Print

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

[roseway]

The chipset which kitz referred to is the DSL chip in the modem (or modem part of a modem/router). At present, only Broadcom DSL chips have the detailed reporting and configuration capabilities which DSLstats needs to function fully.

[OldTimer]

Sorry to have mixed up my original Output Power query with introducing the discussion of different modem/routers, but let me ask one more question on the subject. Whilst not the full future-proof Monty that I aspired to, I came across several reviews of the Billion 7800 DXL today, does anyone have experience of this particular model?

It does appear to have a Broadcom chipset, it seems to be one of the better ones at dealing with longer lines, and it looks well capable of handling two laptops, a desktop, and the wireless network in a small brick terrace property approx 3.6 km from the exchange.

One other question, would it cope with BT Infinity when it reaches us? The cabinet is there, but it will be about May/June I believe before it is available to us.

Regards,

Alan

[roseway]

I came across several reviews of the Billion 7800 DXL today, does anyone have experience of this particular model?

Yes, I've been using one for the last four weeks or so. It's quite simply the best all-in-one I've ever used. It uses the BCM6361 DSL chip, and on my medium length line it's fast and very stable. And it has more configuration options than any other router I've seen. And it's fully compatible with DSLstats.

It's got four gigabit ports, one of which can be configured as a EWAN port for connecting to a separate modem, so you'd still be able to use it as a router with a VDSL2 modem when you get FTTC. At present it isn't capable of functioning as an all-in-one on VDSL2 connections, although I suspect that a future firmware upgrade may make this possible.

[OldTimer]

Hello again,

I'm sorry it has been awhile, but I'm still wrestling with my low (and capped) upload speed on adsl2+. Recently I pushed the boat out so-to-speak and bought the Billion BiPac 7800DXL, so far I'm mightily pleased with it. There are more statistics than you can shake the proverbial stick at, some of which I confess I don't understand, but the 7800DXL is far more informative than either of my previous BT Hubs.

This is just an experiment, but yesterday I monitored the performance on my line and as this forum allows attachments I append it for your perusal and advice.

With the 7800DXL I can at last see a realistic figure for my upstream output power, rather chuffed by that - but whereas the downstream looks to have increased, which is great, the BT Speedtest upstream reflected a consistent 0.25 Kbps throughout the period I monitored the stats. Given that my 'capped' figure is 0.45 Kbps, being on adsl2+ I really need to see something greater than that to allow me to upload audio/video files within a reasonable time frame.

It may be that I just picked a day to monitor the stats which were not the best for my purposes, perhaps another day may be different. I would be glad if some kind soul could take a peek at the attachment and spend an opinion.

I am in an ongoing dialogue with BT about my upload issue, they simply say that my line becomes unstable at the higher speeds. I find that confusing as previously, for about two years I was consistently getting upload speeds of more than double the 'capped' figure. Now I am not even getting the capped speed, yesterday, as you will see if you check out the attachment I was seeing a speed of something like 0.25 Kbps 

Hopefully someone will pick this up and come back with some pearls of Kitz wisdom.

Regards,

Alan

[roseway]

Hello OT

In order to understand what's going on, we really need to see some graphical monitoring results. DSLstats works well with the 7800DXL (I use one myself), and the SNR margin and bitloading graphs in particular will tell us far more than any number of words.

[OldTimer]

Hi,

Tis I once more. I have downloaded DSLstats as recommended, and am just putting my toe in the water seeing what I need to configure etc... Just to test what you require, I have just taken a 'snapshot' of the SNR and have attached it as an image. Please let me know if that sort of thing is what you mean.

How long would you suggest I leave the laptop on to record the performance, so that I can give you a reasonable picture of what is happening so-to-speak?

Here is the 'copy and paste' stats as found in DSLstats

Stats recorded 21 Mar 2014 12:52:32

DSLAM/MSAN type:           IFTN:0x827a / v0x827a
Modem/router firmware:     AnnexA version - A2pD038f.d24h
DSL mode:                          ADSL2+
Status:                            Showtime
Uptime:                            1 day 16 hours 19 min 11 sec
         
                    Downstream   Upstream
Attenuation (dB):             54.5              34.8
Connection speed (kbps):   4731              297
SNR margin (dB):              6.2              14.1
Power (dBm):                  0.0              12.2
Interleave depth:             32              1
INP:                          1.00              0.50

RSCorr/RS (%):                0.0216      0.0276
RSUnCorr/RS (%):           0.0020      0.0000
ES/hour:                      0              0

Regards,

Alan

[burakkucat]

The US SNRM is abnormally high. It seems clear that Beattie is masking the underlying problem by setting the target SNRM for US to a "silly" value rather than finding the fault and attempting to fix it. 

I'll let Eric explain which graphs will be the most useful to see . . . preferably for a 24 hour period.

To save other Kitizens having a problem, I have taken a copy of your one day's statistic file and "munged" it into a PNG image file -- now attached below.

[roseway]

What we need is to see the SNR margin graph over an extended period. You could set the time per page to (say) 4 hours (in Configuration --> Graphs), and take snapshots of this graph (in Configuration --> Snapshots). This will cover a 24 hour period in 6 snapshots. The other very useful graph is Bitloading, and if you set that to snapshot as well you will have a representative selection covering the 24 hours.

[OldTimer]

Here are a couple of Snapshots of the sections that you mentioned. I'm a little anxious about leaving the laptop on overnight, but I guess for once shouldn't hurt. The next snapshots should be taken just after Midnight, about 12:15am.

Something seems to have affected the SNR in the period around teatime. I had left my laptop running and had gone for a walk at the time, so wasn't using it, apart from letting DSLstats run while I was out, and I've left it running since.

Hopefully as you interpret what the graphs are saying you may be able to suggest what might be going on.

Regards,

OldTimer 

[roseway]

Those deep spikes on the upstream SNRM are presumably the reason for your upstream SNRM being set at such a high value. As time goes on it will be interesting to see how often and when these spikes occur, because the timing and frequency might point to a cause. Does something in your home get switched on at around 5 pm (central heating perhaps)?

The bitloading chart looks fairly normal in the downstream for a fairly long line like yours. As you're not getting any bitloading above tone 200 or so, it might be worth doing what b*cat suggested, and forcing the connection mode to ADSL2 instead of ADSL2+. This can be done in the router web configuration utility:   Configuration --> WAN --> DSL.

[burakkucat]

Purely as a means to provide a point of reference, I attach, below, the bitloading graph for my line. (Scaled, hopefully, to the same size as your bitloading graph.)

Due to its length and quality, the circuit is configured to operate in ADSL2 mode. Note the far healthier loading for the US frequencies.

[OldTimer]

As suggested I unticked adsl2+ in the 7800DXL's DSL config option, and the upstream SNR immediately shot up to over 20 dB, the downstream SNR dropped slightly.

Here are stats taken from the 7800DXL:

                                Downstream   Upstream
Line Coding (Trellis)   On                   On
SNR Margin (dB)           2.6                   20.9
Attenuation (dB)           51.5                   34.5
Output Power (dBm)   0.0                   12.6
Attainable Rate (Kbps)   5620                   928
Rate (Kbps)                   4843                   444

These reflect marginal changes in the downstream figures, but the upstream Attainable has increased to what I thought I might get with adsl2, but never have (so far), and the actual Rate has increased, back up to the so-called capped level.

What I have also noticed is that the 7800DXL is now reporting the 'Mode' as ADSL_G.dmt.bis. This is a bit of a surprise, not even adsl2......., what do you make of that? By the way, DSLstats is still reporting the Mode as adsl2+. Will that change in the 4hr cycle that I have set in config?

Regards,

OldTimer

[OldTimer]

Those deep spikes on the upstream SNRM are presumably the reason for your upstream SNRM being set at such a high value. As time goes on it will be interesting to see how often and when these spikes occur, because the timing and frequency might point to a cause. Does something in your home get switched on at around 5 pm (central heating perhaps)?

Sorry, overlooked answering this question ... as far as I can think there is nothing that gets switched on around 5pm that's not already on. And as far as the central heating is concerned, I would have switched it off before now, but she who must be obeyed feels rather nesh, and the 'gas' fired boiler has been on for ages - just controlled by the thermostats on the rads. 

Regards,

OldTimer (aka Alan)

[burakkucat]

What I have also noticed is that the 7800DXL is now reporting the 'Mode' as ADSL_G.dmt.bis.

That's fine, there's nothing to worry about. The following will explain --

G.Dmt == ITU G.992.1
ADSL2 == ITU G.992.3 == G.Dmt.bis
ADSL2+ == ITU G.992.5 == G.Dmt.bis+ == G.adslplus

I would prefer to review 24 hours worth of harvested statistics from your line. Obviously we will see a "step change" when you reconfigured the modulation from ADSL2+ to ADSL2 but that will not be a problem, as it is expected. If the source of the sudden deep drops in the US SNRM can be identified and removed, then the US target SNRM can be reduced back to sensible value (around 6 - 7 dB).

[OldTimer]

I would prefer to review 24 hours worth of harvested statistics from your line.

Hi, Tis I once more,

As requested I left DSLstats running overnight to capture the performance over a full 24hr period. Observation, even to my untrained eye, over the full 24hrs is quite interesting. As you will expect it will take several post to send all the snapshots.

The spike around teatime yesterday you may recall from ny post yesterday evening, but when I got up this morning, I noticed that my laptop had 'gone to sleep' around 5.30am, as a result the graph flatlined around that point. I ought to have thought about it before, but I amended the instruction in Contol Panel, so that if I leave DSLstats running overnight again my machine should 'go to sleep' as before when left unattended.

You will notice that around Noon today (22nd) the SNR graph again shows a deep spike. I was in front of my laptop at the time so I know that at that point I had two dropped connections in very quick succession. It dropped out, but came back quite quickly, only to drop again within a few seconds.

The last captured graph was supposed to be around 6:31pm tonight, and I have included that. The SNR graphs of course tell the story, and you will interpret them I'm sure, but towards the end of the last period the downstream SNR almost dropped off the scale. At one point it was down to 0.1 dB.

I am also including a couple of extra graphs which I captured manually. I have done so because as I was about the stop DSLstats there was another dropped connection . It came back fairly quickly once again, there was no deep spike on this occasion. However when it did come back I noticed that the SNR had improved slightly.

Anyway, enough of my wittering, the graphs are attached. I await your comments with interest.

Regards,

OldTimer