Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2

Author Topic: FTTC issues - stats query  (Read 7491 times)

jacknorris

  • Just arrived
  • *
  • Posts: 7
FTTC issues - stats query
« on: August 25, 2012, 12:23:15 AM »

Had FTTC installed 3 weeks ago, and have been having issues ever since.

The first time it was fit the PPP session would drop if any load was put on the line, speed tests would kill it before they could complete. I managed to get BT to come out and investigate, had a lift and shift completed which did not solve the issue, in the end it turned out to be a faulty ECI modem which was replaced and the connection was stable again. It was short lived as the PPP session dropped again 3 hours after the engineer left, it wouldn't bother me so much if it would automatically reconnect but at the BT end it was still showing me as connected and I had to manually reboot the VDSL modem every time. I managed to get BT out again who then said there was a fault at the exchange and that it had been fixed (though I had previously been told this twice and nothing had actually been done). The issue appeared to be better but then the modem started to physically re-sync every 3 - 12 hours.

I acquired a HG612 and unlocked the modem to get some line stats to see what was going on, surprisingly the HG612 hasn't dropped sync at all in 3 days however I would appreciate it if someone could check over the logs and let me know if things looks odd and if I should get BT involved again. My PPP session is still dropping but only twice in three days and it automatically reconnects. Google maps shows me at approx 110m from the cabinet.

[Current Stats] - https://www.dropbox.com/sh/251ne6443t8si5a/g8sYXWBfdt
[Last 24hrs] - https://www.dropbox.com/sh/97owpvfjis1noog/yqCScS9cMi
[Last 3 days] - https://www.dropbox.com/sh/smnlxnf7zht813w/Y-xwsBdvrc
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: FTTC issues - stats query
« Reply #1 on: August 25, 2012, 04:39:35 AM »

Welcome to the Kitz forum, Jack.

I've had a quick look at your graphs and from the snap-shot set, it looks as if you have quite a good line. As for a more detailed comment, I think it best to let our resident Bald_Eagle make his official proclamation.  :)
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #2 on: August 25, 2012, 07:37:59 AM »

Hi Jack * Welcome to the Kitz forum.


AS mentioned by b*cat, an initial look at the snapshot graphs indicates that you appear to have "quite" a good connection.
However, for the very short distance of 110m (if reasonably accurate), I would have expected to see maximum 79999K sync speed.

Your attainable rate of over 100000K suggests that it could easily sync at the currently capped 80Mb, if "things" weren't restricting it.

Now, looking at your Plink log, a few things stand out as probable reasons for the less than Max sync speed:-

SNR (dB):    9.1       12.8 - Shows some "spare" capacity, but not as much DS spare as I would expect for only 110m.

D:      1549      1 - Interleaving depth - (1 = OFF), so DS of 1549 looks very high, again for such a short distance.

ES:      6827      39 - For around 72 hours connection (AS: 259298), that seems a high count. My connection is upward of 850m & I only see around 100 DS Error Seconds in any 24 hour period.

INP:      4.00      0.00 - Impulse Noise Protection seems quite high, suggesting that DLM has seen quite an amout of Impulse Noise on your connection. MY connection's value is only 3.00 DS.

delay:      8.00      0.00 - This will increase PING times as part of DLM's Interleaving & Error reduction attempts.

MOst other short connections (up to say 200m or so) have zero values for INP & delay, with a values of 1 & 1 (OFF) for Interleaving.


Regarding your ongoing graphs, unfortunately more or less everything looks quite poor.

Apart from an occasion spell of calmness, your SNRM fluctuates quite badly, especially over such a short distance, suggesting some sort of electrical or radio frequency interference, although your QLN graph looks quite good.
Could that be something in your home, I wonder?

Various error counts are cumulatively high (as reported in your Plink log) & minute by minute are very high, again when considering the short line length.
Many users who can sync at the full capped 80Mb demonstrate zero (or certainly less than 10 errors per minute) for RSUnCorr, FEC, HEC & CRC errors.

As far as reporting this to your ISP/BT, I suspect you will be told that "your connection is performing within acceptable limits", especially as sync speeds are high.

I don't know exactly what criteria is used for error count per minute/hour thresholds though.

Additionally we can see from your pbParams band plans that you are connected to an ECI cabinet DSLAM:-

Discovery Phase (Initial) Band Plan
US: (0,95) (880,1195) (1984,2771)
DS: (32,859) (1216,1959) (2792,4083)

Some OR staff may argue that you are not using the compatible provided ECI modem & that in itself is causing the issues.
That would be utter nonsense as we have monitored other connections to ECI cabinets via HG612 modems that do not display the issues you are seeing (unless your unlocked HG612 is slightly "faulty").

Finally, OR may try to discredit your claims of instability/problems as the data you possess has been obtained by equipment that has been unofficially "hacked" & therefore cannot be classed as reliable data.
Again, that would be nonsense as all that has been done is to open the modem in order to access the stats that it has been generating all along.

FWIW, I have attached a couple of graphs from your connection to demonstrate my comments above.

Cheers,

Paul.
Logged

jacknorris

  • Just arrived
  • *
  • Posts: 7
Re: FTTC issues - stats query
« Reply #3 on: August 25, 2012, 01:35:00 PM »

Thanks for looking at this :)

I have tried to think of anything in the house that could be causing this but I am coming up at a loss currently.

I'm not really bothered about the download being stuck at 70, the main reason I switched is for the increased upload - currently it appears as though I am limited to 12mbps. The BT speed test doesn't mention any limits to the upload but then again, I have never seen that change. I managed to get a decent 18mbps speed test before i had issues with the ECI modem.

Could DLM be causing this and do you think it will it relax on this line over time ?
« Last Edit: August 25, 2012, 01:44:27 PM by jacknorris »
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #4 on: August 25, 2012, 02:31:50 PM »

Thanks for looking at this :)

I have tried to think of anything in the house that could be causing this but I am coming up at a loss currently.

I'm not really bothered about the download being stuck at 70, the main reason I switched is for the increased upload -


I wouldn't be too bothered if my connection was stuck at 70Mb either.


Quote
currently it appears as though I am limited to 12mbps. The BT speed test doesn't mention any limits to the upload but then again, I have never seen that change. I managed to get a decent 18mbps speed test before i had issues with the ECI modem.

Well, your connection is in sync at 20Mb US (Path:   0, Upstream rate = 20000 Kbps, Downstream rate = 73322 Kbps)
Are you connected to the router wirelessly or via ethernet cable?

If ethernet, I'm not sure what to suggest other than external contention or routing issues as your US connection stats (apart from SNRM) don't look too bad at all.

Quote

Could DLM be causing this and do you think it will it relax on this line over time ?


I would say that is quite doubtful as US sync speed is capped at 20Mb from 31.5Mb attainable, with spare US SNRM overall, US Interleaving, INP & delay are all OFF & US error counts are quite low.

My US cap is 10Mb, but I can only sync at around 4.5Mb US.

You would no doubt see a lower US sync speed if DLM (or distance from the cabinet) had anything to do do with your lower US throughput.

Do you still see lower US throughput at really quiet times, say in early mornings, between 01:00 & 07:00?

Is this a BT Infinity service or another ISP's offering?

Logged

jacknorris

  • Just arrived
  • *
  • Posts: 7
Re: FTTC issues - stats query
« Reply #5 on: August 25, 2012, 03:32:02 PM »

I have BT Infinity and I am connected via Ethernet, have tried tests right around the clock with similar 12mbit results.

Noticed that since my first post the modem re-synced twice during the night, my sync speed has dropped some more and it looks like interleaving has increased.

[Current Stats] https://www.dropbox.com/sh/zwdc21n9s7n8iyf/aq_K98Oj5J
[Ongoing Stats 24hrs] https://www.dropbox.com/sh/k8vs60opz02oslk/vhFPHn3iGZ
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #6 on: August 25, 2012, 04:08:05 PM »

I had similar fluctuating SNRM levels (DS only though) & random diconnection issues for a few months.

The SNRM issue was directly related to defective filtered faceplates as installed by OR (I have now had 3 of them fitted) & the random disconnections issue was eventually traced to the top of a pole at a poor joint with the underground cable & the DP.

Do you have a dangly filter that you could try in the test socket with all other internal telephone wiring disconected instead of using the filtered faceplate?
That would either rule in or rule out the faceplate as being part of the problem.

If you watch the modem's GUI stats closely , can you see SNRM changes or error increases directly related to telephoning out or receiving calls, or maybe a Sky box being turned on etc?

If so, the phone call thing may be an indication of an external D-side cable high impedance/resistance issue that may need investigation.

Your QLN graph looks fine, but can you detect any sort of noise when conducting a quiet line (17070) test over the telephone?

FWIW, I have attached an updated version of GRAPH6.BAT that you may wish to use.
It is used for plotting the Current_Stats data.

It produces smaller filesized montages that can be directly uploaded here if desired & also includes US data shown in green that is sometimes, but not always provided by ECI DSLAMS. It is NEVER provided by Huawei DSLAMS for FTTC connections.


Logged

jacknorris

  • Just arrived
  • *
  • Posts: 7
Re: FTTC issues - stats query
« Reply #7 on: August 25, 2012, 04:31:30 PM »

It is interesting that you should mention the filter, during the original fault I was testing with the engineers test socket and a normal dangly filter and managed to get 24 hours uptime. I then tested it again with the OR filtered face plate and it dropped within 5 minutes however upon switching back to the test socket it dropped again so I wasn't too sure on what was going on.

I have added the updated script and I will give the quiet line test a go shortly, there is only one phone on the line but there is a UPS very close to the modem which could be causing interference I suppose (it's certainly noisy enough to hear!).

I do find it strange how since 11am today the SNR graph is extremely stable and there are next to no errors being shown, as far as I am aware nothing has changed from any other day. I might wait until the line stars showing SNR issues again before switching the filter over.

Thanks for your help! I will keep you updated :)
Logged

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #8 on: August 25, 2012, 04:59:28 PM »

Mentioning Engineers & filtered faceplates has reminded me of one of the many Engineer visits (or Technician visits as one of our resident Engineers prefers)  ;) :P :lol: :police:

The Technician was dismantling things in order to connect his various line testing gadgets when the last 5mm of the incoming drop wire just fell off, leaving one wire of the pair completely disconnected.

IT appeared that it had been punched down in the connector that firmly by a previous technician that it had severely reduced the cable's thickness, thus weakening it along with any signal trying to pass along it.

Stripping this wire back to good copper & reconnecting certainly improved the stability of my connection by a small amount.

I would mention that it took over 11 months to finally get my connection's stability reinstated, despite having literally hundreds of connection graphs, details of wildly fluctuating BT IP profiles, evidence of re-syncs that aren't detected by ISPS as they are too quick to initiate new PPP sessions etc.

I have a sneaking suspicion that unless your connection completely fails, you could also be in for a long haul  :no:

To be fair to the Engineers/Technicians though, it can't be easy to track down intermittent & irregular faults, especially when they are under pressure to keep visit duration times to an absolute minimum.

Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: FTTC issues - stats query
« Reply #9 on: August 25, 2012, 08:26:06 PM »

Being somewhat more awake, I note that the Hlog graph does show some form of line defect.

Notice the pronounced "roll" and "hump" when transitioning to the higher frequencies?

(No doubt that Asbokid will upload an edited version of the image, with a dashed blue line showing the expected path of the curve!  ;)  )
« Last Edit: August 25, 2012, 08:29:23 PM by burakkucat »
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #10 on: August 25, 2012, 10:22:28 PM »

Being somewhat more awake, I note that the Hlog graph does show some form of line defect.

Notice the pronounced "roll" and "hump" when transitioning to the higher frequencies?



In the meantime, from other connection stats I have seen, the general shape of the Hlog graph does look fairly typical for a HG612 modem connected to an ECI DSLAM (example snapshot graphs attached).

Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: FTTC issues - stats query
« Reply #11 on: August 26, 2012, 03:09:45 AM »

If the Kinky Hlog graph is typical for a Huawei CPE with an ECI DSLAM configuration, then I guess my comment is irrelevant and it is time for b*cat to go to his sleepy place . . .  :sleep:
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

jacknorris

  • Just arrived
  • *
  • Posts: 7
Re: FTTC issues - stats query
« Reply #12 on: August 26, 2012, 06:26:59 AM »

Well I had what looks to be a stable line for most of the day and suddenly its fluctuating again - I have absolutely no idea on that one...
Will try the line test later today and also the filter change.

Can you use the AM radio test for FTTC connections and if so does anyone know what frequencies to look at?

[Ongoing Stats 24 hrs] https://www.dropbox.com/sh/j8na1sc6x0o3bm1/R0ozIb9o6a
Logged

burakkucat

  • Respected
  • Senior Kitizen
  • *
  • Posts: 38300
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: FTTC issues - stats query
« Reply #13 on: August 26, 2012, 06:00:33 PM »

Can you use the AM radio test for FTTC connections and if so does anyone know what frequencies to look at?

Yes, you certainly can.

If you have a receiver capable of such low frequencies, make a start somewhere in the 300 - 306 kHz range. If that is too low a frequency, just try its harmonics, i.e. 600 - 612 kHz, etc.
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

Bald_Eagle1

  • Helpful
  • Kitizen
  • *
  • Posts: 2721
Re: FTTC issues - stats query
« Reply #14 on: August 26, 2012, 10:03:03 PM »

Well I had what looks to be a stable line for most of the day and suddenly its fluctuating again - I have absolutely no idea on that one...

That is indeed a most strange phenomenon.

The timings look all wrong for it to be a central heating pump, air-conditioning unit, street lighting etc.

A fridge/freezer starting to intermittently go on the blink maybe?


FWIW, if you wanted to, you could edit graphpd.BAT to display the full extent of your DS attainable rates that are currently chopped off.

Find :attainable_rate_gnp, then change whatever values are currently there to these:-

echo set yrange [0:130000] >> attainable_rate.gnp
echo set ytics axis 5000 nomirror >> attainable_rate.gnp

Also :-

echo set key at "%DTBEGIN%",129000 left noautotitle box 3 >> attainable_rate.gnp

&

echo set key at "%DTBEGIN%",124000 left noautotitle box 3 >> attainable_rate.gnp


Logged
Pages: [1] 2
 

anything