Print

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

[Jasonkruys]

Hi guys, new to the forum! I have been observing as a guest for some time, and am amazed at the knowledge of some of you guys - I am trying to learn as much as I can from you!

I am on Sky's FTTC product, and achieving a sync of about 42meg (ping 22ms), after being sat at 46meg (ping 9ms) quite happily for a month. Having looked at the connection stats, using some of your awesome work (both firmware for the HG612 and monitoring software), it has become obvious that the doubling of ping is down to the application of Interleaving (depth 800 odd!) and have also noted the associated reduction in sync.

Now as far as I know, there was no change in my house (or outside, such as weather) that might have affected the line such that DLM has interfered - ideally what I would like is for DLM to ultimately return the line to its previous state. I am assuming there is some noise or interference causing this - might my connection stats offer any clues as to where to start looking? Or indeed, is the a particular graph or anything that I can look to improve by trying random things and monitoring the connection?

My line is underground, and comes straight into the side of the house to an NTE 5 socket. The connections look good, and there is no surge filter on this one. I have a VDSL MK2 faceplate fitted, and a 30cm shielded Belkin modem cable running to my HG612. There are no mains cables in very close proximity to the cable.

Any insight you guys can offer would be appreciated. Ultimately I would like to learn as much as I can about interpreting the stats!

Thanks,

Jason.

[Bald_Eagle1]

Ideally, we would like to see more days worth of ongoing stats that MIGHT just have captured what has changed to cause Interleaving to be applied, resulting in lower sync speeds.


Do you have any older ongoing stats in your modem_stats.log?

If so, when running graphpd.exe, could you plot them for 'xx d' where it asks you to enter the period to plot (xx = the number & d = days)?

e.g. 6 d would equate to 6 days

[Jasonkruys]

Hi bald_eagle, thanks for the reply. Sadly I don't have any more stats at the moment. It was primarily the sudden change in connection that prompted me to bite the bullet, be brave, and set everything up lol. I guess it is just going to be a case of monitoring over a period of time and see what happens. I was hoping that something might jump out of the snapshot, perhaps missing bits or a do in the snr or something!
I will keep up the monitoring. Hopefully a positive change will occur, and I can see what has happened. And then should it change again as it had now, I will be able to see what caused it!

Thanks for looking!
Sent from my awesome 1520 using Tapatalk

[Jasonkruys]

Hi guys, me again!

A couple of questions, if some of you experts might be willing to help? I have been uploading data to mydslwebstats (awesome effort on that, thanks!) and comparing my stats with others. I have noted that my stats are very very similar to Ronskis - even to the point where on the Bits SNR per tone graphs, we have dips down to zero in the middle of ranges in the same place.

First Question - are these dips at this particular frequency 'normal' / what might be causing them?

Second Question - Despite the similarity, Ronski is using tones much higher than mine - mine cut dead at 2143. Is this indicative of a particular issue or mode of interference, or just a characteristic of my perhaps slightly longer line?

Many Thanks,

Jason.

[Bald_Eagle1]

The dips to zero are not ideal, but they can be seen on quite a few 'longer' connections.

I have them on my 1100m connection, but more at DS than US frequencies/tones.



It might be best to compare notes with Ronski.
He does frequent this forum fom time to time.

Possibly some aluminium in the D-side cabling, 'noisy' power supply/UPS charger transformer, crosstalk, RFI or a similar set up at home etc?

[Ronski]

I'm 450 meters from the cab (ECI). My modem is next to my router, switch, server and UPS, so probably quite electrically noisy. I have a very short (0.5m) Cat6 cable connecting the master socket to the modem. My line has never achieved my original estimate of 57Mbps, but has fluctuated between 40 - 47 since I've had it, which is just over two years now.

[NewtronStar]

When looking at your BandPlan you have an ECI cabinet the HG612 modem is Huawei, when the ECI cabinet and ECI modem are matched they do seem to give a higher sync and with less errors than an HG612 modem to ECI FTTC cabinet.

Yes I know you will not be able to see your stats with the ECI modem unless it's hardware hacked, my question to you have you got an ECI FTTC modem and are now just using the HG612 to look at your stats ?

[Jasonkruys]

Hi guys, thanks for the replies. NewtronStar - no I don't, it was an HG612 that the openreach engineer installed when they installed fibre.
Sent from my Lumia 1520 using Tapatalk (with free typos)

[Jasonkruys]

For the sake of 20 quid, I have purchased an ECI modem to give it a try! Will miss the stats though 😞
Sent from my Lumia 1520 using Tapatalk (with free typos)

[Ronski]

When I tried an ECI modem my sync was lower. It is now no longer suggested to match the modem to the cabinet, although it is worth trying. Some get better performance with an ECI, whilst some are better off with the HG612.

[boost]

g0t DECT?

[kitz]

Im on an ECI cab and matching it with a lantiq based router hasnt given me any better performance than what I got with the HG612.

For the sake of 20 quid, I have purchased an ECI modem to give it a try! Will miss the stats though

If you dont mind playing, there are ways of hacking ECI modems
http://forum.kitz.co.uk/index.php?topic=11704.0

[Ronski]

g0t DECT?

If that's directed at me, yes we have. It's located in the hallway though.

[Jaggies]

Im on an ECI cab
<snip>

How can you tell what type of cab you're on?

[Black Sheep]

You can tell by the tone bandwidth allocation, I believe ?? ECI and Huawei differ with their allocation. One of the script-graphing members would be able to give a better in-depth explanation, I'm sure ??