snr margin/IP profile

[#1]

Hi,
Anyone out there help clear up some misunderstandings I have regarding snr margins/IP profiles etc.
Using the DMT software I can ‘adjust’ my sync speed to a higher figure.  I then often find that on switching the pc on the following morning that my speeds are down and obviously my router has re-synced during the night when my snr margin will have dropped below a level at which the router can operate.
I’m trying to find an optimum sync speed that whilst giving me the fastest throughput will not cause my router to re sync every couple of days. 
Yesterday my sync speed was 5440 and during the course of the evening my snr margin dropped to ~2.5 dB.  I fully expected the router to have re synced during the night but it hadn’t.  This morning using my old friend DMT I reduced my sync speed to 5120 kbps and my snr margin went up a couple of dB’s.  It’s now at ~ 9 dB , that should equate to about 4 or 5 dB during the evenings which ‘should’ give me a consistent sync.
Couple of questions. 
I notice on the Kitz website they say that any sync speed ‘tween 5120 kbps and 5664 kbps should give an IP profile of 4500.  Would I therefore be right in assuming that reducing my sync speed from 5440 to 5120 should not affect my IP profile?
And another question – It seems quite a co-incidence that when my router was being re-synced it dropped to 5120 – which is ‘just’ inside the 4500 IP profile ‘band’.  I’ve noticed this happen before so I assume it’s not really a co-incidence at all – does the router have certain predetermined sync figures it jumps to when it re-syncs – consistent with the line conditions - or does BT have software at the exchange that determines this? 
Sorry I’m not making myself very clear – probably still a bit too much of the Xmas spirit coursing through my veins 
Thanks for your input and apologies for long posting,
John

[#2]

Suddenly find myself thinking - is it the DMT software that is determining these sync speeds, e.g 5120?
Hadn't thought of that earlier - would make sense - DMT might be 'offering' me the highest sync speed possible consistent with staying within the IP profile I'm on at the moment.

[#3]

Quote from: PNnewbie

And another question – It seems quite a co-incidence that when my router was being re-synced it dropped to 5120 – which is ‘just’ inside the 4500 IP profile ‘band’.  I’ve noticed this happen before so I assume it’s not really a co-incidence at all – does the router have certain predetermined sync figures it jumps to when it re-syncs – consistent with the line conditions - or does BT have software at the exchange that determines this?

Interestring question - I don't know the answer but yes the resynch rate seems to use predetermined steps rather than an algorithm. This is almost certainly done at the exchange as your router doesn't create the synch speed - the exchange equipment does.
All DMT does is to offset the default noise margin not set the synch speed

[#4]

You're right of course - I tend to use these terms rather loosely - and often incorrectly.
>> Interestring question - I don't know the answer but yes the resynch rate seems to use predetermined steps rather than an algorithm. This is almost certainly done at the exchange <<
Someone may 'appear' out of the blue who maybe knows more about these things...
Meanwhile I have to say that I'm very happy with the speeds I'm getting and the overall PN experience. 
I may 'tinker' with putting some ferrites over BT's wire where it comes into the house to see if that might reduce the noise on the line - though I'm not very hopeful...I think I've tweaked my system to the point where I probably won't get much more out of it.
John

[#5]

A handshake between the router and the DSLAM at the exchange take place everytime you reboot or resync the router. The line condition (noise, line length) are the most important factors in determining the sync rate.
The DSLAM assigns the sync rate to the Router after the handshake. this is why someone may get different sync rate using different router on the same line because the handshake relies on readings from both the router and the DSLAM to assign the sync rate.
Having done a quick test on your line, it looks like you've got an internal wiring problem or faulty filters causing the SNR to randomly drop bellow 6dB
Copper line test:
Test Outcome : Fail
Description : FAULT - Loop
if the SNR wont go up (6dB or more) at the main BT socket with a new filter and all other phones/skybox/fax unplugged then you might need to give your home phone provider a ring to run a line test

[#6]

RamiDoodle – thanks for your input.
>> Having done a quick test on your line, it looks like you've got an internal wiring problem or faulty filters  causing the SNR to randomly drop bellow 6dB
I don’t ‘think’ I’ve got a problem.  The SNR drops below 6dB because I’ve ‘tweaked’ the system using DMT…
I was merely posting to try and clear up some things in my mind – the relationship between sn margin and IP profile etc. 
My sn margin normally sits at around 11dB during the day, dropping to ~ 5dB during the evening and I get good speeds – I can synch at ~4000 – 4500.  Lately I have been using the DMT software to adjust my sn margin and by reducing it slightly I have raised my synch speed to 5120.  I’m leaving it like this for a few days to see if it is reliable, if not I’ll reduce it a bit.
My query was how the two things relate and how figures are allocated – for example I have obtained a synch speed of 5120 on a number of occasions – that’s too much of a con-incidence.  So I presume that there are certain values allocated – one of which 5120. 
Not making myself very clear I’m afraid.
Having said all that, the system ‘did’ become un-synched – and stayed that way a couple of days ago.  On PN’s advice I disconnected the router and left it for an hour then re-connected but it did nothing.  I then tried a re-boot and that fixed the problem.
Over the Xmas period my sn margin would occasionally vary by between 2 and 3 dB over a very short period of time.  It would do this for maybe 30 minutes then go back to it’s more usual +/- 0.5 dB variations.  It now appears to have settled down and I’m getting very good results:-

Test1 comprises of Best Effort Test:  -provides background information.
    Your DSL connection rate: 5120 kbps(DOWN-STREAM),  448 kbps(UP-STREAM)
    IP profile for your line is - 4500 kbps
    Actual IP throughput achieved during the test was - 2133 kbps
Incidentally I always have doubts about BT’s throughput figures as immediately after PN’s speed tester says:-
Download – 4462
Upload – 379
Apologies for the long post and thanks again RamiDoodle for your input, I trust you all had a nice Xmas, keep up the good work!
John

[#7]

Hi John,
I thought Rami's explanation was one of the best simplified explanations I've seen. There are a couple of further bits worth explaining that may give you a better understanding.
Dependant on your line conditions, as explained by Rami, the exchange then sets what is called a target SNRM. This will change in 3 dB steps from 6 to 15dB, and the current setting will be determined by what happened in the 10 days training period at the start of your ADSLMax and your recent line history. You can deduce what the current target is by looking at the SNRMargin immediately after a resync not using DMT (best done in daytime hours NOT near sunrise & sunset). It is this target that determines what your sync speed will be (this goes up and down in 32K steps), but as line noise will vary throughout the day (usually worse after dark) the speed will be lower when there is more noise on the line.
It is possible that under stable weather conditions, you have a steady(ish) daytime noise level and so re-syncing to the same sync speed is not out of the question, you are also correct that providing your sync speed remains within each defined band, it will not change your profile. If however the sync speed drops to a lower band and does so for longer than about 75mins. or does it more than once in that period then your profile will drop. Trouble is because of the way the system works it can take several days to come back up again.
You seem to possibly have quite a noisey line at night with that 4 or 5 dB variation, is your line overhead to the exchange or underground?
With regard to throughput speed that can vary by the (fraction of a) second depending on what traffic is going throught your exchange and through PN at that instant.
With regard to your line and the test from Rami, it looks as though you might have a low resistance condition across your line which could be faulty wiring (internal or external) or filter etc.
The best thing to do would be, (during daytime when you have lower and more stable noise), connect your regular filter to the test socket and plug in a corded phone (your regular one if corded), leave your router unplugged initially. If you are BT line rental, dial 151 and do the automated test. If this comes up with a fault condition, repeat the test with another filter and corded phone. If it is still faulty then you would seem to have an external fault which you should report.  If there is no fault condition then reconnect the master socket front plate (which will connect your extension wiring, sockets and other filters) and repeat the test. If there is still no fault condition, reconnect your router (using DMT to sync if you wish) and repeat the test. If still no fault, then any problem may be intermittent or the original test result may have been eroneous.
Let us know how you get on. Regards.

[#8]

Excellent stuff Anotherone,

Quote from: Anotherone

You can deduce what the current target is by looking at the SNRMargin immediately after a resync not using DMT (best done in daytime hours NOT near sunrise & sunset). It is this target that determines what your sync speed will be (this goes up and down in 32K steps), but as line noise will vary throughout the day (usually worse after dark) the speed will be lower when there is more noise on the line.

Very true, street lights and heating equipments, in some cases, introduce noise into the phone line.

Quote from: PNnewbie

I don’t ‘think’ I’ve got a problem.  The SNR drops below 6dB because I’ve ‘tweaked’ the system using DMT…

I've never used DMT tools before, simply because my router doesn't support it  but changing SNR is not really something I would suggest you to do as it's part of the equation.
Glad to hear that the throughput is steady but the Loop Fault is still showing when retesting, try Anotherone suggestions if you can. they will show if its internal or external. if you want me to test it again while having the router and a corded phone at the test socket through a microfilter I'm more than happy to do so 

[#9]

RamiDoodle & Anotherone - thanks for both your long posts.
I shall go and find a quiet corner and have a think (house full of visitors at the moment!) and post on here later...
Thanks again for your inputs.
John