Best 15 quid (2)

I'll start a new post to give an update since the last post deteriorated in a unfounded accusations (I am bearded_malc and not anyone else).

I eventually got interleaving turned off. I'd been told it was off but sync stayed at 7616 with contradicted that. Finally I was believed and it was actually turned off. Now I sync at 8128 and IP profile has gone to 7150. Download/Upload/Ping is 6.75/0.38 Mbs, 17ms. On a good day of course. On old DSL-504 best was 5.26/0.38 Mbs, sync = 6624, IP profile 5500.

BTW TP-Link helpfully gave me the telnet commands doc. Not as nice a DMT gui but the numbers are all there. This was helpful in confirming I was still interleaved. I now mostly use telnet to check my error rates.

In reply to a post by bearded_malc:

I'll start a new post to give an update since the last post deteriorated in a unfounded accusations (I am bearded_malc and not anyone else).

I eventually got interleaving turned off. I'd been told it was off but sync stayed at 7616 with contradicted that. Finally I was believed and it was actually turned off. Now I sync at 8128 and IP profile has gone to 7150. Download/Upload/Ping is 6.75/0.38 Mbs, 17ms. On a good day of course. On old DSL-504 best was 5.26/0.38 Mbs, sync = 6624, IP profile 5500.

BTW TP-Link helpfully gave me the telnet commands doc. Not as nice a DMT gui but the numbers are all there. This was helpful in confirming I was still interleaved. I now mostly use telnet to check my error rates.

What's your attenuation and SNR margin now?

Good question. Attenuation is the same at 24dB. It was that on 7616 which is not much less than 8128 so not unexpected that it reads the same I guess. Readings seem to be to nearest 0.5dB judging by the SNR margin. SNR margin is lower. Peak on 7616 was usually about 8 and it is now about 6 (both with small blips up and larger blips down). AFAIR I didn't have any errors shown on the telnet error report when on 7616. I get some now. Not many, 15min period error rate is 0 or 1 so 1:900 lost at most. And that was playing internet radio to give a continuous work out.

Some time I could try to improve my cabling, get a better micro-filter, etc.. What I have isn't bad, filter is SIN2 standard but again it's old, not a NTE5 for example. Home phones are all after the filter. I've dropped the ring wire and used ethernet cable to the modem phone socket but modem phone cable runs with ethernet and other cables and I'm not sure how good the modem cable is. It's not screened, just flat. But this would be more important if the exchange ever gets ADSL2. At the moment all it might do is make a very small error rate smaller still.

In reply to a post by bearded_malc:

Good question. Attenuation is the same at 24dB. It was that on 7616 which is not much less than 8128 so not unexpected that it reads the same I guess. Readings seem to be to nearest 0.5dB judging by the SNR margin. SNR margin is lower. Peak on 7616 was usually about 8 and it is now about 6 (both with small blips up and larger blips down). AFAIR I didn't have any errors shown on the telnet error report when on 7616. I get some now.

Errors would have been corrected by interleaving - that's what is does. I'm not surprised the attenuation remains the same but the SNR seems a bit low though, so you're right it would affect ADSL2+ badly.

FYI. Interleaving is now off. Connects at 8128. Atten = 24dB SNRM ranges 4-6.5 dB, mostly 4.5-6dB. Max test download speed is 7.11Mbps.

7.11 Mbps sounds like a dodgy tester. 8128 sync would normally give a profile of 7150 kbps. TCP/IP overheads would reduce throughput to a theoretical 6902 kbps. The best speed test I've done on a 8128 sync is 6804 kbps.

Could it be OP is on the latest BT profile system which gives an IP profile of 88% of sync speed. Figures fit. My last 4 IP profiles have been 88 % of my sync speed

I'm assuming his profile (7150 kbps) is 88% of his sync (8128 * 0.88 = 7152.64).

The new 88.2% profile is only available on 21CN Fujitsu MSANS in the exchnages, during Phase 2 and Phase 3 more will be added. << Update from BT mods.