Lower voltage on ~4 channels centered around 585 MHz and also very high number of uncorrectables

Hey, @goldfreaz! Thanks for reaching out to us on the forums. Having a look into the signal to the modem, everything looks to be within specification. As you have stated, you don't have any interruption in service and the up and down speeds are where they need to be. Can you expand more on what you want to know?

@XfinityMichaelC​  Thanks Michael for looking at this.  As you say speeds are fine so why worry.  The modem (SB8200) was purchased to replace and old die hard (TM822) so I want to make sure the new one is not defective before the return window closes.  The old modem downloaded  fine up to its limit, 300 Mbps with only 8 channels.  My concern is the drop in downstream power and S/N in the 500-600 MHz range which result in a high number of uncorrectables.  I am under the believe that uncorrectables reduce speed but I am not sure if all 32 channels need to work perfectly to achieve 480/12 Mbps.  Do these high number of uncorrectables affect me?

Something happened last night and I logged a bunch of events (below).  This morning, all the power levels are 4-6 dB and  S/Ns are 40-42 dB.  Hopefully the levels will stay together from now on.

Thank you for that information. I have the same SB8200 modem and I noticed that it can be recording events that may not actually be going on from time to time. I know when these modems first came out there were some issues with their speed and firmware, but those had been fixed. Checking into the signal again, no timeouts or uncorrectables are actually being reported. A high uncorrectable amount could mean there is noise or interference on the line and could impede the speed. Do you happen to notice any buffering when trying to stream movies or videos?

Michael,

Movies streaming from Amazon have been good for three weeks since we got the new modem.  The old modem worked well, also.  Hardly any buffering and no chronic buffering.  I checked the levels just now and they have dropped again in the same frequency range ~580 MHz).  If it make any difference, the router is an Asus RT-AX86U Pro with WAN aggregation enable (two internet cables to the modem).

No ground wire is on the coax but I assume the outside box is grounded.  The RG-6 coax is about 45 feet long with a splice barrel  and a 6 dB forward path attenuator about midway.  The downstream levels go up 6 dB (<11 dB MAX) without the attenuator but the S/N  and uncorrectables are about the same.  My previous subject  has a picture of the outside box.

Greg

Hey, @goldfreaz, thank you for providing that update. After the improvement to the area take place, please let us know if this issue remains for you. 

Marcus,

I did not notice a change in the rate of uncorrectables after the Tuesday network work.  I added a ground wire to the internal coax add the midway splice point this Friday.  I also moved the coax away from the 120 VAC to 12 VDC transformer at the modem.  I also removed the Cyberpower 1500 sine wave battery back as the power source.  The battery backup was making a high pitched noise.  I do not think the battery back up  is the noise source but I have not added it back yet.

I measured the resistance between the coax middle point and the internal house ground at a nearby power outlet.  It was about 2 ohms so the coax shield is probably grounded  at the outside box/circuit.  The voltage difference was 0 to 0.1 VDC.  The error rate did decrease noticeably (probably ~ 65% with these changes.  The data speeds up and down have been consistently good (always >460 Gbps).  The uncorrectable errors are still in the same frequency range.  The correctables are generally much higher than the uncorrectable now after the changes I made also.

I noticed a buzzing noise (fairly load) at a telephone pole about a half mile from my house.  It is fairly rural here.  The Comcast box on the cable near this pole is about 14"x5"x3".  I could not identify the source of the noise.  It is on my walking route so I will take a picture of the pole number and send it to the power/phone companies.  Let me know if you need the pole location.

Thanks for keeping an eye on this for me.

Greg

@XfinityMarcus​ After 24 plus hours, Uncorrectables have to increased considerably.  Voltage and S/N are adjusting up and down periodically.  Right now levels seem very good to me:

Okay, that does make sense. Let us know here right away if that does become a problem. 

@XfinityJohnG​ I did  a long duration ping test shown below.  The drop in the California server is probably not my modem or Comcast because the other two servers were working.   The channel with the lowest power and S/N is migrating around between 540 MHz and 590 MHz and sometimes the low levels  disappear.  Do you think a new internal coax would help this problem?  I did not see a change with two different internal coax combinations with and without a 6 dB downstream attenuation. Both coax combinations were good RG6 coaxes with quad shielding, but both combinations had a splice (union) about half way over the ~40 foot length (splice and actuator at midpoint shown below). 

I do not think I have a serious problem but I continue with these posts just to document my experience.

<--- to outside   ----  to modem  --->

@XfinityJohnG​ Power spectrum does show the nulls between channels and the dip at 567 MHz.  I cannot control the FFT parameters very well  on the modem and I do not know what to look for...

Transmit spectrum looks okay to me, but I am not an expert.  Four 6.4 MHz channels, two 3.2 MHz channels, and one 5 MHz harmonic 20 dB down.  The highest frequency needs ~ 3 dB more power than the lowest.  The receive spectrum has improved and is nearly flat 60% of the time.  Because this problem appears to me to be intermittent, I have delayed Xfinity from making a house visit.

@jasongil​ Thanks for your input.  I think Comcast knows what is malfunctioning but they only fix their worst problems first and they let minor problems slide to the right.

Thank you for reaching out to us on our community forums @jasongil! Have you noticed these inconsistent speed issues on multiple devices at the same time? Is it impacting wireless connections, hardwired connections, or both?

Hi @XfinityAldrik​, thanks for responding.

I measured the speed only on hardwired devices and was planning on contacting Xfinity support, but the very next day I was notified that maintenance was occurring in my area. Following that maintenance, the problem has been resolved - all channels are within 1 dBmV of each other and there no correctable and uncorrectable errors.

@jasongil thank you for that update. Is there anything additional I may assist you with at this time?

Thanks for providing an update.

Since my previous comment, there was additional maintenance in my area which resolved the wide variances in power across the channels and excessive number of correctable / uncorrectables, though I still get some in the 500 MHz range.

This is 17 days of runtime:

@goldfreaz - As I am looking at my OFDM PLC Channel, I do see the uncorrected errors, that is normal in my experience.

I have read that 700 Mhz corresponds to cell tower frequencies.  You may have weak shielding or grounding issues.

@goldfreaz - you might be onto something, as the 2-way splitter inside the house right before the CM does not have the grounding.

@goldfreaz - I grounded the splitter, we'll see if it makes any difference... ;-)

@EG - there is already a grounding point at the demarc, as the coax comes into the house and where the return filter is installed, so you are saying that the splitter does not need to be grounded?

@EG​ - it looks to be good one (Intersystem Bonding Termination), it goes inside the house together with the main electrical feed, assuming into the breaker panel.