Particularly relevant to the OP might be that on a long line fastpath can actually be slower than interleaved since there may have to be excessive re-transmission of packets due to a greater likelihood of the packets being corrupted by noise?

Thanks, everyone, for confirming my ideas. I shall now have to get back to Talktalk and see how far I get!

Roger

The whole point of interleaving is to improve throughput speed (including speed test speed) at any given sync speed

In reply to a post by RobertoS:

The whole point of interleaving is to improve throughput speed (including speed test speed) at any given sync speed

Can interleaving depth also have an effect on sync speed?

On my partial LLU connection I can sync anywhere between ~12500Kbps and ~13500Kbps on a 6dB SNRM (37dB attenuation.) The TG585 v7 router does not indicate interleaving nor it's depth but I certainly get fewer FEC down errors at ~12500Kbps than at ~13500Kbps. I tend not to take any note of output power up/down but I think it stays at 12dBm and 19dBm. Obviously when there are differences in sync speed (down bandwidth) there are corresponding differences in throughput speeds...

No. The interleaving depth is probably the same at both sync's, the sync varying depending on line conditions at the time you connect.

Lower errors are an expected result of lower sync speed. That's why bumping up the sync-time noise margin is the standard fix on BT Wholesale DLMs. A higher sync-time SNRM lowers the sync speed, so lowers the error rates, which is what the DLM wishes to achieve.

Sync time SNRM's are always very close to 6dB though (according to the router log) would a difference of 0.5dB on the reported SNRM equate to a variation of ~1000Kbps in sync speed on a 37dB attenuation?

Edited by 4M2 (Sun 15-Sep-13 15:30:44)

Sync-time SNRMs are by definition always close to the current parameter setting in the DLM. What is different, and causes sync speeds to differ between sync's, is the SNR.

See my website description of the process .

Edited by RobertoS (Sun 15-Sep-13 15:58:43)

Thanks Bob, but I don't know what my SNR is

Perhaps I got confused when with ADSL Max there can be a reduction in sync speed:

Your website "Connection speed may or may not be affected, but on ADSL Max if the line can connect at 8128kbps on Fast Path this is usually reduced to 7616kbps. A few combinations of modem/router and exchange DSLAM can still achieve the full 8128kbps even with interleaving, particularly TI chipset-based."

And Kitz: "It should also be pointed out that whilst BTw state that applying interleaving shouldn't reduce your line speed, it does reduce the maximum line rate achievable from 8128kbps to 7616kbps due to the additional overhead required for check bytes.
Note: although BT state 7616 is maximum sync speed with interleaving, many instances of higher sync speeds have been reported by users. This is dependent upon your router being able to support S=1/2 mode which effectively combines two RS code words into a larger logical code word of 510 bytes (ANSI T1.413).

Interleaving and Error Correction are always switched on at the same time and the 7616 profile has a 512kbps error correction overhead with an increased latency of around 16 ms.
Whilst Interleaving on its own may not reduce your maximum sync speed, Forward Error Correction can and does affect your maximum sync rate."

Edited by 4M2 (Sun 15-Sep-13 16:23:03)

Very rare to be able to find out the SNR, but it is that which changes all the time. SNRM just tells you how much it is changing, and is reset at every sync.

In reply to a post by RobertoS:

Very rare to be able to find out the SNR, but it is that which changes all the time. SNRM just tells you how much it is changing, and is reset at every sync.

So in other words: a variation in SNRM whilst a router is sync'd to the exchange can be seen as an indication in the variation of SNR that is happening all the time since there is a default SNRM. Thus sync speed would be calculated according to the SNR at sync time together with a further reduction in sync speed corresponding to the default noise margin.

Returning to the Forward Error Correction/interleaving topic: are FEC errors reported by a router an indication of interleaving depth?