Print

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

[bra1nDeaD]

Hi,

I'm new here and relatively new to tweaking my VDSL connection. I've got an unlocked HG612 modem and wired this right next to my main phone socket with really short CAT6 wire. I'm reading the stats with DSLstats, but I do not understand 80% of the lot.

So I was wondering if there is anything else I can do to optimise my VDSL connection any further. It tells me the max down is 39620 Kbps, but I only get 32400 Kbps (up is already at max). Or is it even possible to go even higher? As BT shows VDSL high range (Clean) is 43.6 Mbps. Or am I already at the limit of my copper line?

I've attached the DSLstats and Exchange info.

One thing that strikes me is this (though I don't know why as I don't fully understand what it means):
         Down   Up
SNR (dB):       10.3       6.1
Attn(dB):       25.7       0.0
Pwr(dBm):    13.5       3.4

Attn Up = 0.0 ?

This is from Telnet, not GUI (saw others have similar through a GUI bug).

Not sure this is relevant anyway, just trying to get the max out of my line.

Much appreciated.

[mofa2020]

Welcome to the forum 

Your line is banded for some reason since SNRm raised to 10.3db the normal should be around 6db, how did you get a CAT6 cable fit in RJ11? I tried this before and another kitzen tried this and the service was dodgy as a result of trying to use it.

[bra1nDeaD]

Welcome to the forum 

Your line is banded for some reason since SNRm raised to 10.3db the normal should be around 6db, how did you get a CAT6 wire fit in RJ11? I tried this before and another kitzen tried this and the service was dodgy as a result of trying to use it.

I used a BT master socket MK4 (image attached) where you can hard-wire (without plug) straight into the filter. Then I crimped an RJ11 on the other side (tight fit) and plugged that into the DSL port. The wire itself is only about 10cm long.

Reading into SNR, it seems that this is set by the exchange hardware to make sure the line stays stable (but results in slower speed). I read that some modems can tweak this value, but not the HG612?!? Could this potentially be a result from using CAT6 instead of CW1308? Though it's not a long wire. Is there another way of reducing this value? Should I get in contact with my provider in case they can lower it?

[mofa2020]

Reading into SNR, it seems that this is set by the exchange hardware to make sure the line stays stable (but results in slower speed).

Yes that DLM role to see the line performance and determine the best values to keep service running, in case of something wrong it will kick in and take actions one of them is banding the line speed so service will keep running.

The HG612 can not tweak SNRm AFAIN and even with that you will end with unstable service because you are forcing a SNRm value that service can not be stable at or DLM would set it for your line in the first place, I would give the cable supplied with the HG612 (or any modem/router) a try and see how things go with it "DLM may need a day or two to take action".

*One last question are you on ECI or Huawei cabinet?

[bra1nDeaD]

Yes that DLM role to see the line performance and determine the best values to keep service running, in case of something wrong it will kick in and take actions one of them is banding the line speed so service will keep running.

The HG612 can not tweak SNRm AFAIN and even with that you will end with unstable service because you are forcing a SNRm value that service can not be stable at or DLM would set it for your line in the first place, I would give the cable supplied with the HG612 (or any modem/router) a try and see how things go with it "DLM may need a day or two to take action".

*One last question are you on ECI or Huawei cabinet?

Thanks, I'll try the other wire.

Before I bought the HG612 I checked the cabinet and AFAIK I'm on a Huawei cabinet. I compared the shape of the cabinet with pictures I could find online and did some online searches of my cabinet.

Will that affect the speed as well?

[mofa2020]

Will that affect the speed as well?

No it won't, I thought your line maybe affected from G.INP trial if it is on ECI cabinet that is why I asked.

[burakkucat]

Welcome to the kitz forum. 

         Down   Up
SNR (dB):       10.3       6.1
Attn(dB):       25.7       0.0
Pwr(dBm):    13.5       3.4

Attn Up = 0.0 ?

Don't worry about that. There is nothing wrong. It's essentially a "cosmetic blemish".

There are times when the transceiver units, at each end of the link, fail to exchange all of the parameters. In this particular case, the modem's transceiver unit has sent a signal (at a known power level) to the transceiver unit of the DSLAM's line-card port for your circuit. By measuring the power received, the DSLAM can determine the attenuation for the US direction. However, for whatever reason, the DSLAM has failed to notify the modem of the attenuation it has determined. Hence the modem shows 0.0 dB.

Reading into SNR, it seems that this is set by the exchange hardware to make sure the line stays stable (but results in slower speed).

No, not "exchange hardware" but by the dynamic line management process of the cabinet based DSLAM.

I read that some modems can tweak this value, but not the HG612?!?

No. The DSLAM will reject any attempted tweaks. Attempts to do so will also violate the terms of SIN 498, Section 2.2.1, Dynamic Line Management.

Could this potentially be a result from using CAT6 instead of CW1308? Though it's not a long wire. Is there another way of reducing this value? Should I get in contact with my provider in case they can lower it?

No, no and, er, no.

I've taken a look at the plots contained in the Graphs.zip archive you attached to your opening post. Unfortunately you have over-sized those which have the bandwidth as the x-axis. For a VDSL2 (ITU-T G.993.2) Profile 17a service there are a maximum of 4096 sub-carriers (numbered 0 to 4095). With the x-axis scale set to nearly double that required for the 17MHz bandwidth, the plots are rather compressed and not that easy to interpret.

However all is not lost . . . for you also attached the TelnetData.zip archive. I was able to massage that data into a suitable format and produced the four usual snapshot plots which give an overview of the circuit. Attached, below, is a montage of the four snapshot plots.

Looking at the Bit Loading, the SNR and the Hlog plots I see nothing particularly abnormal. They tell me a mid-length line, with a small amount of downstream power back-off in the DS1 band and with no obvious problems in the metallic pathway. The is no evidence of any significant degree of cross-talk.

The QLN plot is unpleasant.    There are no abnormally powerful discrete signals present, just a systematic rise in overall noise as the frequency decreases. Does the line run near any mains wiring? Or close to any other mains powered equipment?

[bra1nDeaD]

Welcome to the kitz forum. 

Thanks 

Don't worry about that. There is nothing wrong. It's essentially a "cosmetic blemish".

There are times when the transceiver units, at each end of the link, fail to exchange all of the parameters. In this particular case, the modem's transceiver unit has sent a signal (at a known power level) to the transceiver unit of the DSLAM's line-card port for your circuit. By measuring the power received, the DSLAM can determine the attenuation for the US direction. However, for whatever reason, the DSLAM has failed to notify the modem of the attenuation it has determined. Hence the modem shows 0.0 dB.

No, not "exchange hardware" but by the dynamic line management process of the cabinet based DSLAM.

No. The DSLAM will reject any attempted tweaks. Attempts to do so will also violate the terms of SIN 498, Section 2.2.1, Dynamic Line Management.

No, no and, er, no.

Thank you for the clarification, it helps to understand it better.

I've taken a look at the plots contained in the Graphs.zip archive you attached to your opening post. Unfortunately you have over-sized those which have the bandwidth as the x-axis. For a VDSL2 (ITU-T G.993.2) Profile 17a service there are a maximum of 4096 sub-carriers (numbered 0 to 4095). With the x-axis scale set to nearly double that required for the 17MHz bandwidth, the plots are rather compressed and not that easy to interpret.

However all is not lost . . . for you also attached the TelnetData.zip archive. I was able to massage that data into a suitable format and produced the four usual snapshot plots which give an overview of the circuit. Attached, below, is a montage of the four snapshot plots.

I can change the range, but it was already on the pre-selected "VDSL2" range. I've changed it to a custom range now which is closer to 4095. DSLstats don't give me a similar graph as the QLN graph you plotted from the Telnet Data. Good to know the data is there though. Can I assume that the line should show as smooth as possible? Without the spikes?

Looking at the Bit Loading, the SNR and the Hlog plots I see nothing particularly abnormal. They tell me a mid-length line, with a small amount of downstream power back-off in the DS1 band and with no obvious problems in the metallic pathway. The is no evidence of any significant degree of cross-talk.

The QLN plot is unpleasant.    There are no abnormally powerful discrete signals present, just a systematic rise in overall noise as the frequency decreases. Does the line run near any mains wiring? Or close to any other mains powered equipment?

The line coming into the house goes straight through my loft cavity wall into the back of the master wall socket. From there is goes into the modem with only a 10cm wire. I do have a phone line wire plugged into the filter which has a long run to my server rack which might run along a whole bunch of mains wires. Don't know if this might cause issues or not. I will do a test later today and plug the modem straight into the test socket without phone connected and see what data it provides then, and compare.

Thank you so much for pointing me in the right direction here. I'll post back when I have some results. I hope it clears the noise as that means I can do something about it. If the noise is still there then not sure if this can be cleaned up as the wire goes straight to the roadside pole from my master socket.

[ejs]

No. The DSLAM will reject any attempted tweaks.

This is incorrect. The DSLAM doesn't reject anything. Lantiq (and MediaTek) modems can operate with an adjusted downstream target SNRM, Broadcom modems cannot.

Contacting the service provider to try to get a DLM reset to remove the banding might be the most useful thing to do.

[bra1nDeaD]

So I've had my modem connected via separate filter straight into test port (face plate removed) so there would not be any interference of anything else. From what I can tell there is still a lot of noise. This graph shows tone number > 1980 as well, which look even worse.

When looking at the HLog graph there is a bit for upstream missing, and after that the line zig-zags all over the place. Not sure what that is about though.

Does it take a while for these statistics to improve? Shall I leave it a little bit longer and check again? Or doesn't that matter?

If this is what it is, I suspect that I can't do anything myself to improve it further, can I? I suspect that any further improvements require BT Open Reach to get involved.