Hi again Paul and thanks for not giving up me yet!

The modem is Version 3B and it is connnected 24/7....no power supply issues that I've been aware of.

When it was first installed I did a speed test about 5mins after (and then several more in the next few hours) and the max profile I ever saw was somewhere around 22500k....surely DLM wouldn't have interupted within the first few minutes of connection?

I have a engineer booked for tomorrow pm so will ask the appropriate questions and report beack here in due course.

One very worrying discovery this afternoon though.....I used the BT DSL checker to check the estimated speeds in the adjacent street (we are on 4 house development off that road about 30m away) and the estimated speeds were "20.4 to 28.5"....Hmmmm, would tie in with the max 22.5 I saw? Will be interesting to see what JSDU reports and max sync....if only around the 22.5 mark then I guess our 4 year old development has thrown up a curveball hence the over estimate. Seeing that nobody seems to have reported getting less than their estimated speed, I wonder if I might be able to use my experience as some kind of "barginning tool"?!?

Hi Stranex,

In reply to a post by Stranex:

Hi again Paul and thanks for not giving up me yet!

No worries. Some of us are all in the same boat. The least we can do is try to help each other out from time-to-time.

The modem is Version 3B and it is connnected 24/7....no power supply issues that I've been aware of.

When it was first installed I did a speed test about 5mins after (and then several more in the next few hours) and the max profile I ever saw was somewhere around 22500k....surely DLM wouldn't have interupted within the first few minutes of connection?

It COULD do, especially if it saw incredibly high error count bursts.

I have a engineer booked for tomorrow pm so will ask the appropriate questions and report beack here in due course.

Fingers crossed for the "right sort" of engineer turning up & for a (hopefully) successful outcome, but don't hold your breath.
(I'm afraid I have become rather cynical over the last few months)

If the engineer REFUSES to let you see his JDSU stats for YOUR connection, just ask him for his/her Controller's/Manager's direct telephone No. & discuss it directly with them.
IMO it is completely unreasonable to hide connection stats from users who actually pay for a service.

I had one engineer who refused to let me see the stats, supposedly on the grounds that he could get sacked for showing me.
What utter BS. I wish I had contacted his Manager at the time.

One very worrying discovery this afternoon though.....I used the BT DSL checker to check the estimated speeds in the adjacent street (we are on 4 house development off that road about 30m away) and the estimated speeds were "20.4 to 28.5"....Hmmmm, would tie in with the max 22.5 I saw? Will be interesting to see what JSDU reports and max sync....if only around the 22.5 mark then I guess our 4 year old development has thrown up a curveball hence the over estimate. Seeing that nobody seems to have reported getting less than their estimated speed,

Apparently, BT will be updating speed estimates 28th/29th January to take the 80Mb/20Mb roll-out speeds into account.

I wonder if I might be able to use my experience as some kind of "barginning tool"?!?

Stranger things have happened............


Paul.

Edited by deleted (Wed 18-Jan-12 19:11:19)

The only way you could keep an eye on your own connection stats would be to unlock the modem, but that is an unofficial method, probably in breach of your terms & conditions.

If you do intend to unlock a modem, I would suggest obtaining a spare one from Ebay, or at the very least wait until after the engineer's visit

If you did acquire access to an unlocked modem and if you are a Windows user, the links in this post may be of some assistance to you over the coming weeks/months:-

Unlocked modem Stats post


Paul.

In reply to a post by Bald_Eagle1:

If (when the engineer visits) the JDSU reports a mac sync of 27.8 as per BT speedchecker then why aren't I seeing that at my end? Surely (aluminium or not) whatever the JDSU sees is what I should be able to achieve?

Your connection may have re-synced at a lower speed since installation, possibly due to errors/DLM intervention?
The engineer will confirm your connection's speed on arrival, but it does involve disconnection of the modem & syncing with his JDSU, which in itself may force a higher or lower speed.

The only way you could keep an eye on your own connection stats would be to unlock the modem, but that is an unofficial method, probably in breach of your terms & conditions.

This is really to the OP, but this is the right bit of conversation to quote...

The engineer would (probably) connect his JDSU instead of your modem, and check the state of the connection on there - which makes this the one occasion you can get a "window" on those line statistics that stay hidden within the standard (un-hacked) modem.

In particular, this will let you see what kind of rate the errors clock up at.

BaldEagle1 is right that the JDSU might sync at a different speed, but it will only do so under the limits of the setup allowed by DLM - so it will be very similar before the engineer orders a reset. Once the engineer has reset the line, then gettign the line stats becomes even more important (especially the errors), because that will setermine whether you line will be immediately throttled again.

It still looks like you need the engineer to actually do something to the line. Perhaps by replacing some cable, or re-making some connections. Everything else we've talked about is *interesting* to know, but isn't going to actually fix anything - it still needs the man with the pliers

One engineer told me that if anything, the JDSU reports a slightly lower sync speed than most modems (something to do with resistance within the JDSU itself).

That fits with my observations on the new line. The JDSU gave my line a top speed of 85Mbps; the modem reports 90Mbs. They're closer on upstream though (24 vs 25).

I will be sure to ask engineer what distance I am from cabinet and try to take some photos but I suspect I am well under 800m from cabinet. Even if wrong, why would speed checker estimate a significantly faster speed if it knew which cab I was connected to???

The checker can only estimate based on the data held in the database at Openreach. If someone stored duff data to describe your line, then the estimator will make duff estimations.

Cheers Guys,

So basically when he turns up Fri pm (did I say Fri before? Think I may have said tomorrow!) I need to ask to look at the JDSU for sync speed and what? Error rate or similar? Does anyone have a pic of a JDSU screen?

Presumably after that, he'll hopefully go to the cab and reset the DLM and then come back and recheck the line....time for another JDSU reading?

Wombat....you mention the engineer might actually have to do some "engineering" rather than screwing modems onto walls and plugging a computer into the socket (my old man was a "proper" BT engineer......none of this reading data from a JDSU malarky!). Do you have any idea how he / I will know if this is required? Presumably, if he resets the DLM and then comes back and the JDSU reports the max sync as the same as what I'm getting, he'll just say that is all I can expect? Or would errors show up on the line indicating that "something" needs to be tampered with?

Edited by deleted (Wed 18-Jan-12 22:59:47)

In reply to a post by Stranex:

Does anyone have a pic of a JDSU screen?

Apologies....just seen the screenshot from BaldEagles earlier post

Hi Stranex,

In reply to a post by Stranex:

Cheers Guys,

So basically when he turns up Fri pm (did I say Fri before? Think I may have said tomorrow!) I need to ask to look at the JDSU for sync speed and what? Error rate or similar? Does anyone have a pic of a JDSU screen?

The only photo I have is the one mentioned previously ( JDSU Photo ). A different screen would show error rates (if they are actually present at the time)

Presumably after that, he'll hopefully go to the cab and reset the DLM and then come back and recheck the line....time for another JDSU reading?

The type of Engineer that will be visiting your home doesn't actually have access to the fibre cab where the DSLAM & DLM are located.
They only have access to the old cabinet. The broadband is connected via short patch leads from the fibre cabinet to your copper D-side in the old cabinet.

If the engineer does decide your profile needs resetting, he will phone the "Fibre Centre" or whatever it is called, for it to be reset remotely.
He would recheck your stats as soon at that has been done.

Wombat....you mention the engineer might actually have to do some "engineering" rather than screwing modems onto walls and plugging a computer into the socket (my old man was a "proper" BT engineer......none of this reading data from a JDSU malarky!). Do you have any idea how he / I will know if this is required? Presumably, if he resets the DLM and then comes back and the JDSU reports the max sync as the same as what I'm getting, he'll just say that is all I can expect? Or would errors show up on the line indicating that "something" needs to be tampered with?

Yes, the errors are shown on a different screen from the one shown on the previous post.

After I had the problem on my first line - where the engineer left after telling me the sync was 40Mb, but DLM later decreased it due to the 'obvious' errors on the line - I struggled to make anyone interested in helping me get my line "better" while the actual speed was still better than the checker's predictions. My problems weren't as bad as BaldEagle's - but I sure wish I'd not let the engineer leave quite so quickly.

I thought about it a little, and decided that the only way I'd know (while the engineer was present) was not to look at the sync rate alone, but to look at the error rates (mainly CRC errors, but FEC errors too if the forward-error-correction process was turned on alongside interleaving).

So, I tried asking about that during the second install (in true 'locking the gate after the horses have gone' fashion). His display showed zero errors over the 4-5 minutes we left it connected, which I took to be a rather better sign

The question that would naturally be raised is "How many errors are too many?" I don't know the answer to that - and not many people do. The part that "error rate" plays in DLM is not well understood.

However, someone found a graphic on "DLM Thresholds" and posted it into a Plusnet thread. On the "standard option" (rather than stable or super-stable), it quotes an MTBE (mean time between errors) of 60 seconds as the "red threshold" and an MTBE of 600 as the "green threshold".

I interpret that to mean: more than 1 CRC error per minute (60 per hour) is bad, and would cause DLM intervention to downgrade; between 6 and 60 errors per hour is OK; less than 6 per hour would cause DLM "de-intervention" to upgrade.

The "stable" option would set thresholds that allowed 1/10th of these error rates, and "super stable" a further 1/10th of the "stable" ones.

Note we have no indication that those thresholds belong to BT, or whether they apply to 20CN, 21CN or fibre (or even if they apply at all). I'm pretty sure the language of a "super stable" option applies to 21CN, so I figure this graphic applies to that class of products. From old 20CN experience, I can certainly say that a rate of 10 per hour is utterly stable.

Wombat....you mention the engineer might actually have to do some "engineering" rather than screwing modems onto walls and plugging a computer into the socket (my old man was a "proper" BT engineer......none of this reading data from a JDSU malarky!). Do you have any idea how he / I will know if this is required? Presumably, if he resets the DLM and then comes back and the JDSU reports the max sync as the same as what I'm getting, he'll just say that is all I can expect? Or would errors show up on the line indicating that "something" needs to be tampered with?

The $64,000 question. The JDSU is the thing that tells him what the line is capable of; unfortunately the policy of DLM "not intervening" for the first 48 hours means that just looking at the sync speed on there *during an installation visit* is almost a waste of time. To me, it allows the engineer to sign off & walk away under false pretences.

IMO they need to be taught to look at the error rates, and have some guide for knowing what are "too many". If your engineer doesn't know to look at these, and doesn't know how to interpret the numbers, then you might need to prompt him about whether X sounds high.

But all that is merely a "trigger" to get them to investigate the line properly - to determine if there is a fault in a cable somewhere, or in a joint somewhere. There are tools to help them there (some of which are line testers situated at the exchange), but I'm not an expert on that part.

I know that when the neighbour's had their FTTC install done, the engineer had his head poked down the manhole cover at the nearest distribution point - looking for the joint to try to fix a problem. He said they'd sync'ed at 40Mbps, so (in hindsight) I assume he was investigating a problem with the error rate.

What would I expect to see when the engineer turns up?

If the engineer plugs in the JDSU before resetting the DLM, I'd expect you to see the current sync speed, and an error rate within an "OK" range - hopefully somewhere near the numbers I mentioned earlier, but you might have to play it by ear.

If he orders DLM to be reset, then the line should resync as fast as it can (IIRC, you said this was at 40?), and this should be visible on the JDSU - as would a higher rate of errors.

Just a point of information re " unfortunately the policy of DLM "not intervening" for the first 48 hours means that just looking at the sync speed on there *during an installation visit* is almost a waste of time". I agree with the sentiment, but it isn't 48 hours. SIN 498:-

On the first day of operation, DLM will intervene if severe instability is detected. Otherwise, DLM will wait until the day after provision before intervening, provided that the line has been trained up for at least 15 minutes during the preceding day.

I think a JDSU can also run a TDR test. Not sure if the engineer has to plug in a different module though.

From what I can gather, most engineers still prefer to use the older 'Hawk' or 'Mole' testers for TDR testing, but I believe these are being phased out in the name of progress?

The only time my connection actually had a TDR test (using a 'Hawk'), an OBVIOUS high resistance problem was identified at a distance "somewhere near the cabinet".
All other previous "standard" tests until then had resulted in LTOK!!!!!!


The engineer dealt with a double-jumpered connection INSIDE the cabinet, which appeared to lower my connection's DS attenuation value from 30.2dB to approx 21-22dB.

However, that engineer did NOT re-test the line to check for anything else less obvious (he actually REFUSED to) as he said it wasn't necessary - he had fixed the fault.

The very same afternoon, my telephone started crackling again, resulting in modem disconnections.

Surely, it would be far more cost-effective if FULL tests were carried out at the outset, rather than keep having another "dabble" at things.

e.g. around 100m of drop wire was replaced in November.
It reduced my error counts a little, but I still haven't got my higher speeds back & I still suffer from random disconnections & error count bursts of 2Million or so at a time.