Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[tbailey2]

mydslweb stats show other oddities that I had not noticed.  Your upstream CRC per minute counts seem to be 0,24,46 or 70 almost all the time    --I wonder why other numbers don't seem to happen.

Well other numbers do happen but you are not seeing them due to lack of pixels to plot them on for the viewing period/vs pixels available. It does appear strange I admit but looking at the raw database figures, they do tend to be around those values but they are +/- in the detail. If you zoom in to a level that shows individual minutes then you can see there are differences as shown in the zoomed attachment graph.

But there is a definite grouping around the values you mention from the actual uploaded values (US_OHFErr_delta currently as noted elsewhere and used as per BE1's recommendation). Also attached some of the the raw DB figures against time.

[NewtronStar]

looking at kitz's US errored second counts on MWS it seems fine now since 13:00 hope thats the start of line going back to normal 

[kitz]

Your upsrteam CRC per minute counts seem to be 0,24,46 or 70 almost all the time 

If you could see my upstream SNRm then its not too surprising why the errors follow that pattern.   It doesnt show up anything in the combined SNRm, but as I mentioned a few posts back, my upstream SNRm per band paints a pretty picture and far more insight of whats going on.  The 3 bands have separated and its like the blade of a saw swinging 1dB either way.   

If you think the SNRm graph looks spikey, thats nothing to what I see on the SNRm per band - thats real pretty - not 
I wonder if the SNRm bands is something that could perhaps be something for Tony to think about if he wants to develop things further.

 

Its what Id equate to consistent static noise.. then every so often theres a burst spike..  and its only the larger spikes that are showing up on MDWS.

Looks like somethings changed today and its a lot more stable.   Not had chance to look properly yet whats been going on.. and I need to look on the PC really but it appears like Ive had no errors for a large part of the day.

[kitz]

But there is a definite grouping around the values you mention from the actual uploaded values (US_OHFErr_delta currently as noted elsewhere and used as per BE1's recommendation). Also attached some of the the raw DB figures against time.

Yep I agree it does, but if you could see my upstream SNRm broken down into bands it would become clear that its also following the same pattern.  Take for eg U1 it was swinging between something like 10db - 12db in a jagged saw edge pattern.   I cant do a screen cap atm, but I think you can see the saw blade pattern starting to form from an attachment in one of my earlier posts.

[tbailey2]

If you think the SNRm graph looks spikey, thats nothing to what I see on the SNRm per band - thats real pretty - not 
I wonder if the SNRm bands is something that could perhaps be something for Tony to think about if he wants to develop things further.

Done I haven't double-checked the displayed data against the Database but looking at yours, since it appears to represent a Magnitude 9 on a Seismometer I guess it's correct 

The entries are at the top of the list under the normal SNRM entry. Hit Update if you don't see them at first on a pane or F5 to get all panes at once.

SNRM
SNRM/Band DS
SNRM/Band US

[les-70]

  I am also on an ECI cab and with no speed capping my upstream snrm's are always about 6db at sync, with my varying crosstalk the power adjusts to ensure this at sync.

 With speed capping if I cap the upstream below 17mb/s the upsream snrm's are always about 9db- The power increases to achieve this -- maybe in an effort to do better. 

 Variations have always been much as yours is now. 

   I have never understood why your upstream power and snrm magnitude behaviour is so different. Even when my line first started and the downstream attainable was 108 the upstream was held done by a very low upstream  power..  It makes me wonder if all ECI DSLAMs are the same or have the same settings

[kitz]

If you think the SNRm graph looks spikey, thats nothing to what I see on the SNRm per band - thats real pretty - not 
I wonder if the SNRm bands is something that could perhaps be something for Tony to think about if he wants to develop things further.

Done I haven't double-checked the displayed data against the Database but looking at yours, since it appears to represent a Magnitude 9 on a Seismometer I guess it's correct 

The entries are at the top of the list under the normal SNRM entry. Hit Update if you don't see them at first on a pane or F5 to get all panes at once.

SNRM
SNRM/Band DS
SNRM/Band US

Thanks - thats brilliant.

That particular graph isnt as bad though as the day before..  I couldn't do a screen grab last night as I wasnt at the PC and I hadn't even checked out properly what the line had been doing as I'd been out and busy trying to re-arrange things so I was free on Wed if a BT engineer needs to come.. and I'd typed out the line behaviour from memory.

Yesterday the line was a lot better no CRCs/Err secs and although U0 is still jagged.. previously all 3 bands were behaving erratically.   Something changed yesterday and the line is now pretty stable.
Whether the added INP has affected that or not Ive no idea, but I dont think that would account for U1 and U2 loosing their jaggedness.

What is strange is that today the separate bands look in much better shape than they have been for the past 5 days..  yet for the first time the combined SNRM graph for this morning is more bumpy than it has been since this started.


--
Edited to add graph

[kitz]

  I am also on an ECI cab and with no speed capping my upstream snrm's are always about 6db at sync, with my varying crosstalk the power adjusts to ensure this at sync.

 With speed capping if I cap the upstream below 17mb/s the upsream snrm's are always about 9db- The power increases to achieve this -- maybe in an effort to do better. 

 Variations have always been much as yours is now. 

   I have never understood why your upstream power and snrm magnitude behaviour is so different. Even when my line first started and the downstream attainable was 108 the upstream was held done by a very low upstream  power..  It makes me wonder if all ECI DSLAMs are the same or have the same settings

I should imagine they do have different settings.  Ive noticed that my PSD mask doesnt always follow the same shape as others. Note how on my D1 bitloading I dont get as much PCB at around tone 255. 

Im very close to the exchange which I can see from my window [well I could, but in summer a neighbours tree in full leaf has grown taller and mostly obscures it atm].  Its already known that BT had 4 possibly 5 different PSD masks depending upon distance from the exchange for adsl2.  So theres bound to be different profiles on the FTTC dslams.  These will probably have to take into account not just distance from the exchange but also distance from the cab.

[tbailey2]

That particular graph isnt as bad though as the day before..  I couldn't do a screen grab last night as I wasnt at the PC and I hadn't even checked out properly what the line had been doing as I'd been out and busy trying to re-arrange things so I was free on Wed if a BT engineer needs to come.. and I'd typed out the line behaviour from memory.

You probably know but if not, you can get any graph for any single day by going to the Range date option (rather than current) and then select both days the same. Attached is the 19th for instance. Along with some -ve spikes....

[les-70]

  In spite of DLM variations I will be very interested in how your line goes from now on and whether there is a fixable fault.  For the last 24 hours all Your SNRM wobbles and spikes are quite similar to those on my line. To be more similar you will need to get odd ES/SES/CRC spikes on top of your fairly low background of errors.  My background is about 5 Es per hour if you remove the 10-20 or so minutes each day with spikes that can contribute most of the days error count.

[kitz]

Thanks tony I was aware of that handy feature.   It was because I was looking at DSLstats anyhow that I grabbed a copy of that.

For the last 24 hours all Your SNRM wobbles and spikes are quite similar to those on my line.

Yeah since the Errors stopped it is very wobbly and not its usual behaviour.  Its currently wavering on a downward trend.   

It is very strange that the wobbles in the total SNRm only started to show when the vast majority of the errors stopped.  For now the individual bands seem much healthier, yet the overall SNRm isnt.   Im not sure how much behaviour is being masked by DLM intervention.

[kitz]

Ive just been advised that an engineer visit isnt deemed necessary.  Instead BT have raised an exchange task "to check a number of issues related to your circuit".  The next update from BT is due after 48 hours.
 

[burakkucat]

Ive just been advised that an engineer visit isnt deemed necessary.  Instead BT have raised an exchange task "to check a number of issues related to your circuit".  The next update from BT is due after 48 hours.

If I was in exactly your situation (knowing, as I do, the precise "lay of the land" of Maison Kitz, the PCP through which your line is connected, its fibre twin cabinet and the telephone exchange) I would have replied to that message to say "I'll be waiting outside the main door of the telephone exchange at 0900 hours, precisely, ready to accompany your technician whilst (s)he investigates the issues with my circuit." 

[kitz]

I may take a stroll to replenish food supplies at some point .  but its not my intention to hover outside a cab for most of the day.   Although Im in sight of the exchange.. the cab isnt.

[NewtronStar]

I may take a stroll to replenish food supplies at some point .  but its not my intention to hover outside a cab for most of the day.   Although Im in sight of the exchange.. the cab isnt.

Just to let you know your back on US fastpath and US sync is rising again