CM700 high US/UnCorrectables

Hey FURRYe38,

I’m having a hard time deciding between:
attenuator/splitter or NO attenuator/splitter.
Could I get your opinion? Opinions from others are welcome as well.
The ‘attenuator’ is just a 2-way splitter (-3.5dBm) with only 1 used Output, I left the unused terminal OPEN. If you feel I should connect a 50/70 Ohm resistor between signal conductor and GND rather than leave it floating, let me know. This attenuator is connected at the drop.

Here are some details/stats.

NO ATTENUATION:
_________________________
Downstream:
My DS PWR Level for Channels 1-24 range from +2.4 to +3.5 dBmV.
SNR/MER Range = 41.4 to 43.1dB
The following stats are from 50 Hours of UpTime with ~3.85E9 UnErrored CodeWords/Channel:
Correctables range from 500-2500(as frequency increases, Correctables INCREASE)
UnCorrectables range from 400-600(as frequency increases, UnCorrectables DECREASE)

Upstream:
PWR Levels:
46.8 dBmV
47.0 dBmV
48.0 dBmV
48.0 dBmV
_____________________________________
WITH ATTENUATION:
Downstream:
My DS PWR Level for Channels 1-24 range from -0.8dBmV to +0.0dBmV.
SNR/MER Range = 42.0 to 43.5dBmV.
The following stats are from 18 Hours of UpTime with ~2.87E9 UnErrored CodeWords/Channel:
Correctables range from 13-378 (as frequency increases, Correctables INCREASE)
UnCorrectables range from 62-98 (seemingly no correlation with deltaFrequency)

Upstream:
PWR Levels:
49.8 dBmV
50.3 dBmV
51.3dBmV
51.3 dBmV

I would copy/paste my logs, but all I have are screenshots of the old log, split between 2 separate images.

I noticed that WITHOUT the -3.5dB drop, my DS levels were all ~ +3.5dBmV, so a 3.5dB drop here would be perfect to hit unity. But as I expected, installing the splitter brought my (Already borderline HIGH) US levels From ~47.5dBmV up to ~ 50dBmV. I vaguely remember reading specs for the CM1000 indicating it was capable of ~54dBmV. This may be standard for all Modems, but still.


So based on the trade-offs described above.. would I be better off WITH or WITHOUT the splitter? Thoughts? Concerns?

I’ve noticed no differences in Download/Upload speeds via Speedtest.net.
However, I just noticed that PingPlotter is currently showing some problems. Like an elevated Latency Floor referenced to when I first set up my CM1000. The Jitter directly following CM1000 setup was <5mS. Currently it is regularly jumping to 30mS, averaging around 10-15mS. I’m going to remove the splitter and try again right now to remove any possible changes caused by temperature/time/neighbor loading, etc. Still, thoughts?

Disregard the PingPlotter stats I mentioned above.... I’m an idiot lol. I connected my laptop to the Ethernet cable connected to the TV. I left the WiFi Enabled, so those stats were via WiFi connection..

Here’s an update:
I noticed that my Correctables/unCorrectables were still reaching >1,000 / channel within 24-48 hours following CM1000 install.
My US Power levels were typically ~47dBmV/channel and were periodically maxing out and causing T3/T4’s just like before.

Adding a splitter to attenuate any potential noise at the drop did not change my SNR/MER levels much, they remained 39-43.
So I bought a female-female adapter that had terminals for proper grounding. So I connected the drop directly to my coax line upstairs where my modem is connected (just like before) but this time with proper grounding. - NO CHANGE -

I researched a bit and ended up buying a 100ft Quad Shielded Coax from Amazon so I could run this temporarily from the drop to my modem upstairs. If this worked well and problems went away, I would know that my coax in the walls is the Root Cause.

Here’s the cable I bought:
100ft Weather Seal Quad Shield Outdoor 3GHZ RG-6 Coaxial Cable 75 Ohm (Satellite TV or Broadband Internet) Anti Corrosion Brass Connector RG6 Fittings Assembled in USA by PHAT Satellite INTL
https://www.amazon.com/dp/B01GBX8HIG/ref=cm_sw_r_cp_api_i_5HENEbFNQ5AMD


Unfortunately, ~113 hours of UpTime, and each Channel has >2,000 correctables/unCorrectables just like before.
Adding the splitter in series with this new cable causes the US to increase to max and cause T3/T4’s as well.

So this means that the Root Cause must be somewhere on Spectrum’s end. Whether it be someone in the neighborhood introducing noise, an issue at the node or something else.
I feel better having ruled out issues on my end, and I don’t expect this to be resolved in a timely manner, as is no surprise I’m sure based on TimeWarner/Spectrum’s history with resolving these types of issues.

Just for fun/reference I will include a capture of my log today. I’ve been taking a screenshot of my log/cable connection every day, so I have the daily history too which may come in handy.

Let me know if you notice anything I missed!
Thanks.