Why is a "degraded" line not a faulty line?
A BT Openreach person visited on 15/02/13 (after another no show on 09/02/13).
From the initial conversation I had with the Openreach person, and the amount that I've learned during the life of this issue, I'm not sure the extent of technical understanding of the Openreach person was better than mine, except that he had been trained to use his automated test instrument.
For example, when I showed him the SNRM graphs that indicated how my noise margin dropped during the day, his interpretation was that was a good thing, as "an increased noise margin would indicate a problem".
His test instrument advised him that the battery voltage was a little high on one wire, which was a what the last boost engineer had noticed as well. He said this could be because 2 wires were touching at some point. So he left to look at the cabinet and carry out a test there.
His instruction from Plusnet was quite clear about he was only to consider the d-side and not change the e-side at all.
When he returned, he said nothing about the battery voltage, but that he had spoken with Chris at Plusnet and BT Wholesale at the exchange.
He advised that upstream interleaving had been turned on by BT wholesale. He was was also told by the exchange that they could see my connection had been dropping at 1am every day. However my data only showed that the SNRM dropped about 0.5dB at 11am everyday. It may be that the BT wholesale person misread a record and said 1am instead of 11am. However, I wonder what kind of data history is actually available within BT wholesale, and if it would be any use. Accorinf to my router interleaving had been on for the last 5 weeks, however, I noticed that the target upstream SNRM had now been increased from 6dB to 9dB.
He had opened two small junction boxes in the pavement and one larger manhole. The records showed that the larger manhole had not been opened since 2003. He found a joint in there with green corrosion, so re remade the joint and returned.
The Openreach person returned just before 11am, so I was going to have an opportunity to observe any effect on the above 11am noise margin drop that I had been experiencing. At that point my downstream sync speed was 5728kbps and SNRM 6.1dB.
When 11am came, there was no SNRM drop as had been clear over the previous weeks. In fact the SNRM showed a gradual increase.
The Openreach person said this was a Special Fault Investigation not a boost visit and not a boost visit, which explained why he appeared to be distracted form considering the technical aspects of the job.
So, he carried out his closedown checks and left for his next job. Unfortunately the close down checks required my line to be disconnected again and it reconnected at 5337 kbps.
My SNRM continued to gently rise until early afternoon where it sat around 6.4-6.6dB and then gradually decreased towards evening, which seemed better.
So, I planned to do a disconnect/reconnect around mid-day today and see if I got as higher sync speed.
However this morning at 11am , the SNRM drop recurred, so things may still be vulnerable to whatever interference causes that drop.
I went ahead with the reconnect at about 12:25pm and found my speed stayed around the same 5359kbps (ADSL2+). So I resynced, forcing the connection with ADSL2 and it connected at 5287kbps. So I then resynced forcing the connection with ADSL1 and it connected at 5632kbps.
I waited a short time and the errors did not climb as per the last time I tried switching to ADSL1.
These are the current router stats 4h 6m after connection.
From the initial conversation I had with the Openreach person, and the amount that I've learned during the life of this issue, I'm not sure the extent of technical understanding of the Openreach person was better than mine, except that he had been trained to use his automated test instrument.
For example, when I showed him the SNRM graphs that indicated how my noise margin dropped during the day, his interpretation was that was a good thing, as "an increased noise margin would indicate a problem".
His test instrument advised him that the battery voltage was a little high on one wire, which was a what the last boost engineer had noticed as well. He said this could be because 2 wires were touching at some point. So he left to look at the cabinet and carry out a test there.
His instruction from Plusnet was quite clear about he was only to consider the d-side and not change the e-side at all.
When he returned, he said nothing about the battery voltage, but that he had spoken with Chris at Plusnet and BT Wholesale at the exchange.
He advised that upstream interleaving had been turned on by BT wholesale. He was was also told by the exchange that they could see my connection had been dropping at 1am every day. However my data only showed that the SNRM dropped about 0.5dB at 11am everyday. It may be that the BT wholesale person misread a record and said 1am instead of 11am. However, I wonder what kind of data history is actually available within BT wholesale, and if it would be any use. Accorinf to my router interleaving had been on for the last 5 weeks, however, I noticed that the target upstream SNRM had now been increased from 6dB to 9dB.
He had opened two small junction boxes in the pavement and one larger manhole. The records showed that the larger manhole had not been opened since 2003. He found a joint in there with green corrosion, so re remade the joint and returned.
The Openreach person returned just before 11am, so I was going to have an opportunity to observe any effect on the above 11am noise margin drop that I had been experiencing. At that point my downstream sync speed was 5728kbps and SNRM 6.1dB.
When 11am came, there was no SNRM drop as had been clear over the previous weeks. In fact the SNRM showed a gradual increase.
The Openreach person said this was a Special Fault Investigation not a boost visit and not a boost visit, which explained why he appeared to be distracted form considering the technical aspects of the job.
So, he carried out his closedown checks and left for his next job. Unfortunately the close down checks required my line to be disconnected again and it reconnected at 5337 kbps.
My SNRM continued to gently rise until early afternoon where it sat around 6.4-6.6dB and then gradually decreased towards evening, which seemed better.
So, I planned to do a disconnect/reconnect around mid-day today and see if I got as higher sync speed.
However this morning at 11am , the SNRM drop recurred, so things may still be vulnerable to whatever interference causes that drop.
I went ahead with the reconnect at about 12:25pm and found my speed stayed around the same 5359kbps (ADSL2+). So I resynced, forcing the connection with ADSL2 and it connected at 5287kbps. So I then resynced forcing the connection with ADSL1 and it connected at 5632kbps.
I waited a short time and the errors did not climb as per the last time I tried switching to ADSL1.
These are the current router stats 4h 6m after connection.
DOWNSTREAM (Rx)
Noise Margin: 6.2 dB
Connection Rate: 5632 Kbps
Line Attenuation: 48.0 dB
Power: 19.6 dBm
Max Rate: 5920 Kbps
SuperFrames: 868421
SF (CRC) Errors: 122
Reed Solomon: 59052628
RS Corrected: 46318
RS Un-Corrected: 1292
HEC: 96
Errored Seconds: 83
Severe ES: 0
Interleave Depth: 32
Bitswaps: 2143
UPSTREAM (Tx)
Noise Margin: 7.0 dB
Connection Rate: 864 Kbps
Line Attenuation: 27.5 dB
Power: 12.3 dBm
Max Rate: 1056 Kbps
SuperFrames: 868419
SF (CRC) Errors: 16
Reed Solomon: 7380558
RS Corrected: 164
RS Un-Corrected: 0
HEC: 6
Errored Seconds: 0
Severe ES: 0
Interleave Depth: 4
Bitswaps: 245
TOTALS
Total Uptimes (From SF counts):
WAN: 0 days, 04:06:03
-2013Feb18-1216.jpg){kind=link}
-2013Feb18-1217.jpg){kind=link}
This is definitely a case of see how it goes and maybe a bit of trial and error to reduce the errors - I'll try and omit any further puns 
First observation is that the US errors are very much better, but I was somewhat surprised at the US upload speed, ought to be nearer 0.7Mbps.
Try running the BTw speedtest at varying times. You can run the Diagnostics every 10 minutes upto 3 times an hour, but the first stage can be run as often as you want.
The DS errors aren't too bad, but difficult to say how much of an improvement there may be or whether ADSL2 or 2+ may be better. I'd stick with this a while and see how it goes.
The Bits/Tone allocations are somewhat better at the Low Frequency end both in the US and DS and that chunk 212-238ish as well.
In case I haven't mentioned it before, missing tones 133, 147 & 152 are quite common on 21CN for ADSL1 or ADSL2. Different tones are missing for ADSL2+ .
I suspect there might be a bit of a trade-off between the CRC count (and the others) actual throughput speeds, sync speed and mode as well as day/night SNRM swing.
The 1100 ish SNRM drop definitely seems like something being switched on. It's difficult to see when it's being turned off, it looks as though it could be around 2230 on your most recent graph. See if you can spot anything similar on the SNRM plots for other days.
First observation is that the US errors are very much better, but I was somewhat surprised at the US upload speed, ought to be nearer 0.7Mbps.
Try running the BTw speedtest at varying times. You can run the Diagnostics every 10 minutes upto 3 times an hour, but the first stage can be run as often as you want.
The DS errors aren't too bad, but difficult to say how much of an improvement there may be or whether ADSL2 or 2+ may be better. I'd stick with this a while and see how it goes.
The Bits/Tone allocations are somewhat better at the Low Frequency end both in the US and DS and that chunk 212-238ish as well.
In case I haven't mentioned it before, missing tones 133, 147 & 152 are quite common on 21CN for ADSL1 or ADSL2. Different tones are missing for ADSL2+ .
I suspect there might be a bit of a trade-off between the CRC count (and the others) actual throughput speeds, sync speed and mode as well as day/night SNRM swing.
The 1100 ish SNRM drop definitely seems like something being switched on. It's difficult to see when it's being turned off, it looks as though it could be around 2230 on your most recent graph. See if you can spot anything similar on the SNRM plots for other days.