Can anyone make sense of this?

I have two ADSL lines with ZEN and both go down when there number is rung. it doesn't matter if a phone is plugged in and I have tried the usual e.g. different faceplates, modems etc

The following happens

1. When first connects stats look fine
https://www.dropbox.com/s/4967oct9bd8kiw8/line1.png?...

2. After a few seconds without the phone actually ringing the output power under downstream drops to 0 and stays on 0.
https://www.dropbox.com/s/sew99u8r4gkbaze/line1%20-%...

3. As soon as phone is ringing the upstream SNR margin goes crazy going to 6552.2
https://www.dropbox.com/s/v96goaw5mf7dyeq/line1%20-%...

4. After a few seconds of ringing the upstream SNR margin drops right down
https://www.dropbox.com/s/j63nuo5sijnelrs/line1%20-%...

I had the same thing happen last year which is why I ended up with a second line. The only way it went away was to have the line disconnected (terminated) and a new line put in.

If you ring and quickly hang up the modem can sometimes hold on to the connection.

Any help much appreciated.

HR fault, High Resistance fault

Presume you have tried different corded handsets

Both HR ? That would be most unlucky.

Thanks for the reply. The phone can be unplugged and it still does.

When I had issues last year the engineer said the issue still happened at the cab so it was nothing to do with my line between cab and house.

So, why did he not track the issue back from cabinet to exchange?

I have no idea.

Last year I had 10 engineer visits, 4 no shows. I was without a phone line twice for 5 days. I ended up writing to my MP who got in touch.

They tried moving me to a different cab which did fix the issue but it severely degraded the connection which is already pretty slow.

In reply to a post by kingbiscit:

3. As soon as phone is ringing the upstream SNR margin goes crazy going to 6552.2
https://www.dropbox.com/s/v96goaw5mf7dyeq/line1%20-%...

6552.2 I would guess that is actually an SNR of -1.4dB and the field has not been identified as being potentially a Two's complement *number which will allow for negatives.

65522 is fairly close to 65535 or 2¹⁶-1 or 1111111111111111₂

I other words the noise level on upstream is greater than the signal level.



* In Two's complement (16 bit):

+32767 >> 0111111111111111
+10 >> 0000000000001010
+2 >> 0000000000000010
+1 >> 0000000000000001
0 >> 0000000000000000
-1 >> 1111111111111111
-2 >> 1111111111111110
-10 >> 1111111111110110
-14 >> 1111111111110010
-32767 >> 1000000000000001
-32768 >> 1000000000000000

And as you see -14 in Two's Complement would convert to 65522 in normal binary

Wow! Impressive.

Any ideas why this is and how it can be fixed? Is it an issue in the exchange?

No idea. Just that was an obvious comment to explain one of the problems where a modem bug is causing an incorrect display meaning that your UP SNR is not going silly just dropping from normal 6 dB to -1.4dB.