Fibre upload speeds and general questions

Are the PN upload speeds thus:

76 / 19.5
38 / 19.5
17 / ??

And can anyone point me to a fairly accurate distance from cab, speed estimate guide? I assume that the most likely distance in new-build estates is to map the roads. If this is so, my DfC is about 0.5 / 0.6 km.

With that in mind, I'm probably going to go for 38 / 19.5 but any input on why 76 / 19.5 is worth it (I assume my DfC will mean it will be closer to 50 meg or so).

Lastly, what are the 18-month contract get out costs likely to be at any given point?

Ta!

The ADSL2+ upload maximum is 0.8 to 1 Mbps

The faster download FTTC depends on whether you want to benefit from a bit of extra download speed. Many just opt for the cash saving.

http://www.thinkbroadband.com/guide/fibre-broadband....
The often copied and made prettier chart, and while people say the speeds are too slow they are taking into account cross talk and without other factors like copper cable thickness and which ANFP mask your cabinet is using it is all guess work until you stick a VDSL2 modem onto the line.

Don't forget the upload speeds are all "up to" as well. I am about 600m from the cab (cable length) and I get between 40-50 down sync and 13-15 up sync. I am with BT but the tech is the same with PlusNet.

On BT I pay for the 80/20 package mainly just to get a bit more on the upstream for when uploading photos.

I considered changing to PlusNet and would probably have gone for the 40/20 with them as the loss in the downstream wouldn't have mattered that much to me (only reasons I didn't was the retention deal BT gave me and PlusNet were going through a bit of a rough patch for support).

http://www.camieabz.co.uk/broadband

Scroll down to line stats. They are the best-effort stats. My ADSL2+ speeds are generally around 12Meg throughput. Yes, that list is a bit stingy on line speeds. Not sure about the external wiring, obviously, but it's CAT6 copper internally, to the wall port at least.

Hmm. I forgot about the up-to on the upload. There's always something.

Without knowing details of the cabinet and your exact address and the copper used largely irrelevant to guessing your FTTC speeds

Is there anywhere a table of (down) attenuation against expected speed (or distance) like there is with ADSL? That would be really useful.

If I don't sleep I might manage it today

Sleep? Next thing you know you'll be wanting to eat at some point as well. Don't you realise you are operating within a 19th century workhouse and things like sleep and food should be forgone in order to show your proper commitment???

The TBB data:

Downstream - http://i.imgbox.com/QtXc2q2k.png
Upstream - http://i.imgbox.com/asUGND0f.png

/had 5 minutes to kill

Edit: No attn data...just meddling with charts

Edited by camieabz (Mon 16-Mar-15 13:18:45)

If I was to declare my working hours it would dissuade anyone from entering IT as a profession

Thanks but I was after it as a function of attenuation so that when someone presents their router stats you can work out if they are actually getting a reasonable speed.

IMPORTANT: We are basing speeds on generally worst case, i.e. being as honest as possible so as not to set expectations high, i.e. impact of bad crosstalk is name of the game. In essence we expect people to be above the figures given in the vast majority of cases. The use of three power masks means you can get different speeds as the power is cut back to not kill ADSL2+ speeds in the bundle, which may explain some variations. A cunning plan is underway to increase the data we have on the speed/distance side.

Distance to cabinet (metres)	Downstream Attenuation (dB)	Estimated downstream connection speed	Estimated upstream connection speed	Cumulative %'age of premises at this distance
100m	5dB	100 Mbps	25 Mbps	5%
150m	10dB	80 Mbps	20 Mbps	10%
200m	12dB	65 Mbps	18 Mbps	20%
300m	15dB	45 Mbps	17 Mbps	30%
400m	16.5dB	42 Mbps	16 Mbps	45%
500m	18dB	38 Mbps	15 Mbps	60%
600m	19.5dB	35 Mbps	14 Mbps	70%
700m	21dB	32 Mbps	11 Mbps	75%
800m	22dB	28 Mbps	10 Mbps	80%
900m	24dB	25 Mbps	9 Mbps	85%
1000m	26 dB	24 Mbps	8 Mbps	90%
1250m	28dB	17 Mbps	5 Mbps	95%
1500m	33 dB	15 Mbps	4 Mbps	98%

If this passes the test of peer moaning, then can commit to the site pages

I don't know my downstream attenuation (don't have an unlocked modem) but at an estimated 600m line length your estimate for downstream is about the same as BTs (although I get between 40 and 50 throughput depending on what DLM is up to) and the upstream is more hopeful than BTs estimate of about 9Mb but pretty much spot on for what I actually receive.

In reply to a post by camieabz:

Lastly, what are the 18-month contract get out costs likely to be at any given point?

Full price list including early termination costs.

Takes a while to get used to the layout of the list, so make sure you don't pick up a wrong figure .

In reply to a post by camieabz:

Hmm. I forgot about the up-to on the upload. There's always something.

Yes :-/

Line attenuation (dB):  	21.1		0.0
Signal attenuation (dB):	Not monitored		
Connection speed (kbps):	57775		11487
SNR margin (dB):        	6.0		5.8
Power (dBm):            	12.6		7.1
Interleave depth:       	8		4
INP:                    	47.00		47.00
G.INP:                  	Enabled		

RSCorr/RS (%):          	0.0006		0.0077
RSUnCorr/RS (%):        	0.0000		0.0000
ES/hour:                	17.7		0.21

easiest way is to look here for your account summary https://portal.plus.net/my.html?s=0&action=account_s...

Thanks.

wow those are super conservative numbers

They are very conservative, aren't they? Better than being over-optimistic but they seem to be allowing for masses of crosstalk. When vectoring comes (if it ever does) they will hopefully be revised.

Edited by kasg (Tue 17-Mar-15 09:19:58)

Once we see how vectoring behaves yes we will.

One does not want to be accused of over stating the speeds, though we do get some strident campaigning to say the figures are still too high.

According to this list, I should get about 38mb/s download, which most of the time I do more or less, but I should be getting 15mb/s upload, I get about 4-5. Any ideas why this may be? Neighbour gets around 10 with BT, but a lot slower download.

ANFP power mask issues.
Cross talk issues.
Openreach don't like you.
Oddity in terms of some interference that is impacting an upload band only.

Getting 38 Mbps and only 4 or 5 Mbps upload is odd.

In reply to a post by MrSaffron:

One does not want to be accused of over stating the speeds, though we do get some strident campaigning to say the figures are still too high.

If you are aiming to show estimates that represent the bottom end, wouldn't it be better to give the column a title that reflected this? Rather than just an estimate? Perhaps a "minimum estimate"?

BT produce a range derived from the 20th and 80th percentiles, but fail to tell anyone what the range actually represents; because the bottom of the range is still only the 20th percentile, some people could be below the minimum...

Given that BT show a range, would it be better for TBB to also have a range, but to go on and explain why there is a range (crosstalk), and how you can expect variation?

I like the way in which the Broadband Forum's document on vectoring shows speeds vs distance - which distinguishes what people can normally get vs the worst case and vs the best case vs the vectored case.

See figure 6 in http://www.broadband-forum.org/marketing/download/mk...
Essentially, the blue crosses there represent the same kind of thing as BT's 20th and 80th percentiles. I'm not trying to say anything about vectoring here ... just that the depiction of the "normal" by the blue crosses is actually useful.

That graph represents a 17a profile that looks very similar to BT's 998ADE17, but on 0.4mm copper; it probably understates the average line capability in the UK. Unfortunately, it only goes up to 1,000 metres.

I have seen similar graphs go up to 2,000 metres, here:
http://cms.comsoc.org/SiteGen/Uploads/Public/Docs_Gl...
Pages 10-11.

Unfortunately, those graphs are for the 17a profile known as 998E17 profile - which differs from ours considerably: there is no US0, so it has limited upstream range; and 2MHz is swapped from DS to US (the 12MHz-14MHz portion), so at closer distances, the DS speeds appear low compared to the UK, while the close US speeds look high.

On the positive side, the graphs use TP100 cable, which is 0.5mm solid copper, PE insulated, multi-pair cable as used in BT's D-side. That part ought to match up fairly well.

Overall, perhaps the longer-range part of the downstream graph (1km-2km) ought to be similar to the UK experience.

and I can believe that, its certianly possible given the poorly maintained/built state of BT's local loop those speeds can be above what people get.

After all my own line did manage to drop to below 50 (on fast path) before they changed the pair, and thats on a line estimated above 70 and started with 110 attainable.

In reply to a post by MrSaffron:

ANFP power mask issues.
Cross talk issues.
Openreach don't like you.
Oddity in terms of some interference that is impacting an upload band only.

Getting 38 Mbps and only 4 or 5 Mbps upload is odd.

I will go with the third one, mainly that I do not really like BT openreach either.

I have decided to try power line adaptors again and got myself a TP-LINK one from Argos, hoping it will be better than the last one I used a few years back.
Now I get a better upload speed than I did with Wi-fi. I do not understand it, I got a great download speed with Wi-fi, naff upload, now both are good.

37 down, 9 ish up.

Very strange, unless me switching off the wrong switch and turning the modem off done anything, but I can't see it myself. I must put some tape on that switch.