How to work out max cable modem connection speed

I know a fair bit about BTs lines and connecting kit, the router/modem syncing with the exchange at a given speed and which decides what sort of reliable speeds you can get through the phone line.

Do cable modems have some equivelent? More specifically Virgin Media 50mbit cable modems. I've look at my stats, there's 3 of 4 channels active, they're rated at

Modulation 256QAM 256QAM 256QAM N/A
Symbol Rate
(Msym/sec) 6.952 6.952 6.952 N/A

From what I've read the symbol rate depends on (is a multiplier for) the modulation width (?) i was just wondering if I was on the righ track with this, and how to actually turn this in to realisitc download speeds.

What I'm getting at, is there any way from the VM router stats to see how fast your line can operate, i guess it's only set to operate at a max of the package you're on, but for example if you cannot reach full 50mbit download on the VM 50mbit package, could you basically tell from the router if that was to do with the coax run from your home to the cabinate being a bottleneck.

I'm interested in finding out what sort of speeds FTTC can provide maximum, especially considering the impending 100mbit upgrade on the horizon, in terms of line quality the coax run will be the bottleneck but Im wondering at what sort of speeds, I know VM have tested up to about 200-300mbit when implimenting it, after speaking to the engineers who put my line in.

In reply to a post by anon123456:

I know a fair bit about BTs lines and connecting kit, the router/modem syncing with the exchange at a given speed and which decides what sort of reliable speeds you can get through the phone line.

Do cable modems have some equivelent? More specifically Virgin Media 50mbit cable modems. I've look at my stats, there's 3 of 4 channels active, they're rated at

Modulation 256QAM 256QAM 256QAM N/A
Symbol Rate
(Msym/sec) 6.952 6.952 6.952 N/A

From what I've read the symbol rate depends on (is a multiplier for) the modulation width (?) i was just wondering if I was on the righ track with this, and how to actually turn this in to realisitc download speeds.

What I'm getting at, is there any way from the VM router stats to see how fast your line can operate, i guess it's only set to operate at a max of the package you're on, but for example if you cannot reach full 50mbit download on the VM 50mbit package, could you basically tell from the router if that was to do with the coax run from your home to the cabinate being a bottleneck.

I'm interested in finding out what sort of speeds FTTC can provide maximum, especially considering the impending 100mbit upgrade on the horizon, in terms of line quality the coax run will be the bottleneck but Im wondering at what sort of speeds, I know VM have tested up to about 200-300mbit when implimenting it, after speaking to the engineers who put my line in.

Can't answer the question itself, but there is no relationship between the speed you get from VM cable (and the connection from its cabinet) and what you would get from BT-based FTTC.

On FTTC the only variable affecting speed, once the package is settled, is the existing copper quality and length from the existing BT cabinet.

Your service on FTTC will be on BT kit all the way, including the physical connection to your house. Virgin cable is on a Virgin connection to the house.

The way the broadband works is also completely different. FTTC is VDSL2. As I understand it, Virgin is variations of DOCSIS, but I have no idea what that is.

I'm not really interested in comparting FTTC between VM and BT, I'm more interested in what sort of bottleneck does the copper/coax run from the cab to the home presents in these FTTC setups.

I became interested in this after the TBB news items on Fludiata calling out BT and VM on advertising their "fibre" connections when its not really FTTH but only FTTC. In the comments I made the argument that the coax run is not a significant bottleneck in these runs because it's so short and that most people could achieve the maximum speed on their line of 40mbit/50mbit of these services from BT/VM.

Another recent article on speeds showed VM 50mbit users achieving 46-47mbit on average during favourable times (at night with multi threaded downloads), showing that the coax run in this current setup probably isn't a bottleneck for the majorety of customers.

obviously some kind of measurement from the actual modem indicating low level connection speeds would give us some good info to go on.

The modem will connect at 50Mbps - if it is unable to do this is won't connect at all.

If you are seeing less than expected speeds this would be due to contention on your serving cabinet or within the network...

Your cable modem config file is what determines how fast it will go and speed is not affected by the coax to any significant degree - if it was marginal then you'd maybe get a lot of errors but mostly if the signal levels were out of spec it just wouldn't work at all.

If you don't get download speeds close to your stated package speed it isn't the coax to blame it's contention (the server, the route, or in the VM network) or a problem with your own kit.

there are a number of 8MHz (channels) in use to carry y Mbits/s each using fancy maths ie 256QAM

Google for DOCSIS 3 and have a good read. From memory 2 channels gives you 50 Mbits/s

There is a parallel with VDSL / FTTC in that the latter probably also uses about 18 MHz for 40 Mbits/s. Coax is better at carrying RF over distance though.

Maybe the question is confused by

I'm interested in finding out what sort of speeds FTTC can provide maximum, especially considering the impending 100mbit upgrade on the horizon, in terms of line quality the coax run will be the bottleneck

as BT doesn't use coax and the cabinet distance is hugely more significant.

The BT copper may even be rotting aluminium for a distance, but in general the probable attenuation is a calculable and highly relevant figure. The commonality of what effect that will have on BT FTTC and the effect coax has on Virgin from closer cabinets is zilch.

Trying to analyse what you get from Virgin in order to estimate what you would get from BT is utterly pointless.

Again I'm not comparing the 2 services, I dont care which is faster, what I care about is understanding the copper/coax run from the home to the cab.

In the article i've discussed, Fluidate have critisesed BT and VM for selling FTTC as "fibre" because it's not fibre all the way. What i'm trying to assertain is if that last copper/coax leg is actually a bottleneck with these 40/50mbit packages.

if you check the news article on this I am discussing this point with CARPETBURN in the comments.

In reply to a post by Anonymous:

In the article i've discussed, Fluidate have critisesed BT and VM for selling FTTC as "fibre" because it's not fibre all the way. What i'm trying to assertain is if that last copper/coax leg is actually a bottleneck with these 40/50mbit packages.

The problem will be uplink for all cable connections. Since cable was designed as a broadcast medium (for TV) the frequency plan allocates more towards the end user than away from. This is why the VM 50mbps download service has a 1.5mbps upload.

DOCSIS 3 gave some really clever ways of using the coax, but doesn't solve the fact they need to squeeze in quite a lot of homes on the line, all using upload.

FTTC / VDSL still has an individual connection for each house, shared back at the exchange; whereas DOCSIS is shared on the cable back to the UBR.

BE and BT (shortly) are offering Annex M on ADSL2+ which can achieve 2mbps upload, try asking VM for that

OK, I begin to see where you are coming from, but you haven't half complicated the discussion by introducing several diversionary points .

As jchamier says, the upload can be a factor, as all downloaded packets need acknowledgement in one way or another.

Attenuation is a major factor in the case of BT, as I said. Less so it seems with shorter coax runs on VM. But that affects connection speed if anything, not actual download.

Actual download speed even now on the vast majority of BT Openreach line-based circuits, even if it is LLU, is mainly constrained at the other end. In the past on BT Wholesale, the BT Centrals and Central Pluses, and more and more now MSILs, on WBC and WBMC. These are hugely expensive (to the ISP) pieces of kit and the under-provision of these on cost grounds the main reason for the big players to have traffic management.

Capacity on their internal systems and peering with the wider web are of course other factors.

Some LLU providers, whilst not subject to the Central/MSIL bottleneck, are still finding traffic management necessary to reach the desired level of profitability. TalkTalk being the obvious case.

The effect of the final bit of copper on throughput at any given sync is very minor. In particular on ADSLx and VDSLx, I don't know about VM, there is a constant stream of packets travelling at the connection speed, frequently empty. The issue is how much data gets into each packet, and that is nothing whatsoever to do with the physical connection and its characteristics. Perhaps you were not aware of this.

I maintain you are wasting a lot of effort on a pointless discussion. And I flatly refuse to read a totally different thread on what is for this purpose a totally different site in order to find out and consider arguments you are putting forward on the matter.

That's like writing letters to people with have the info sent in one to their home address and half to their work address. Or more closely by having a discussion via email but saying "look what I have written on so-and-so's blog".

Impossible. As answers to points in one place without all the points at issue being immediately obvious are not practicable.