dave Burstein (G.FAST NEWS) who attended the event.

Copper Line Length Distribution (Trials)
Lines Shorter than 100 Metres = 38.4%
Lines of 100-150 Metres = 35.3%
Lines of 150-200 Metres = 13.7%
Lines of 200-250 Metres = 8%
Lines Longer than 250 Metres = 5%

As anybody with a basic understanding of the technology will already know, longer copper lines mean slower speeds due to signal degradation over distance (interference etc.). G.fast tries to push extra data down the line by using more spectrum (G.fast 106MHz+ vs VDSL 17MHz) and power etc., but this in turn only works best over much shorter distances.

Last year BT showed that it was theoretically able to deliver 300Mbps over a 350 metre copper line using G.fast (here), although the trials indicated that Openreach’s current systems are not yet able to achieve this goal.

Recorded Download Speeds (Trials)
Speeds Greater than 300Mbps = 75%
Speeds of 200-300Mbps = 12%
Speeds of 100-200Mbps = 10%
Speeds Less than 100Mbps = 3%

At this point it becomes even easier to understand why the final G.fast product will initially come in two flavours: 1) ‘up to’ 330Mbps download (50Mbps upload) and, 2) ‘up to’ 160Mbps download (30Mbps upload). Aside from 160Mbps being double the 80Mbps FTTC / VDSL2 maximum of course.

The 160Mbps option should be achievable by the vast majority of people within the service’s initially expected coverage and more improvements are coming, which means that the original 300Mbps at 350Mbps expectation should still become viable. Linney notes the following expected enhancements.

Future G.fast Developments
* Enable higher bits per tone (12>14)
* Improve the receiver sensitivity (<-150dBm/Hz)
* Increase the transmit power (4>8dBm)
* Optimise the frequency usage with VDSL
* Increased vectoring group sizes (>96)

BT currently aims to make G.fast available to 10 million premises by 2020 (roughly 40% of the UK), with “most of the UK” likely to be done by 2025 (this probably equates to around 60% coverage).

The operator did originally suggest that its top download speed could also be upgraded to 500Mbps by 2025, but this aspiration was set when they envisaged a more FTTdp orientated roll-out and only time will tell if they are still able to achieve this (it’s viable on very short and stable runs of copper in the tens of metres, but that seriously limits its coverage).

Now we just need to know how much Openreach will charge ISPs and the new pilot should soon give us a good indication.