In reply to a post by RobertoS:

Have it your own way.

I wouldn't say it's my way. It's the correct way. I'm not entirely clueless, you know

[attainable rate] is a calculation based on what the modem thinks if could do if no limits were imposed by the exchange based hardware [source]

The modem actually measures things like line and signal attenuation and actual power and uses these to calculate the maximum physical speed the line is capable of regardless of IP profile or banding. As such, banding is a red herring in this thread.

Things which will affect attainable rate include line and signal attenuation, the VDSL profile the circuit is on (e.g. 8a, 17a or 30a) and the target SNR margin.

I suspect that routers typically calculate attainable based on a target SNR margin of 6dB.

My router at home is currently synced at 80000/19999 with an SNR Margin of 4.5dB. Attainable is less than this at 73080/26049.

Thanks for that, I'll have to check. The HH5 syncs at a lower rate that the Draytek 2850 which definitely doesn't support G.INP. The Draytek provides much better stats, which is why I took it with me.

How sure are you given the mess of wiring, that there is nothing before the master socket test socket?

Attenuation figures (though may have missed them) and what is distance to the VDSL2 cabinet?

The Openreach estimates are just estimates and are not always correct. The bin plot I suspect will show that the signal is fading away in the middle of one of the downstream bands, but upstream is holding up or there is a local noise source causing issues in a download band.

If the physical copper environment is suspect, then a modem that can provide the Hlog, QLN, SNR and bit-loading graphs all help. They are a sanity check, at least.

The attenuation (single, and per-band) then offers comparable material on MyDslWebStats.

I wondered about something before the master socket too. The line comes into a BT 201D and that seems as though it was installed by someone from BT. The circuit is labelled with details of the PCP. I wouldn't expect that if what Zarjaz calls a 'have a go hero' had wired it up. There shouldn't be anything before this.

From there to the master socket is a few centimetres of cable. I've checked there aren't any loose connections. Well, there was just one; on the connection to the burglar alarm, which had been giving intermittent fault warnings anyway. That's now fixed.

Maybe I'll put a tone on one of the spare pairs on the incoming and see if I can follow where it goes. It's not obvious looking at it. It disappears through a hole in the ceiling and on the floor above there's no sign.

I'm a bit limited as far as diagnostics are concerned as the HH5 is beyond useless and my unlocked HG612 is currently being used elsewhere.

I haven't definitely located the cabinet yet but I think I know where it is. I'll do that on my next trip. For upstream to show an attainable of 22808, it must be reasonably close.

As for REIN, I've looked for possible sources of interference. This is a domestic situation and the only thing that I can see that's close to the incoming cable is the central heating controller. I'll take my AM radio next time.

Have you determined if you're on Infinity 1 but paying for Infinity 2 yet?

You can get decent stats from HH5 including sync speed, line snr etc, you just need to go into advanced & technical log.

EDIT: Checked in a hh5 go to troubleshooting and helpdesk for stats.

Should see similar to:
1. Product name: BT Home Hub
2. Serial number: +068343+NQ33689394
3. Firmware version: Software version 4.7.5.1.83.8.173.1.4 (Type A) Last updated Unknown
4. Board version: BT Hub 5A
5. VDSL uptime: 4 days, 05:15:14
6. Data rate: 19999 / 79995
7. Maximum data rate: 30324 / 98992
8. Noise margin: 15.2 / 8.8
9. Line attenuation: 0.0 / 14.7
10. Signal attenuation: 0.0 / 14.7

Edited by ukhardy07 (Wed 17-Aug-16 12:23:08)

No, because that won't change the downstream attainable, and that's the problem I'm dealing with initially.

So you're doing it the wrong way round - solve a problem that may not be a problem before determining what package you're actually on

Thanks for that, but those are still very limited stats compared to the HG612 or the Draytek. They don't give any idea of error rates or whether particular frequency bands are more affected than others.