Upload capped again?/ Something screwy

[#1]

Moderator's Note:
I have merged together two threads that I believe cover the history of the original poster's upload speed problem.
David (spraxyt)

[#2]

Hi James
And why can you not work a 7 day week, some people eh   (just kiddin, I appreciate your help).  Thinking about it, it was around the middle of last week when these issues started I think.  However...... since I have moved to connect directly into the XTE faceplate, I have not once dropped sync, and uptime is just under 1 day, but I have had a loss of the PPP connection once.  I will monitor this, it may be my Cat5E cable that was slightly faulty, and maybe the higher frequencies of ADSL2+ are showing it up, also it may be the london issues as well.  The Cat5E cable is not proper CAT5E, it was bought from Clarity as an Engineers install kit. http://www.clarity.it/xcart/product.php?productid=16139&cat=250&page=1. It is rather stiff, and it could be breaking down I suppose. Was stable on ADSL though. Will wait and see......

[#3]

Does this mean anything to anyone.... I have used DMT. I know SNR etc is now locked out, but I wondered if it would mean anything to someone more knowlegable than myself..... (and thats not hard!!)

[#4]

Hiya - Now is probably not the best time to take readings.  Customers on ADSL2+ are seeing massive packet loss (esp on upstream), which is why I suggested waiting until BT had fixed their kit.

[#5]

Cheers James.  As an aside, had no resyncs since the move to connecting directly to the master faceplate (not the test socket).  This appears to confirm that the Cable to the RJ11 box is faulty. Will wait until the cable is fixed in London, before I go any further
Cheers
Chris

[#6]

Things seem better now. Had a resync event yesterday afternoon, and Upstream has gone back up significantly. Also line appears alot more stable. Not sure if this is due to the router now being connected to the master socket, or the line card fault repaired.  ANyhow, will continue to monitor, and close the ticket.  BTW, whats the target SNRM Downstream on ADSL2+.  I resynced at 3db 
Also the FEC/CRC error reporting appears to have gone back to normal.
Attached are the stats, do these now seem OK.

[#7]

The re-sync might have been due to the DLM changing your downstream target noise margin to 3dB. I think the options it has on 21CN range from 3 to 18dB in 3dB steps. It will go back to 6dB in 24 hours if it feels the number of errors is then "excessive". That could manifest itself as low throughput, because the number of retransmissions is large.
David

[#8]

The msan's seem to be very quick in upping and lowering snr target's.
my line on adsl ran at 12db with a sync of 3500-3600 and stayed up for weeks with only 1 db variation over night.
if i ran sync at 4000+ overnight the variation in snr was 3db+ and would most likely drop connection but would always re sync with a 12db target i did the same on adsl2 and it lasted the night with the same 3db+ drop in snr but dropped the following day at about 4pm and re synced with a 15db target.
Its now running the same as this

Quote

my line on adsl ran at 12db with a sync of 3500-3600 and stayed up for weeks with only 1 db variation over night.

but on a 15db target

[#9]

Well things seem pretty stable now. Still connected to Faceplate, and I think I will keep it that way, and have a rethink of the network.  Little worrying that SNR goes down to around 2.5 during the night, but on the upside, connection does not drop.  Just not used to seeing it so low, but I think its how ADSL2+ is meant to be. 
Link Information

Uptime: 2 days, 3:02:50
DSL Type: G.992.5 annex A
Bandwidth (Up/Down) [kbps/kbps]: 1,060 / 7,267
Data Transferred (Sent/Received) [kB/kB]: 0.00 / 0.00
Output Power (Up/Down) [dBm]: 12.5 / 0.0
Line Attenuation (Up/Down) [dB]: 24.0 / 44.0
SN Margin (Up/Down) [dB]: 7.5 / 2.0
Vendor ID (Local/Remote): TMMB / IFTN

[#10]

Yes, it seems a reasonable assumption that ADSL2+ management wouldn't use a 3dB target noise margin if it was not expected to be viable. What are the FEC, CRC and HEC error counts now? The downstream FEC count was 18 million last time, so probably around 50 million now. High FEC errors seem to be a "feature" of these Thomson routers.
David

[#11]

David
Full Stats below.
Link Information

Uptime: 2 days, 21:49:53
DSL Type: G.992.5 annex A
Bandwidth (Up/Down) [kbps/kbps]: 1,060 / 7,267
Data Transferred (Sent/Received) [kB/kB]: 0.00 / 0.00
Output Power (Up/Down) [dBm]: 12.5 / 0.0
Line Attenuation (Up/Down) [dB]: 24.0 / 44.0
SN Margin (Up/Down) [dB]: 7.0 / 3.0
Vendor ID (Local/Remote): TMMB / IFTN
Loss of Framing (Local/Remote): 11 / 0
Loss of Signal (Local/Remote): 1 / 0
Loss of Power (Local/Remote): 0 / 0
Loss of Link (Remote): 0
Error Seconds (Local/Remote): 5 / 0
FEC Errors (Up/Down): 0 / 85,731,008
CRC Errors (Up/Down): 0 / 999
HEC Errors (Up/Down): 0 / 891
Those odd upstream errors seem to have gone.  Not yet resynced, so fingers crossed.....

[#12]

Thanks, your connection seems to be doing well now.
David