Print

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

[kitz]

CAnt see any real pattern yet other than perhaps it appears to be worse when people are home?

>>  It does appear that your line suffers from a significant amount of RFI during the hours of darkness.

yep..  problem could be tracking it down :/ 
Bows to the black cats advice-  do you reckon it may be worthwhile trying the radio trick?
Im not sure if its just too fuzzy to pick anything up.

I think if it was me, I'd be going round the house turning everything off whilst leaving DSLstats running to see if theres any improvement.

Culprits in the past have been TV's, Monitors, etc.  Weve even had a music centre with a faulty plug that was wiping out adsl for 6 local houses.

[kitz]

PS...  before I forget..   although its not relevant atm, you may be interested in the following thread where users on a HG612 are testing capping dsl speeds.
iirc your router has a similar chipset so you may be interesting in reading the results.   I only mention it because it was something you had brought up earlier.

http://forum.kitz.co.uk/index.php?topic=13130.msg248345#new

[burakkucat]

>>  It does appear that your line suffers from a significant amount of RFI during the hours of darkness.

yep..  problem could be tracking it down :/ 
Bows to the black cats advice-  do you reckon it may be worthwhile trying the radio trick?

Hmm . . . At the moment, I am just suspecting significant amounts of RFI during the hours of darkness (something that I also suffer from) and not a source of REIN. That said, if Brian has access to a radio capable of tuning (down) to around 600 kHz (or, even better, 300 kHz) then a sweep of the house may help to settle the question: "RFI or REIN?"

Come to think of it, sight of the snapshot graphs -- QLN & Hlog, in particular -- may be illuminating. As an example I attach the relevant montage of data for my own ADSL2+ line, data harvested last June. The QLN graph shows how much RF 'noise' is present in the west Suffolk environment.   

[Brian Coat]

I managed to log 24h+ and compare good/bad bandwise snr

Interesting ...

[roseway]

Brian, I hope you don't mind my suggesting that those images are a bit fuzzy after combining the two into one and saving as a .jpg. It's generally better to post the original .png snapshots, which will be much clearer (and probably smaller as well).

[Brian Coat]

Thanks.

The jpg is only intended to highlight the frequencies where there are big night/day differences, not to provide hi-res quantitative data.

I've found and old Sangean SW tuner which also has 535-1705 kHz AM tuner on it so I'm going try to to tune it into the (receivable) frequencies where there are noticeable gaps between black and pink (e.g.Bands ~305-385, ~1300-1650 kHz)??

[Brian Coat]

Logs as suggested.

[burakkucat]

From the latest set of graphs we can see the typical pattern in the DS SNRM and maximum attainable DS connection speed plots, which result from increased RFI during the hours of darkness. (Note that yet another re-synchronisation event occurred shortly after 0600 hours this morning.)

It would be really nice to see the Hlog and QLN graphs, plotted from the data obtained at synchronisation time. If they are in anyway comparable to my graphs (remembering that would be an attempted comparison of VDSL2 with ADSL2+), then there is nothing that could be done . . . 

[Brian Coat]

Thanks very much for this.

"QLN and HLog at time of resynch" : before/after event logs attached.

I'm not sure how the modem/DSLStats s/w  are generating the HLog file but it looks suspiciously smooth?

"Yet another resynch": I've only seen/logged two since reconnecting the Zyxel a few days ago, both occurred 6-8 a.m. when the SNR curve is rising. Is this resynch rate a problem, it does not seem to be degrading the IP Profile?

"Normal SNR change night/day": I don't know anything about this - is it due to atmospheric effects? domestic rfi?

[roseway]

I'm not sure how the modem/DSLStats s/w  are generating the HLog file but it looks suspiciously smooth?

In DSLstats you can see the raw data reported by the modem under Telnet Data --> HLog.

[c6em]

"Normal SNR change night/day": I don't know anything about this - is it due to atmospheric effects? domestic rfi?

During the hours of darkness there is increased radio frequency interference in the atmosphere
All BB lines pick this up
So during the evening/night the SNR trace becomes jittery and the SNR value drops
It recovers the next morning
There are some lines where the RFI affects are so severe and the SNR drops so much that the line may re-sync at a lower value during the night.  These will mainly be rural properties with long overhead line sections.

The basic reason for this is the sun's solar wind which drags the inosphere section of the atmosphere upwards on the lee side of the earth -ie the nighttime side
This causes the "Heaviside Layer" high up in the atmosphere to rise up (its approx 75 miles up)
This layer reflects radio waves
This higher positioning of the layer allows RFI to propagate over the horizon far easier.
Hence the increased RFI during the hours of darkness coming from far and wide all around.
This is the same reason why short wave radio signals can be received from a long way away during the night while they cannot be listened into during the day.

[NewtronStar]

There was me thinking it was the street lights coming on earlier and heating systems coming on and house hold lights being switched on and eletric cookers and so on and totally forgot about solar activity and sun spots.

[kitz]

Nice explanation c6em.  Thank you.

[burakkucat]

. . .  the HLog file but it looks suspiciously smooth?

It certainly is smooth. Nice and smooth. So no obvious defects in the metallic pathway. 

However your QLN graph shows significant spikes between tones 150 and 370. By cross-referencing the frequencies of those peaks with the list of UK broadcast transmitters we might be able to deduce your location!

The peak at just above tone 160 will be due to the Radio 5 transmitter, broadcasting at 693 kHz.
The peak at around tones 244 - 245 will be due to the Talk Sport transmitter, broadcasting at 1053 kHz.
The somewhat broader peak at around tones 285 - 288 could be due to the effects of both the Virgin AM and Capital Gold transmitters, broadcasting at 1233 kHz and 1242 kHz.

Carefully watching the map as the markers were added makes me wonder if you are located in south east England . . . The county of Kent, perhaps? 

[Brian Coat]

Wow, C6em: Thank you for an excellent explanation of the ionospheric effects!

Yesterday’s night/day AM radio tuner experiment certainly bears this out – in addition to some of the channels which Burakkucat predicted, at night you start to pick up all kinds of ‘continental’ stuff plus a great deal of other noise.

Thanks, Burakkucat for your noise source map. I am based in the Midlands, not the South East as predicted by Burakkukat (so no need to apply to the feline section of MI5/GCHQ just yet ) but the map also has plenty of dots near where I live, so the logic is spot on.

And I do live in a “rural property with long overhead line sections” as mentioned by c6em.

So you have helped me to understand that … I am fairly near to transmitters broadcasting at problem frequencies and I have a large AM antenna in the form of overhead lines.

I do not think there is much I can do about these fundamental effects, except wait for BT to bring FTTC to my own (small) village. This could be a long wait!

That said I am very happy that my d/s IP profile is now over 6x what it was before I got a modem that allows access to the stats.
I am really grateful to Kitz forum members whose expertise and generous advice showed me how to access the data and what it means. Thank you very much.

This allowed me to convince BT there was an IP Profile problem, (something I’d not been able to do before, despite trying) and it also allowed me to learn a little about how it all works, which to an engineer is always a bonus.

THANK YOU!

By the way, this Zyxel 3300N modem is a nice bit of kit – it works well with BT Infinity and has many excellent features on the LAN/WLAN router side.