FTTC understanding sync speeds and overheads etc.

I had FTTC installed yesterday and I have to say that I'm most impressed. 73mbps down and the top end of 17mbps up with a 10ms ping.

I have the ECI modem, so I need to make a UART connection to hack it. Not wanting to risk going without internet should something go wrong, I've ordered another ECI modem to try it on (I chose to stick with ECI as they seemed cheaper, and it was my understanding you get an ECI modem if the VDSL kit in the cab is ECI).

Plus net report my line speed as 78mbps, estimates gave me 69mbps.

So what can I expect my sync to be? Can I expect it to get better or worse as time goes on? Just trying to familiarise myself with how FTTC works really.

Probably not get better, you are pretty much at the top end now, so down is more likely, particular as cross talk increases as more people take FTTC in your area.

If the IP profile is 74,000 Kbps then the sync speed will be 74000/0.962= 76923 Kbps

It is very much a case of letting the connection run and see how if copes.

In reply to a post by mikehiow:

So what can I expect my sync to be? Can I expect it to get better or worse as time goes on?

Worse, probably. I don't mean that in a snide way it's just that most people who originally syncd below 80 have seen a significant drop as time went on. It's probably due to increasing cross-talk as more people adopt the service. I started off syncing at 80 (went live last May) but am now down to 71. Some of that might be due to the cold weather because speed usually is better during the summer.

Edited by Andrue (Thu 04-Apr-13 12:55:51)

In reply to a post by MrSaffron:

Probably not get better, you are pretty much at the top end now, so down is more likely, particular as cross talk increases as more people take FTTC in your area.

Oddly, when I had FTTC installed Sept 2010, my speedtest results were 27-29 ish, and they stayed that way until last autumn, when I started noticing 32-34 was typical.
I don't run speed tests that often though, call me old fashioned, but I just sort of use the connection, and don't worry about every last Mb/s

(line length at least 1km)

Mine started at 66 Mbps and now it's down to 54 Mbps which is now quite a bit below the estimated speed. Mind you it hasn't been helped by the local council replacing the sodium street lights with LED lights.

Run the BTW performance test and when it finishes click the Further Diagnostics at the bottom. That will give you a text box with the downstream IP Profile. Mutilply that by 1.033 and you get you approximate sync. (IP Profile is 0.9679 of sync).

In reply to a post by simon194:

Mine started at 66 Mbps and now it's down to 54 Mbps which is now quite a bit below the estimated speed. Mind you it hasn't been helped by the local council replacing the sodium street lights with LED lights.

Mine is more in line with initial estimates and my actual throughput hasn't dropped as much as the sync. That tends to suggest that I've mostly lost speed that was being lost to errors anyway.

When I had an 80Mb/s sync I was seeing 73Mb/s throughput.
On 71Mb/s sync I see 67Mb/s speeds.

That change may have been the move from Profile 8c to Profile 17a in April 2012 which added more MHz to the signal used and benefited people

In reply to a post by simon194:

Mind you it hasn't been helped by the local council replacing the sodium street lights with LED lights.

Why should LED street lighting impact VDSL (or ADSLx) performance?

I can see why "noisy" high pressure sodium "may" have an impact but LED should be relatively electrically "quiet".

And LED should be massively more energy efficient and last much, much longer.

In reply to a post by Andrue:

In reply to a post by mikehiow:

So what can I expect my sync to be? Can I expect it to get better or worse as time goes on?

Worse, probably. I don't mean that in a snide way it's just that most people who originally syncd below 80 have seen a significant drop as time went on. It's probably due to increasing cross-talk as more people adopt the service. I started off syncing at 80 (went live last May) but am now down to 71. Some of that might be due to the cold weather because speed usually is better during the summer.

There's supposed to be enough bandwidth at the cabinet for every connection at 80mb. Obviously the copper twin reduces the eventual speed to the home.

Not the LED but the switching circuitry that drives them:

http://www.emcrules.com/2011/07/radio-interference-f...

Green by name, green by nature! What are the LED drivers like in terms of noise?

cheers, a

In reply to a post by wobblyheed:

There's supposed to be enough bandwidth at the cabinet for every connection at 80mb. Obviously the copper twin reduces the eventual speed to the home.

The problem isn't the capacity, it's the cross-talk from adjacent copper. Which gets worse as more circuits go live on the cabinet. With luck, in a while Vectoring will be introduced and everyone should benefit from that.

At the weekend I'll dig out the quiet line plots for a night time resync before the LED lights and after.

am I going to sound like a rotten apple?

I started off at 80 and had a snr margin on that sync of almost 20db, since then something has clearly gone wrong. I have been down to a 46mbit sync during this week, I got told I have had a IP profile of 27mbit meaning at some point I had a sync as low as 30mbit or so. Prior to the last 3-4 weeks tho my sync was stable around the 69-73 mark with occasional huge jumps up to nearly 90 on attainable. So even prior to the recent few weeks I had lost a ton of signal strength.

Engineer is due this afternoon, so I may report back on his/her findings.

Edited by Chrysalis (Fri 05-Apr-13 07:30:17)

In reply to a post by Chrysalis:

am I going to sound like a rotten apple?

No, just lucky on two counts, one to have a very good speed in the first place, and two to have such big erratic changes that you're ISP and BT accept that there is a fault.

Mines always been low for my distance and does fluctuate, usually between 40 to 45, when my estimate was 57. Don't think I'm affected by cross talk as I was first on the cab last August and it's been the same all the time.

and there was me thinking you can drive a LED with 5V DC

In reply to a post by yarwell:

and there was me thinking you can drive a LED with 5V DC

.... and a series resistor to limit the current unless it is built in!

.... but the series resistor wastes a lot of power unless there is an active driver circuit which takes us back to the beginning .........

You can certainly run Leds on 12 & 24 volt, in fact most of the led lamps we fit to the trucks are dual voltage, not a lot inside them either.

And how many streets have a DC power supply, i.e. the switching is what is doing the AC down to DC and step down of voltage.

Time will tell if the engineer accepts.

Edited by Chrysalis (Fri 05-Apr-13 13:37:14)

Right,

My IP profile is 77.44, which would seem to indicate that I have a sync of 80mbps.

If that's the case though, why don't I see download speeds closer to 77mbps?

how much interference does a simple bridge rectifier produce ? Is switching the right choice of technology for a simple power supply if the result is excessive RFI ?

Firstly it's a contended service, you're only guaranteed a maximum of 30M from the cabinet to the exchange although if lightly loaded this isn't an issue. For BTW based services "For End-User Accesses selecting the "Standard" traffic option, BT Wholesale will provide a throughput level of at least 12 Mbit/s for 90% of the busy 3 hour period"

Secondly you have the overheads and layers to consider - "For example, if the reported downstream VDSL2 data rate is 40,000kbit/s and the IP packet size is 1500 bytes(i.e. Ethernet frame size at End User LAN is 1514 bytes) the maximum throughput achievable is actually 39,178kbit/s (when measured at the EU LAN i.e. no VLAN header, but including Ethernet header)." (Openreach FTTC SIN)

If that 1500 IP packet carries a useful payload of 1460 bytes of application data (wrapped up in various headers) then 39178 becomes 39178*1460/1500 = 38133 or with PPPoE possibly 1452/1500 giving 37924 or in your case 75848 of useful data throughput.

A long time since I did any EMC formal testing...

prefer ye olde 50 Hz comforting hum of a transformer

engineer I guess normally would have just left, as he been telling me 80mbit sync not possible on FTTC at all (justify current situation), but he got a call whilst here from high level of BT telling him to persist, so he is off to cabinet now and ringing me back in a bit.

He doesnt seem keen to do a pair swap.

Edited by Chrysalis (Fri 05-Apr-13 14:30:38)

In reply to a post by mikehiow:

If that's the case though, why don't I see download speeds closer to 77mbps?

There are other overheads that mean you won't see more than 75-76 at best.

Edited by kasg (Fri 05-Apr-13 14:48:34)

In reply to a post by yarwell:

how much interference does a simple bridge rectifier produce ? Is switching the right choice of technology for a simple power supply if the result is excessive RFI ?

I used to have LED lights in my kitchen but they obliterated DAB reception so they went in the bin and were replaced with a flourescent strip light.

Nearly all the lights in our house are now LED, with no issues whatsoever that I'm aware of.

Edited by RobertoS (Fri 05-Apr-13 15:35:10)

yes highest I ever seen I think was in the 76 range.

In reply to a post by Chrysalis:

engineer I guess normally would have just left, as he been telling me 80mbit sync not possible on FTTC at all (justify current situation), but he got a call whilst here from high level of BT telling him to persist, so he is off to cabinet now and ringing me back in a bit.

He doesnt seem keen to do a pair swap.

So, how did you get on Chrysalis?

BTW, did you see any changes in your Hlog graph (especially a 'V' shape) when your sync speeds dropped right down, or were you using the ECI modem at the time?

hlog and attenuation didnt change, there is no active faults found on the line (during the test).

Case is currently not closed, they said will book a 2nd engineer next week but they didnt book it yet so I am not convinced. The first engineer was supposed to report back to them but he didnt, I couldnt get him to do a pair swap,

It does seem currently that its probably very bad crosstalk. on the sync speed but I have also outages.

The engineer wasnt particurly great either badly trained or deliberatly trying to mislead me, claiming the 80mbit sync I had originally was not possible as requires FTTP.

Edited by Chrysalis (Sat 06-Apr-13 01:42:44)

Considering we know EXACTLY what the Huawei line cards are capable of (in controlled conditions) its pretty worrying that the engineers are so uninformed.

For the record, I also have mostly LED lighting in the house but am getting a 99Mbit sync on my Digital Region line. Although they are the ones that screw straight in with their own PSUs in each one, rather than having a single bigger PSU running a bunch of them.

Edited by alexatkin (Sun 07-Apr-13 23:19:26)

The engineers doing the domestic installations and maintenance don't have anything to do with the FTTC cabinets or what's inside them.

All they have to do is jumper the lines from the exchange and the premises for the specific customer to the patch panel installed in the PCP, to make a loop to the FTTC cab and back.

Most of our LEDs are GU10 5w spots which have inbuilt drivers. 23 of them. Three 8w E27s in one fitting and an 8w B22 the same. Something else with a similar legacy fittting in a table lamp - I forget what. Plus a 1w bedside lamp (a waste of ~£5 from Sainsbury's).

The bathroom has two banks of 3 assemblies each on a common dimmable driver, (no dimmer switch installed), each assemby having 3 x 3w in a recessed tilt housing. So 54w there.

As I said, no FTTC problems being caused so far as know.

claiming the 80mbit sync I had originally was not possible as requires FTTP

perhaps he meant you can only have an 80M IP Profile on FTTP.

maybe but I did say clearly connection speed and that throughput was around 76mbit.

Then there is the bit where he asked me what my estimated was and his JDSU was 11mbit below the estimate which he didnt bat an eyelid at.

In reply to a post by yarwell:

claiming the 80mbit sync I had originally was not possible as requires FTTP

perhaps he meant you can only have an 80M IP Profile on FTTP.

Which you can't either.

In reply to a post by RobertoS:

The engineers doing the domestic installations and maintenance don't have anything to do with the FTTC cabinets or what's inside them.

All they have to do is jumper the lines from the exchange and the premises for the specific customer to the patch panel installed in the PCP, to make a loop to the FTTC cab and back.

Most of our LEDs are GU10 5w spots which have inbuilt drivers. 23 of them. Three 8w E27s in one fitting and an 8w B22 the same. Something else with a similar legacy fittting in a table lamp - I forget what. Plus a 1w bedside lamp (a waste of ~£5 from Sainsbury's).

The bathroom has two banks of 3 assemblies each on a common dimmable driver, (no dimmer switch installed), each assemby having 3 x 3w in a recessed tilt housing. So 54w there.

As I said, no FTTC problems being caused so far as know.

In line with your comments.

All my incoming circuits end up in a small purpose built loft above an upstairs hallway. In there are the various LTUs for analoig and anlaog + VDSL; the modem; router; additional WAP; burglar alarm control; and a few other bits.

The lines come through the adjoining attic and into the floor space of the loft, That floor space has a few power cables - all low power usage running through along with several high quality GU10 downlighters which use LED lamps with inbuillt converters. My VDSL line passes within maybe 400-500mm of a couple of the lamps and I see no interference from them.

Previously when I had ADSL I did put the incoming line close and noticed a drop in speeds.

But I have seen cases where there have been some problems caused by LED converters - often the type which are remote from the LEDs.

In reply to a post by yarwell:

how much interference does a simple bridge rectifier produce ? Is switching the right choice of technology for a simple power supply if the result is excessive RFI ?

Down to good design and price.

I will build you a 240AC to 5V DC switching converter that will obliterate everything! And cost £2 to produce

Give me four times as long and I will design ones which will be almost "silent". And cost £3 to produce.

Unfortunately everyone want everything cheap - they decide on price and NOT performance in too many cases.

In reply to a post by MHC:

Unfortunately everyone want everything cheap - they decide on price and NOT performance in too many cases.

Sounds like the UK broadband market..

Exactly ... whether it is the overall price where the ISPs have to outsource to India to cut a few pence of the costs, or supply bottom end modems/hubs/routers.

In reply to a post by RobertoS:

Which you can't either.

Dunno, haven't seen it defined precisely enough yet.

IP Profile is 96.79% of sync. There are no sync's above 80Mbps.

true

In reply to a post by RobertoS:

IP Profile is 96.79% of sync. There are no sync's above 80Mbps.

FTTP ?

I think I got in a tangle with your introducing a surmise about what he meant as opposed to what he said. Hence my reversion to the Subject of the thread and thinking we were talking about FTTC IP Profile.

Reverting to your original

In reply to a post by yarwell:

claiming the 80mbit sync I had originally was not possible as requires FTTP

perhaps he meant you can only have an 80M IP Profile on FTTP.

even by your own argument here and in this post the engineer would have been talking twaddle if he had meant what you suggest. Which is what Chrysalis was saying and you were replying to.

The engineer wasnt particurly great either badly trained or deliberatly trying to mislead me, claiming the 80mbit sync I had originally was not possible as requires FTTP.

that depends how the 80M FTTP product is defined, which I don't know. It could be for all I know an 80M IP profile, which was where my loop started.

my line is been monitored for a week now as was stable over the weekend, no engineer was booked as I expected.

DLM also swapped me back to fast path which has my sync just below my estimated speed now so if the line is stable between now and then I think they will close my case, even tho the engineer visit was a bit of a sham.

case is as good as closed now unless the stability is compromised again but has been stable on fast path all week.

I am on a 62mbit IP profile, originally was the full 77.43. Also I am not even seeing variations on noise margin and attainable now. So its possible they did do something but just not saying so. BT have told me tho as far as they concerned the 77.43 down to 62 profile is not something to investigate even tho my line could do 110mbit on day 1. So I havent got anywhere on that.

So in my case I guess I just wait for vectoring, that will prove if its crosstalk or not.

The stabilisation may also be related to the fritzbox which I put in since sunday, it has incredibly low errors compared to the hg612.

I am still in 2 minds, realistically 60mbit is pretty good for the uk, and for me it is good enough. But I have a mind that wants the best and it has bugged me the lost attainable but the main reason I did contact them was the long outages which have stopped.

Also my line is on a 74mbit banded profile, which is irrelevant currently as its synced below that.

However if it falls down nearer to 40mbit, I may downgrade, whilst 20mbit upload is nice to have I dont actually upload often, and I wont be paying for missing performance.

Edited by Chrysalis (Thu 11-Apr-13 10:05:45)