T3 Timeouts (and corresponding internet drop-outs), RNG-RSP CCAP Commanded Power in Excess of 6 dB, and other issues.

In the spirit of information sharing; a second tech was scheduled after I dumped all my information with support in direct messages. I did ask for someone on the outdoor/plant side, but, they sent another regular tech.

He looked over/verified everything that had been done and looked over my information as well. He agreed with my assessment about removing all attenuation. It was likely appropriate for local plant conditions based on what he's seen in the area and my particular kind of modem (Arris.)

Phantom Modem

After I scheduled the appointment, the chatbot wouldn't show me the phantom modem any more since it wants to use a different troubleshooting workflow.

I'm waiting for all that stuff to clear out and will have another round in private messages to try and get that removed. It's an account-level thing, so, nothing anyone here can really benefit from seeing (and there's nothing we can do on our end as customers to fix it.) Support chat requested I send a screenshot since they cannot see it in their normal account tools, so, I'll continue to work with them to sort that out once I get the bot to show it again.

RNG-RSP CCAP Commanded Power in Excess of 6 dB

Both technicians I've been sent were unfamiliar with this message or its ramifications.

From my reading and research, including having to read parts of the DOCSIS specification...it is my understanding that this is when the CMTS tells the modem to broadcast at a higher power and the modem refuses. This is likely the primary indicator/cause of my packet loss and drop outs.

Once I removed all the attenuation on my line, this event stopped showing up and it was likely the most important piece of information I could get out of the modem. If the CMTS needs more power and the modem won't provide it, your connection is going to suffer. If you have a modem that doesn't show your event log, I don't know how you'd figure this out.

Splitter

The second tech said that the splitters they use tend to be higher quality than a simple pad, so, that's why the previous tech used one. -7.5dB is what would be expected given my local connection.

I have +10dBmV signal on the downstream and according to the second tech, they're trained here to simultaneously try to get downstream power closer to 0 and upstream to broadcast a bit higher.

In my case, a -7.5 would put me around +3 on the downstream and +47 on the upstream. But that's pretty close to my modem's maximum (according to the datasheet) for 4+1 upstream channels, so there's not much room for error if the CMTS needs more power.

To be clear, I've always had heavy attenuation on my line from when Xfinity buried a new one for me 5 years ago. My issues got worse relatively recently, so it's simply possible that local plant conditions have changed. Fortunately, my modem can handle +15dBmV on the downstream. Therefore, I can get away with having no attenuation at all and keep its upstream broadcast power well within its capabilities.

If local conditions change such that my downstream goes higher, I will be forced to attenuate the line somehow and it'll be a fight to figure out what the modem will be happy with, because so far, even a relatively small -3dB pad has proven problematic (even though it shouldn't be). That said, the one -3dB pad I have might be faulty. If I end up having to attenuate the line I intend to try a forward-path attenuator. Failing that, I'll get the best -3.5 splitter I can find.

This won't work for everyone, you need to know what your modem can handle, what the local plant will allow, and you have to have enough headroom on your upstream to get away with it if you're out of spec in the other numbers. The second tech told me that +14 is the local maximum so I'm realistically cutting it close.

T3 Timeout Events

I still see these a fair number of these, but, they don't seem to be impacting my connection. Apparently I'm not missing enough in a row to cause the modem to reset the connection, at least.

Interestingly, I did find an instance of a T3 timeout with no change in corrected & uncorrected symbols. I originally thought they were always associated with my bursts of uncorrected symbols, but now that I'm very closely monitoring things, I can see they are happening all the time.

Something with the mgmt layer in the local CMTS? Could even be normal given the number of profile changes and various balancing things that happen.

Can't be noise related if they happen when the counters aren't ticking up.

Bursts of Uncorrected Symbols

Neither technician could offer much help here since it's both sporadic and could be anything from a failing amplifier to a someone two houses down using their microwave. If it doesn't happen while they're hooked up the line with the analyzer, it's impossible to trace. It has not happened either time they've been here and when hooked up.

Time will tell if this is resolved or not by the modem no longer denying the CMTS more power. Based on my research, it could simply be that when the CMTS needed more power and the modem refused, this was the fallout.

I will continue to operate my monitoring script and collect data.

Drop Outs

Since everything is technically working, the technician wasn't comfortable making changes to anything (I agreed with this assessment.) He did double-check everything, talked to me about my understanding about a few things, and we're in wait-n-see. So far, so good, though since removal of all attenuation.

If the drop outs come back, there is some super old equipment in my pedestal that they might want to swap out (not the tap, that was new.)