Comcast XFINITY → Re: What to do about 11 weeks of node/network problems?

Some clarification about my notes:

1. By "line tech" I am referring to an actual network line tech (the guys who drive around with bucket trucks and fix problems), not a service technician (the guys who come to your house and check everything between the cable modem all the way up to where the cable comes in off the pole).

2. Line tech was not able to see the issue in the field (was unable to see bad SNR for 747MHz or higher). I wasn't give any details other than that; just that "in the field" he couldn't see the problem. He stated that he was going to go to the head-end later that day and see if he could see it from there.

3. The follow-up call I got consisted of him saying the following, which didn't make any sense to me (this is why I put "at the node" in quotes): "I went to the head-end and yeah I can see the problem there at the node". I said "the node? Did you mean the head-end?" and he said "no, at the node".

As I understand it, "the node" refers to the cable node/segment (which correlates with a small or large geographic area, i.e. a "segment" of the cable network from which customers get their drops) while "head-end" means where (or near where) the CMTS is. So this is why I put "at the node" in quotes because I'm not sure what exactly he meant by what he said, given the terminology I've learned. I often get the impression all these techs (service and line/network) "re-use" terms to describe multiple things.

4. Line tech stated that after chatting with fellow techs about this problem, one reminded him of this issue happening in San Carlos (about 15 miles from here) but he himself had forgotten about it. He then told me the story about having to replace a long run of cable up there which looked fine but obviously was taking noise. I asked how they managed to figure out it was that cable. He said there's a device they have which allows them to look for leakage, but it's a long and drawn out task because it requires 2 people and a lot of walking.

Now for a question:

Doesn't a "truck roll" refer to sending out a line/network tech? If so, from whom do I ask this? Comcast? I have talked to Comcast CSRs (800-COMCAST) more times than I can count, and have requested line/network techs twice already -- what they send is a standard service tech every single time, who goes "yeah you have your ducks in a row, sounds like a network problem, not much I can do, you should talk to that network/line tech you've met". (Yeah that's nice, I would if I had his surname/extension...)

Or is this something I'm supposed to ask my LFA?

The node is a glorified media converter essentially. It takes the optical feed from the hubsite and converts it to the coax running around neighborhoods. The CMTS generally exists at the hubsite level not the headend level. Some might reuse headend to mean hubsite but with everyone having their own ways of saying things it's really anyone's guess.

Think of the headend as the main aggregation point for a geographic area. It feeds many different hubsites that themselves feed many different nodes. There may be more than one headend for an area depending on size, Atlanta has two that I know of: Stone Mountain and Vinings but I believe there are two others that probably came from other cable system acquisitions and general expansion.That is a cable leak detector and there can be vehicle and hand-held versions. Looks like they have really changed since the last one I had my hands on. The one I had really resembled a radar gun and had two telescoping antennas coming out each side. It could require two people but sometimes will be done with one depending on what they are looking for.

If they are looking for signal egress (signal leaking out) that can be done with one person driving around until he/she gets to the general area and then going on foot. If they are looking for signal ingress (signal making it's way in), that really does require two people. One to sit at your house and/or designated test points with a sensor and the other to drive around broadcasting a specific signal that sensor will be looking for.

Suffice it to say, depending on the area this issue could be fixed either from you pushing them or the FCC/FAA pushing them. If you live near a major airport or are in a major flight path, the FAA will do flyovers and identify hot spots that need attention. As such, your area may get flagged and fixed eventually even if you don't push them (not that I'm saying you shouldn't try, just that it may one day "fix itself" so to speak).

He could have had the headend techs check the transmitter for your node. They could then see whether or not those last two QAMs had MER/BER issues or if the signal from the NSG was degraded on those last two QAMs. If he truly saw the problem at the headend for your node, then that means those last two QAMs are degraded at the NSG because of a faulty card. A simple card swap at the headend and it's fixed. However, I doubt that's the problem or else it would have been fixed already. So, honestly I'm not sure what he meant there.

What you call "line tech" is what Comcast calls "Maintenance Techs". Maintenance techs do have the ability to walk the line with sniffer guns to locate where interference is entering the plant or where signal is leaking. I don't know how CSRs dispatch calls these days, but next time you call in, specifically request that you want a maintenance tech or a CommTech 5 tech to come diagnose an issue with the plant.

Also still contact the LFA and report the problem. Municipalities love fining cable companies for not being in compliance with maintaining their plant. Fines tend to jump start maintenance efforts.

Isn't it basically their S.O.P. that the phone reps can not do this and that only the premise facing techs can escalate an issue to the maintenance / network dept.

This is correct. The CSRs will open up a new CR (ticket/case) and will only schedule a standard service technician to come out. That technician and his supervisor are the ones who make the decision to "escalate the problem to network" (as in make a case and hand it off to network/line techs (maintenance techs) to deal with), and that's quite literally the last the customer ever hears of things (i.e. the assumption is made by all involved that network/line will resolve the issue). I speak from experience on this one. :-(

In my case, I was able to meet the network/line tech (maintenance tech) solely because of a Comcast employee on the official Comcast Forums (who shall remain anonymous per their request) "putting in a good word" for me behind-the-scenes + reaching out to someone at my local repair office (in Menlo Park) to look at the case. It was total chance that I actually *met* the guy -- I happened to be looking out my window the next day and saw a Comcast truck (with bucket) outside, where the driver got out and wandered around near our property looking up at utility poles. I went out and talked to him, asked him if he was handling my CR #, and he was. That's the guy, by the way, who I mentioned above (re: went to head-end, reproduced issue with 747MHz and higher, determined issue was same as what was seen in San Carlos).

If I call 800-COMCAST, all the CSRs are going to do is open up a new CR/case, link it/tie it to the old ones, and schedule to have a service technician come out. I've been down this road 3 times already, and all it does is waste service technicians' time since they can't solve the problem.

Thanks, HeadendJoe, and JoelC707 (hi dude ).

As surprising as it might sound, I actually understand most of the terminology you've used, sans two terms: MER and NSG. I know what QAMs are, and I know what BER is (bit error rate; from my experience with DS1/DS3/OC3/OC12 and SONET), but the other two are new to me.

I put in a call to the "general manager" who I'd spoke to in the past (about service technicians). I think I'm going to let things sit until Wednesday.

If I don't get a call back from him by then, I'm going to try an alternate approach -- I do have a contact at Comcast who handles the PR and IP networking side of things, and I have a good relationship with that person. Possibly they can get me in touch with someone who can drive this.

Otherwise if that goes no where, I'll give my LFA a call and approach it from that angle.

I guess on the positive side, there's lots of hard data/evidence that I can provide, and there is a part of me (the engineer part ) that's happy that Comcast was able to confirm existence of the issue. Sometimes tracking down a problem is the hardest part. In this case though it seems to be getting it fixed that's a pain.

Will let folks know what transpires.

This is a NSG »harmonicinc.com/product/ ··· -9000-6g

Essentially its an edgeQam that receives optical info and generates the info into RF QAM channels. So how Comcast works is that their CMTS routers are uBR10012 Cisco routers. The Downstream info (signal from the headend device to the node) leaves the router via fiber to the Harmonic NSG. The NSG outputs the info into multiple RF QAMs. Those QAMs eventually reach a transmitter where its converted back to optical, sent to the node, and then converted back to RF. Comcast also uses NSGs to convert VOD optical streams to RF QAMs.

Just a few definitions for you. MER (Modulation Error Rate) is the digital equivalent to SNR, which is the baseband measurement equivalent to Carrier to Noise. MER measures the ratio of error power to average power in an ideal QAM signal. MER measures the difference in quality between the transmitted modulation of a digital signal and the received modulation. Carrier to noise ratio is the measurement of the distance between a modulated RF carrier and the inherent noise floor. You can find a little more detail about these on Wikipedia.

Keep us updated for sure, hopefully it will get fixed. If you end up calling the FCC or your LFA, make sure you ask about Comcast's POP (proof of performance) testing. I believe by law the public has the right to review the outcome of those tests. They do the tests twice a year. May be worth glancing over if you're interested in where they found leaks in the plant during the test.