We have this strange issue with 10dB or more of US level fluctuations and it eventually causes the modems to drop one or more of its US channels. This does not seem to be an issue with the hardline network since it is only affecting random modems while a neighbors modem on the same tap is fine. The problem seems to go away for awhile if we reseat the drop at the tap or ground block. We have tried changing fittings and it still seems to come back after awhile (next day or next week). But the problem immediately fixes itself just by disconnecting and reconnecting one of the fittings in the drop circuit, seems like any fitting will do it. It almost appears to be like some static charge builds then discharges when we break the connection. The other odd thing is the downstream is not affected. We are monitoring levels on US and DS and can see only large fluctuations in the US. This did not seem to be an issue until we moved to D3 with bonded US. A large amount of Arris and some Moto D3 modems found so far. Firmware is up to date. Any thoughts?
Thank you
Jon
How many upstreams are you running in each node?
Do your nodes have the old FP lasers, or the newer DFB lasers?
We are running between 3 and 4 US channels depending on CMTS licensing. All of our FP's have been upgraded to DFB lasers (laser diode swap out - upgrade).
The house the noise was coming from. This is due to the tap value loss decreasing the ingress noise into the plant as well as the signal from devices to and from the drop.
Air Conditioning Cambridge
How is your MGN looking at the poles? Good bonding to the ground blocks at the affected houses? Have you tried any reverse sweep at the affected houses to see how the noise floor looks?
What is MGN at the pole? - No we have not verified the bonding, but that sounds to be a good idea. Not sure about the reverse sweep from the house showing the noise floor. Do you mean looking at an analyzer to see return noise from the home? It is my thought that a high noise floor would affect all the modems on that return path not just one. Another thing to mention here is that most of these issues seem to be comming to light as we upgrade the customer from a D2 modem to a D3 modem.
MGN = Multiple Ground / Neutral. This is where everything (power, utilities) bonds to ground at the pole.
Noise at one house would indeed affect multiple devices in that network segment, but to a lesser degree than the house the noise was coming from. This is due to the tap value loss decreasing the ingress noise into the plant as well as the signal from devices to and from the drop. So noise at a house degrades the system as a whole, but may cause instability at the source.
Bad headend pad? Bad Splitter? I ran into both of these in my D3 turnup, new out of box's too.
Also, when you upgraded your return lasers, did you optimize your return links at the same levels they were, or did you redo it based on your new channel loading? Because you're going to have alot more RF energy hitting your return laser with your bonded upstreams, you could be clipping your lasers.
Look at these, (optical starts on slide 25 I think) they're basicly the same, but the 2nd one has some different slides past the halfway mark. In particular, it starts talking about composite power on slide 35.
http://www.sctedakotaterritory.com/library/Cisco%20Return%20Path%20Optim...
http://www.scte-rockymountain.org/Documents/Seminars/Return%20Path%20Fam...
Capm - thanks for those slides, great info.
We are finding these are all individual home issues. Several bad fittings (especially RG11), F81's and ground blocks. We have had good luck using our TDR to find and fix these and are batting 1000 on repairing the last 15 issues we rolled out on. Symptom is unstable return levels and occasional offline. Reason bad connectors affecting mostly the low freq return band. We have found that if the return transmit moves more than 3dB - there is usually a problem that should be fixed. Most modems are very stable in the US power level. Lucky we have software that allows us to see the history in a graph which makes it very easy to see this issue. So far we have not found any plant issues other than one bad tap, the rest are from the tap to the home. The reason we thought it was only a D3 issue is becacuse we noticed many D3 modems dropping US channels and this is what caught our attention. D2 modems did not have any US channels to drop so they stayed online. We have since seen the US level issue on D1.1 and D2 modems (they seem to handle the problem better) Like I said before we have more than a hundred of these just on the D3 modems.