Internet Stalls

[neoapple]

A while ago I posted a topic about my internet stalling. Well i figured i would start a new topic as the old one was crazy old.

Basically my Internet will randomly just stall, i wont be able to access anything sometimes not even the cable modem.

I phoned up Last week and the people where running tests and basically blaming my equipment...... Even though I had point blanked said many times it not my equipment.

Was not until the guy said I'm going to run tests and what was awesome timing on my cable modems part was it decided to stall right then. he comes back and says did you reboot the modem and I was like no, this is what it does. So he then arranged for an engineer to come out and replace the cable modem as my upstream was fluctuating like mad, would not stay stable.

So the tech gave me a brand new modem as I was running some proper old school original 250 series ones, now been given a 256v2.

Anyway I had to activate the router etc as expected. Everything worked fine then suddenly it stalled again. I phoned up and tech support said it's apparently it's a fault in my area that would be fixed. Got a cal the next day and was told it's been fixed etc.

Well sadly I'm now reporting a week later that it's still playing up and I'm getting frustrated by phoning up Tech support all the time.
I have noticed when the internet stalls the cable modem don't give out a dhcp lease.

Things I have tried include the following:
Different pc's
Different routers even resorted to nabbing a spare CISCO Business router and configuring that from console.
Wiresharked traffic
Tried many different cables, cat5e, cat6, made my own as well.
Taken the routers out the equation and plugged the cable modem directly into the PC.
Ran cable tests on all my network equipment
When the internet froze i figured what happens when I plug a PC direct into it, thats when i noticed the modem refused to assign a DHCP.

So it's now safe to assume the issue is between the cable modem and where ever the hell it goes.

I have noticed that there is now alot of this in my CM Log
Sat Sep 25 08:51:38 2010 Sat Sep 25 08:51:38 2010 Information (7) New UCD in effect
Sat Sep 25 08:51:38 2010 Sat Sep 25 08:51:38 2010 Information (7) Starting on-the-fly UCD change.

Downstream Lock : Locked
Downstream Channel Id : 53
Downstream Frequency : 314750000 Hz
Downstream Modulation : QAM256
Downstream Symbol Rate : 6952 Ksym/sec
Downstream Interleave Depth : taps12Increment17
Downstream Receive Power Level : 7.2 dBmV
Downstream SNR : 40.7 dB

Upstream Lock : Locked
Upstream Channel ID : 4
Upstream Frequency : 35800000 Hz
Upstream Modulation : QPSK
Upstream Symbol Rate : 2560 Ksym/sec
Upstream transmit Power Level : 32.7 dBmV
Upstream Mini-Slot Size : 2

It's now becoming a major joke


Edit
Also noticed that now my upstream power transmit level is changing AGAIN!
1 min after i posted this it's now at
Upstream transmit Power Level : 43.7 dBmV

[jb66]

Is it also changing between qpsk and qam16?

[Peter_]

The upstream is a tad low.

[moaningmags]

If it's jumping from 32 to 43 in the space of a minute you need a tech visit to sort it out.

[neoapple]

I have no idea if its changing from QPSK

What should the upstream be for the Bromley area?
Also my upstream is sitting at 39.7 and has been for some time now.

But the thing is what will an engineer be able to do round my house? They already replaced the cable modem and the cable run seems fine anyway.

I dont want to wast a days holiday again to basically find out "this *may* fix it etc"

[jb66]

The tech won't have a clue, second line might know why your upstream is changing, all the tech can do is notify network dept.

[Ignitionnet]

Par for the course for VM technical support blaming your equipment I'm afraid. The issue is on the cable network, upstream problems. Upstream shouldn't be QPSK and shouldn't bounce around like that in terms of power.

The UCD messages you see are the cable network adjusting parameters to try and keep you connected.

This should already be on the job list of the network guys. Certainly was an area fault, obviously not yet fixed.

[Kymmy]

Should any healthy connection be on QPSK for the upstream?? I only ask as mines always been QPSK since I got cable

[Sephiroth]

I keep a full record of my event log going back to November 2009.

On the rare occasions that my network has frozen, I too can't get into the Cable Modem even though it shows ready and the router says it's connected to the CM. Only a cold reboot of the CM resolves that situation.

In July 2010, I had been switched from cpc2-rdng21 to cpc8-rdng20 and my modulation changed from 256QAM to 64QAM. Shortly afterwards, the on-the-fly UCD changes started to occur and I was regularly switched betweem 16QAM and QPSK.

I asked Igni for an opinion and he suggested that we might be beinbg resegmented for the 10:1 programme.

On 03-Sep, we were moved to another subnet on rdng20, since when, despite several on-the-fly messages, we have remained on 16QAM upstream - at least so far as my daily sampling has revealed.

Furthermore, we've never had such good performance. It was never poor before but never so good in peak time as it is now, although yesterday there was a little bit of upstream "reluctance".

Through all of this, the upstream power remains around the 35 - 36 dBmv mark and doesn't jump wildly. When people say the upstream power is low, if what you send delivers 15 dBmv to the first active node on your route (AFAIK), then job done if nothing else is wrong. On the other hand, a duff modem (which I'm sure you don't have in the light of your story) might not deliver enough power.

There are better techs than me on the forum who know the architecture detail. But my networks mind tells me that if it fluctuates wildly, there is an electrically unstable situation. This could be between home and street cabinet (like upstream noise ingress, or dipole effect due to corrosion which will vary with humidity and temperature). It could be the same at the CMTS end, although that would affect a larger number of users.

So IMHO, jb66 has called it right in your case.

HTH.

[Ignitionnet]

Should any healthy connection be on QPSK for the upstream?? I only ask as mines always been QPSK since I got cable

Yes, it can be absolutely fine for connections on the 'older' network. This area has no upstream congestion despite being on it.

On the 50Mbps DOCSIS 3 overlay network though it's unusual and indicates issues.

[neoapple]

So what am I going to have to do again?
The Upstream power level is still sitting at 40.2, with more New UCD stuff in the Log.

There are no known area faults for my area. I'm at a loss.

[Peter_]

So what am I going to have to do again?
The Upstream power level is still sitting at 40.2, with more New UCD stuff in the Log.

There are no known area faults for my area. I'm at a loss.
Does the connection seem stable at the moment.

[neoapple]

It's currently stable, but i have no idea for how long etc. It can stay stable for a while then wham it goes mad etc!