wholesale checker accuracy

just wondering...

does the checker know how far the customer is from the cabinet, is it a measured figure, or is it a guess based on postcode or similar?

There is no known. Until u see on the BT FTTC Modem to tell u how far from cabinet.

Edited by adslmax (Mon 02-Dec-13 17:01:14)

so it is pretty much what our american friends might call a SWAG* then?



*Scientific Wild Assed Guess

What? That just doesn't make sense and the BT modem certainly doesn't know how far you are from the cabinet.

Please reread your posts to make sure they make sense.

In reply to a post by cookie:

just wondering...

does the checker know how far the customer is from the cabinet, is it a measured figure, or is it a guess based on postcode or similar?

It does indeed. It'll either use records of line length if you give the address or phone number, if postcode it'll base it on the property in centre of the postcode.

Please disregard his reply it's incorrect, and confusing at the very least.

The estimates are getting better, I'm guessing they are based on known information but BT will only know what info that is. Distance, line quality, allly or copper etc.

Going by recent posts I would suggest most people seem to get somewhere in the area where the two speed ranges overlap, but it comes down to distance, quality of line and any interference issues.

So the only real way to know is to try it, which is what adslmax probably meant.

Interestingly I just checked a neighbour's address. (They are on Sky). Reason being they have a serious problem, under 400kbps sync on ADSL2+!

On the same cabinet as me, it shows:-

FTTC Range A (Clean) 72.7 54.6 20 16.3 -- Available
FTTC Range B (Impacted) 62 34.9 20 11.4 -- Available

WBC ADSL 2+ Up to 3.5 -- 2 to 6 Available

Mine shows:-

FTTC Range A (Clean) 63.8 46.9 18.6 12 -- Available
FTTC Range B (Impacted) 53 29.5 18.6 8.4 -- Available

WBC ADSL 2+ Up to 4.5 -- 3 to 7 Available

given the number. Given the address the FTTC figures are the same as with the number, but the ADSL2+ is only 4Mbps.

I'm pretty sure that BT have records of the attenuation value for lines, at 300kHz - the documents that estimate the nationwide picture of line lengths give a reference back to some BT statistics based on such attenuation.

If that is true, and the estimates are based on the attenuation values, then they take into account length, diameter, metal-type and quality of copper.

In reply to a post by WWWombat:

I'm pretty sure that BT have records of the attenuation value for lines

What about lines that never had BB on them? Or even ones that never had BTw BB? Wouldn't imagine BT had attenuations for those.

In reply to a post by XRaySpeX:

In reply to a post by WWWombat:

I'm pretty sure that BT have records of the attenuation value for lines

What about lines that never had BB on them? Or even ones that never had BTw BB? Wouldn't imagine BT had attenuations for those.

My comment is based on section 2 of this Sagentia report for Ofcom. It states:

The main data required for this study was the statistical distribution of line lengths and
types for the millions of copper loops in the UK. We have obtained this data from
public domain sources.

and

A paper for NICC DSLTG by John Cook presents the cumulative distributions for BT�s
network of line length for exchange to customer and cabinet to customer in the form of
attenuations.

We know of attenuation being determined by DSL modems, and being tested by the diagnostic woosh tests. However, they aren't the only way in which attenuation can be figured - there are indeed single-ended line tests that can be performed at the exchange that give this result.

So it is certainly possible for BT to measure the attenuation of every line in the country, whether it has DSL or not. And we know that BT perform routine line tests every 24 hours too.

But whether they *have* performed tests for attenuation on all their lines is unknown.

i'm trying to anticipate what speeds i can expect when the cabinet i'm on goes live and i'm surprised at the high figures on the checker as i'm approx. 650M from the cabinet.

FTTC Range A (Clean) 77.5 59.1 20 18.6 -- 31-Mar-14
FTTC Range B (Impacted) 66.8 38.1 20 13.3 -- 31-Mar-14

from what i read on here, i'd have thought anything over about 30 would be doing well at that distance. in fact i think they have risen again today.
i know that BT can measure the line attenuation for the whole distance, but wasn't convinced that these estimates are taken from cab to subscriber meterings.

In reply to a post by cookie:

i'm trying to anticipate what speeds i can expect when the cabinet i'm on goes live and i'm surprised at the high figures on the checker as i'm approx. 650M from the cabinet.

Your figures do seem rather high Perhaps the lines in your area are newer or of higher quality than average though, the sort of thing the checker is probably aware of.

Are you sure you have the correct cabinet?

Does it have your number on it?

In reply to a post by cookie:

i'm trying to anticipate what speeds i can expect when the cabinet i'm on goes live and i'm surprised at the high figures on the checker as i'm approx. 650M from the cabinet.

FTTC Range A (Clean)	77.5	 59.1	20	18.6	--	31-Mar-14
FTTC Range B (Impacted)	66.8	38.1	20	13.3	--	31-Mar-14

My old house was 650 metres from its cabinet, on 1990's copper. With the old 8c profile (using 8MHz) it could get 40Mbps, but was reduced to around 36Mbps by DLM. Once the 17a profile was introduced, this jumped to a clear 40Mbps. We never had the option to try 80Mbps on it - but the recent estimates have been around 48 Mbps.

The estimate today is:

FTTC Range A (Clean)	68.1	50.7	20	14	--	Available
FTTC Range B (Impacted)	57.4	32.3	20	9.9	--	Available

i know that BT can measure the line attenuation for the whole distance, but wasn't convinced that these estimates are taken from cab to subscriber meterings.

I suspect not too. More likely that they measure from exchange to sub, and then subtract an amount for the exchange-cab path.

On the other hand, I've also seen something that says the estimates are really based on the cab-DP portion alone. I have no idea about the truth of this, though.

Testing the line with a TDR will tell you the electrical distance to each discontinuity in the twisted pair, so identifying the electrical position of the cabinet, the DP and the NTE should be pretty easy.

No time to dig out stuff in Google, and it's far too long since I did transmission line theory, but with some clever maths I'm pretty sure you ought to be able to get a pretty good estimate of the AC attenuation of each section of the line. I'd put good money on standard exchange/central office test kit doing this.

yes, it is the right cabinet and it is numbered as per the checker.

A paper for NICC DSLTG by John Cook presents the cumulative distributions for BT�s network of line length for exchange to customer and cabinet to customer in the form of attenuations.

BT are aware of what all the line lengths should be, looks like John Cook converted that to attenuations using a formula.

BT won't have tested the attenuation of every loop because they can't - what about the unbundled ones?

In reply to a post by mpellatt:

Testing the line with a TDR will tell you the electrical distance to each discontinuity in the twisted pair, so identifying the electrical position of the cabinet, the DP and the NTE should be pretty easy.

Unfortunately lines generally aren't one long contiguous piece of copper from exchange to PCP.

my line is now below the estimated clean range. Of course I am still trying to find out if I am the only one who's previous estimate si the same as the new impacted low estimate.

I now have the new firmware, and sync speed is down, errors up, and lots of weird data on my new graphs such as

lower attenuation on the upstream tones than downstream tone, whats that about?
higher snr strength on the upstream tones vs downstream tones, even when QLN is equal.
clear power cutback on my U1 tones, probably explanation for my 9mbit of lost US attainable.

On the old firmware I couldnt see the upstream data so didnt know this before.

Also that I have a engineer booked for HR check, but it got cancelled as openreach says is major fault in area affecting many lines but wont say what it is.

The BT checker thinks my clean rate is the full 80/20:

FTTC Range A (Clean) 80 63 20 20 -- Available
FTTC Range B (Impacted) 71.7 44.1 20 15.4 -- Available

I don't actually get anywhere near to the clean or impacted maximums. I haven't looked in a while but it's something like 65/20, or 71-ish after a power cut where (presumably) there is minimal crosstalk before all of the modems sync up.

Don't forget those estimates are for the connection speed, not speed test results which on highish FTTC lines are always several Mbps lower.

You can get a very close approximation to your connection speed by running the BT Performance Test and clicking the Further diagnostics button at the bottom of the initial results page.

If you divide the downstream IP Profile by 0.9679 you get the connection speed within a very small range. This doesn't work for the upstream, which is always 2, 10 or 20 depending on the product you are on.

The speeds I refer to are the sync speeds as reported by the modem - it has never synced at 80/20.

In reply to a post by sor:

I don't actually get anywhere near to the clean or impacted maximums.

But you are within the range of both of them.

I am also syncing below my clean rate.

But its important people dont get confused with my question.

The question is are people's new impact low estimate the same as their previous estimate before BT wholesale rolled out the latest checker.

not if they sync below or above it.

Most people on kitz forum seem to have a clean low estimate the same or very similiar to their previous estimate suggesting mine been at impacted low is unusual.

whenever i run the BT Performance Test Diagnostics i get:

"The Performance Tester is currently unable to run a speed test for your broadband connection. Please try again shortly, however if this problem persists, raise the issue with your service provider."

The diagnostics second stage, or the first test?

That was for sor. What ISP are you with when you try to run the performance test?

If your TBB profile is still correct you are on O2 LLU moving on to Sky LLU and so could never run BT Performance Test as it's for BTw based connections only.

I was wondering about the accuracy too....

About a month ago it stated on the Wholesale checker that the max I could get on FTTC was up to 68.1mbps on a download and 20mbps on an upload.

I've recently checked again and am now getting this:

FTTC Range A (Clean) 80 79.9 20 20 -- 31-May-14
FTTC Range B (Impacted) 80 73.9 20 19 -- 31-May-14

Is there any possible reason as to why this may have changed?

Thanks guys!!

Mine is:

FTTC Range A (Clean) 80 74.9 20 20 -- 28-Feb-14
FTTC Range B (Impacted) 80 68.9 20 19 -- 28-Feb-14

They have revised the estimates several times or more, the previous time they seemed very low, now they seem more inline with what people are getting.

FTTC Range A (Clean) 80 78.1 20 20 -- Available
FTTC Range B (Impacted) 80 67.2 20 19 -- Available

what is sor?

it's the second part of the test i get an error with, and i am on o2 for broadband with a BT telephone line.

sor is the member RobertoS was replying to.

As I already told you as you are on O2 LLU BB, the BT Performance Test Diagnostics doesn't work for you. The 1st part is an ordinary speed test independent of ISP but the Diagnostics are for BTw BB.

In reply to a post by cookie:

what is sor?

You mean who is sor

BT are such teases:

FTTC Range A (Clean) 80 79.9 20 20 -- 31-Mar-14
FTTC Range B (Impacted) 80 77.6 20 19 -- 31-Mar-14


Line length is maybe 50-75m or so. I wonder what the max would be without the 80mbit cap, and/or with vectoring.


Not that I'm holding my breath for the March deadline though - it's now four years of being constantly shifted back by three months.........

Edited by arfster (Thu 05-Dec-13 11:34:12)

My impacted low estimate from the checker is 2Mbps below, at 59, the previous figure. My line is syncing some 8Mbps below that.

The clean low estimate is 72.7Mbps.

I am now below my new estimate, ordered a MAC of BT, will join plusnet and immediatly open fault when I start off below estimate.

my apologies to sor

In reply to a post by Chrysalis:

I am now below my new estimate, ordered a MAC of BT, will join plusnet and immediatly open fault when I start off below estimate.

How did you get out of the 18month min term contract with BT?

remember when I recontracted to turn of the traffic shaping?

that replaced my 18 month contract with a 12 month contract.

I was also given an opt-out earlier in the year but that opt-out was temporary.

If I migrate this month I will still have to pay off approximate 2 months as the recontract was done in feb.

Edited by Chrysalis (Thu 05-Dec-13 20:17:36)

In reply to a post by Chrysalis:

remember when I recontracted to turn of the traffic shaping?
that replaced my 18 month contract with a 12 month contract.
I was also given an opt-out earlier in the year but that opt-out was temporary.
If I migrate this month I will still have to pay off approximate 2 months as the recontract was done in feb.

Nice, hope you get what you want at PlusNet. I'm still looking for my new ISP, when I leave BT in Feb, as I want static IP and preferably IPv6, but without the insane high cost that AAISP would be. Zen may be my destination as I fundamentally disagree with PlusNet's traffic management ideas. (that someone watching iPlayer is more important than someone watching a streamed HD movie from Apple, or downloading the move, or even installing an OS X update)

Because BT gave me a new number on existing line and got it wrong I had to do some research to even get FTTC.

Locally to me all the houses with the same postcode and house numbers (address search) and in the same sequences of numbers have the same speed based (bt wholesale checker) no's 1,3,5,7,9 might have 80Mbps, then no's 11,13,15,17,19 might be 76 Mbps.

Using phone numbers produces the same results (if people don't have FTTC ... I believe).

I observe the groups of numbers seems to correspond to Telegraph poles, all the houses going to the same telepgraph pole had the same speed prediction, new pole, further away, slower, new pole nearer faster. I believe the numbers are based on the telegraph poles, whether its measured by a test to the pole or based on distance to the cabinet (presumably BT have GIS system with pole positions) I don't know.

However when someone gets FTTC its all change...

Now I have FTTC my results change to my actual... indeed my phone number and my address but there is more....

Not just my house, but all my neighbours phone numbers and address searches, have been updated using my actual results to adjust those nearer and those further away i.e. one person getting a real speed for FTTC updates the database for many people.

I suspect someone else getting FTTC nearer than me would cause it to change those lines and address around it based on that line. How it would cope speeds being slower near, faster further away I don't know but its not just my postcode, the effect hits postcodes further down my road as well and near. I suspect BT understand how addresses correspond to the phone lines from the poles and their position along them.

So the speed is either based on a pure estimate where there is no nearby FTTC customer or where there is a nearby FTTC customer it cause the checker to refine the estimate over a geographical area (line related in some way ?). They are not testing individual lines without FTTC.

I might be wrong but that is my best guess based on my observations.

BTW The engineer who did my FTTC install implied that there weren't many FTTC installs near me and when I check phone numbers nearby I get a lot of unrecognised numbers ... Sky, Virgin, etc I believe. So my belief my FTTC install has effected 100+ properties isn't just me playing with checker before and after.

Net of this is, if you know someone nearby has FTTC, chances are the estimate is going to be close (assuming the cabinets are the same), if you don't it could be a lot different ... my estimate was 64/20 before, the engineer said 80/20 and I am seeing 77+/18+.

Hope this helps someone understand the estimates better.

Ignitionnet
I can remember reading that Openreach insisted that even unbundled pairs had a testhead under Openreach control so that they could test/fault the lines. Otherwise how would they know where faults are.

Some CPs want to remove this under the impression it will reduce the cost and there was something on the OFCOM site about this a few months ago.

If it is removed CPs will have to provide their own linetests to Openreach when there is a fault so that there is some idea where the problem is. I see that as a disaster as most ISP just want to pass the blame when something breaks not have to help fix the problem.

cja

I suspect you are correct, certainly once you have FTTC the results are much more accurate as are others around you when in a dense area.

Where there are long distances from the distribution point ( could be many poles to a house!) I think the problems arise. This is why we are seeing a few inaccurate estimates in rural areas a long way from the cab.

Rule we should be using is, " if there are many spans of wire to YOUR house the estimate may be higher than the actual when delivered" "If you are direct off a pole with many others the estimates are likely to be correct"