Maybe someone who actually has one will confirm if the noise margin is fixed or changes?

In reply to a post by BatBoy:

Maybe someone who actually has one will confirm if the noise margin is fixed or changes?

Would be helpful, yeah.

just tried, and can confirm that noise margin, and max rate are dynamic.

5. VDSL uptime: 0 days, 09:47:01
6. Data rate: 8263 / 41844
7. Maximum data rate: 8207 / 42553
8. Noise margin: 5.8 / 6.0
9. Line attenuation: 0.0 / 22.7
10. Signal attenuation: 0.0 / 20.4

4. Board version: BT Hub 5A
5. VDSL uptime: 0 days, 09:48:59
6. Data rate: 8263 / 41844
7. Maximum data rate: 8242 / 42724
8. Noise margin: 5.9 / 6.0
9. Line attenuation: 0.0 / 22.7
10. Signal attenuation: 0.0 / 20.4

5. VDSL uptime: 0 days, 10:00:15
6. Data rate: 8263 / 41844
7. Maximum data rate: 8199 / 42808
8. Noise margin: 5.8 / 6.0
9. Line attenuation: 0.0 / 22.7
10. Signal attenuation: 0.0 / 20.4

happy to help John/vwlowen however I can...

edit, another set of stats, showing downstream margin stats are dynamic too
4. Board version: BT Hub 5A
5. VDSL uptime: 0 days, 10:16:21
6. Data rate: 8263 / 41844
7. Maximum data rate: 8207 / 42601
8. Noise margin: 5.8 / 6.1
9. Line attenuation: 0.0 / 22.7
10. Signal attenuation: 0.0 / 20.4

Edited by ggremlin (Sat 23-Nov-13 19:30:41)

Max rate varies because the noise margin is doing. The noise margin varies because the external noise varies - that's the whole point of it .

There are some odd changes in those stats you give though.

Edited by RobertoS (Sat 23-Nov-13 20:12:33)

In reply to a post by RobertoS:

Max rate varies because the noise margin is doing. The noise margin varies because the external noise varies - that's the whole point of it .

There are some odd changes in those stats you give though.

no, the noise margin is derived from max rate and current rates. NOT the other way round

max rate changes due to current conditions. (including external noise/ crosstalk)
current rate is what was established and attainable on connection. (with a margin so its dosn't have to reconnect all the time)
noise margin is the -live- difference.

I would have thought the noise margin varies because the noise varies and the signal remains the same.
I also thought the max rate was some calculation based on the noise margin and varies also because the noise varies..

In reply to a post by ggremlin:

no, the noise margin is derived from max rate and current rates. NOT the other way round

The noise margin (SNRM) at connection time is a setting in the BRAS for the line. It is an imposed value and not dependent on the line conditions at the time.

From that point on it and the SNR can vary continually, unless SRA is in operation which seems for some HH5 users with updated modem firmware may be the case.

In reply to a post by BatBoy:

I would have thought the noise margin varies because the noise varies and the signal remains the same.

Exactly . However:-

I also thought the max rate was some calculation based on the noise margin and varies also because the noise varies..

It's certainly based on some esoteric calculation and varies, but I've yet to work out what and how.

As it is normally close to the actual connection speed it cannot be directly related to the noise margin, though on ADSLx lines with very high noise margin it is always much higher than the actual. It is also frequently below the actual sync even when the SNRM appears normal. It's almost as though it is determined using the BRAS SNRM value during what we see as the discovery stage on FTTC, but by the time the medley stage is completed it is history, retained and shown for diagnostic purposes.

But then on VDSL2 it changes, and not apparently in direct relationship to the noise margin .

At some point I shall have to get the latest unlocked HG612 firmware and stats program. It used to be fascinating to watch the Max figures. I think I have loads of the logs from before the OR firmware update. If so I should be able to provide some info for ggremlin. Bigger things than that on my mind though at the moment.

Edited by RobertoS (Sat 23-Nov-13 23:29:11)

umm sort of, you got a bit confused.

The "target" noise margin is set dslam side. (can be overidden by modems in some cases as seen on adsl).

The actual noise margin does vary and is based on the actual snr and bitrate.

So eg. lets say a tone has 9 bits, each bit requires 3 db of snr. There is 32 snr, that means there is 5db of snrm on that tone, as 27db of snr is required to maintain the bitloading. So as the snr varies the snrm of course also varies. Most modems dont reveal the snr in the gui, so the end user will typically only have the snrm to go on.

As far as I can tell the modem guestimates the max sync rate based on the target snrm, existing sync speed, number of tones in use and spare noise margin.

No, I didn't get confused in the least, unless it was in my explanation to ggremlin, who is definitely wrong. I did only cover the essential error in his post, not attempt a full explanation.

Yes, I'm addressing the "averaged" SNRM over all the tones.

As I said, the Max rate is generally close to the actual. It is net of margin, not the gross before margin is applied to obtain actual.

My web page has explained since September 2009 how SNRM works. Improved in January 2010 and a few additions because of FTTC in 2013. Your single tone explanation doesn't seem particularly well written, seeing as the BRAS SNRM figure is the overall result required. Hence the widely varied loading of individual bins.

Re the Max rate, we are agreed it is a fluctuating figure. Your theory doesn't really add anything to my post, as certainly on VDSL2 its variation is far from closely linked to the overall SNRM fluctuation. IIRC it can even move in the opposite direction. But as I said before, I have more important things to do at the moment than to connect up and use a different machine to get back to the logs I have from Bald_Eagle's prog.