Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[NewtronStar]

I'm not sure who the question was aimed at, but mine are attached.

Thats kind of lowish rtx_tx per min BE1 though i'll have to take your current long line with a low sync into consideration, Im going to call it a day at my end the RPi is just not Co-operating as it feels like a hardware issue on the RPi b+ and it's only six months old 

[pettaw]

I don't harvest my modem for stats and so I didn't notice that my line had G.Imp enabled at all until I was fiddling about downloading something and I noticed my downstream bandwidth looking a bit high, so I logged into the modem. I have always been on fastpath although I have a line which has weird spiky drops in SNR. Some of these spikes would cause sync to be dropped and I would see regular resyncs approximately once every 3 days or so.
When I logged in I mistakenly thought that my SNR had gone up to 16db so I rebooted the modem to see if I could increase sync speed without seeing how long my connection had been up for. Because I didn't realise that the webGUI reports incorrectly I didn't realise until I looked at the stats using BEs scripts....oops. It was then I noticed that interleaving had been applied and I have this new Bearer 1 thing.

Anyway. I do have the luxury of having a Samknows whitebox on my line so I interrogated it and it turns out that I went on G.Imp on the Saturday of the Easter weekend.

I have an interleaving depth of 8 D/S and 8 U/S and my sync speed has gone up from approx 66Mbps to approx 70Mbps but my whitebox data shows it has NOT affected my latency in any way.

I'll keep you updated as to how stable my line is now that G.Imp has been enabled.

[burakkucat]

Just one small point -- it is G.Inp (or G.INP) not G.Imp (or G.IMP)! 

On the subject of "weird spiky drops in SNR", do they tend to occur when the telephone is used? Assuming that there is a telephone connected to the circuit, of course. If yes, then perhaps you would please post a few example graphs that show the effect.

[Black Sheep]

Just one small point -- it is G.Inp (or G.INP) not G.Imp (or G.IMP)! 

On the subject of "weird spiky drops in SNR", do they tend to occur when the telephone is used? Assuming that there is a telephone connected to the circuit, of course. If yes, then perhaps you would please post a few example graphs that show the effect.

Freudian slips ??

[pettaw]

So sorry. Serves me right for not reading my stats carefully. And no the spikes are nothing to do with the telephone. Seem to occur at random.

Freudian slips ??

Very possibly.

I have no recent graphs because as I said I don't run the scripts any more but this is from 2012.

[burakkucat]

Hmm    That is certainly "not right". But accepting your word that there is no correlation with usage of the telephone, I would suggest that it should just be categorised as "one of life's little mysteries".

Ideally I would have liked to have seen the graph for just one 24 hour period, as that would give a better "feeling" for what is occurring.

[NewtronStar]

Thought this may be helpful for G.INP lines regarding the figures you see  INP: 47.00  41.00 DS and US and how this effects your CRC's and ES seconds and note during my observation the Interleave depth has not changed it's still DS 8 and US 4

I have had three changes on the DS and just 1 on the US, initially when G.INP arrived on my line the DS INP: was 46.00 and US INP: 43.00 and then the DS INP: increased to 49.00 this value was able to decrease most CRC's & ES's seconds and after a week or so the DS INP: decreased by 01.00 (48.00) with a wee bit more ES's.

Now the DS INP: is 47.00 and US is 41.00 the ES's have increased 2 fold on the downstream instead of 4 ES per day it's now 8+ .

So that should give us some idea on how noisy the line is for example DS INP value of 41.00 low line noise and DS INP value of 54.00 very high line noise.   

[Bald_Eagle1]

Remember, INP = Impulse Noise Protection.

I have fairly high level of background noise (crosstalk) that robbed me of almost 10 Mbps over quite a few months.

However, my INP values are only 41 for both DS & US, which suggests I'm not affected much by impulse noise.


What sort of daily LEFTRS values are you seeing?

I see around 72 per day, usually in very low delta values of 1, 2 or 3 (see attached graph).

These are apparently some sort of indication of interference.


I recall that your connection's SNRM is somewhat affected in the evenings by Radio China (or other radio station(s).
I wonder if that could be classed as INP? i.e it's not there all the time.


In contrast, see WWWombat's LEFTRS graph over 6 days - only 2 LEFTRS in total.

He has INP values of 46 DS & 47 US

His INP values MAY be higher than mine simply because he has a much higher speed connection or that my connection is at such low speed that it doesn't need higher INP values.


It really isn't as easy to suss out good/poor connections and/or faults etc. on G.INP active connections.

[NewtronStar]

It really isn't as easy to suss out good/poor connections and/or faults etc. on G.INP active connections.

All I am saying here is that a pattern is evolving and that's good

As for the LEFTRS values you will need to ask Roseway to supply this value on his program, but still we are awaiting for these results to be implemented into tony's MDWS, use guys don't seemed to have found a common ground relating to the values that is important for your average G.INP line.

[NewtronStar]

Having a look at the Leftrs value all it seems to do is increment, I take these values are coming in from the dslam in realtime ?

Code: [Select]

Downstream Upstream
General
rtx_tx          293197          57             
rtx_c            15355            1023           
rtx_uc          235              122             
LEFTRS          396              33             
minEFTR          32898            7636           
errFreeBits      124200589        376781790       
Bearer 0
RxQueue          26              12             
TxQueue          13              6               
G.INP Framing    18              18             
G.INP Lookback  13              6               
RRC Bits        24              24             
Interleave depth 8                4               
INP              47.00            41.00           
INPRein          0.00            0.00           
Delay            0                0               
Bearer 1
Interleave depth 1                1               
INP              2.50            4.00           
INPRein          2.50            4.00           
Delay            0                0               


[roseway]

Yes, the LEFTRS values are counters which increment, not instantaneous values line SNRM.

[Bald_Eagle1]

As for the LEFTRS values you will need to ask Roseway to supply this value on his program, but still we are awaiting for these results to be implemented into tony's MDWS, use guys don't seemed to have found a common ground relating to the values that is important for your average G.INP line.

We have been discussing this over the weekend & have now more or less agreed what is worth uploading to MDWS (there's quite a lot of new, relevant data to consider) & Roseway will be including some of it in the graphs locally generated via DSLStats.


Regarding your other message, yes. LEFTRS is indeed reported by the modem as a cumulative value.
We have to calculate the delta values via our programs.

A lot of LEFTRS delta spikes & high total daily values seem to indicate a 'nearside' interference issue/effect (such as general background noise (crosstalk), Rein when interference occurs for longish periods of time & then ceases or sporadic Impulse noise spikes).


FWIW, I do remotely monitor a connection where a water authority pump kicks in at exactly 23:00 each night & switches off at exactly 05:00 the next morning.

I'd like to see how that connection is handled by G.INP, but it is connected to an ECI DSLAM which as we know, are not yet G.INP capable.

[roseway]

I've added a graph of LEFTRS to the G.INP section of DSLstats. It will be in the next release.

[NewtronStar]

I've added a graph of LEFTRS to the G.INP section of DSLstats. It will be in the next release.

That will be good to see this graph on DSLstats it should make easier to understand how this line is performing, I still run hg612_modem_stats but as it not monitoring 24/7 it's very hard to see a trend with all the gaps.

One other thing regarding INP: values they only seem to change when you turn off/on the VDSL2 modem not sure if other notice this to, It could be the time of day when it resyncs due to my variable DS SNRM.