Modem loses connection to Rogers network every month

[toolcubed]

Hi everyone,

Even though I'm on Ignite, I'm posting this here because my issue is NOT related to the Ignite platform, and I've had this problem even when I was on legacy internet with the CODA-4582 modem.  I have reported this problem to Rogers three times - once in June when they opened a "monitoring" ticket for 7 days and didn't find anything wrong; again in July when they opened another monitoring ticket for 7 days, and again, didn't find anything wrong; and again today (because it just happened again), but this time they escalated a ticket for further investigation since this issue occurs monthly like clock work.

The problem that occurs monthly is that the modem will completely lose its connection to the network for about 10 mins.  Once it re-establishes the connection, everything is fine for another ~month until the exact same thing happens again.  It's never on the same day of the month but fairly close.  For example, in June, it occurred on the 22nd.  In July, it occurred on the 19th.  And it occurred again today (20th of August).  And again, the internet is down for only about 10 mins each time it happens.  I also had this issue in May, April, March and likely earlier although I can't remember that far back.  I did not start reporting this to Rogers until June.  When the issue occurs, my Ignite modem's light will change from solid white to flashing amber for 10 mins, then it will flash green for a few seconds before turning solid white again.  On the CODA-4582 modem, the light sequence/pattern was the same as if I had power-cycled the modem except that the downstream/upstream lights would continue to flash for 10 mins before turning solid and locking onto the network.

I don't have any problems with speed or latency.  My signal power levels and SNR values are all VERY good.  And all three Rogers techs that I spoke with (each time I reported the problem) all confirmed that there was absolutely zero noise on my line and in my area.  They also confirmed that other homes in my area have the same problem (i.e. a bunch of modems in my area will lose connection at exactly the same time as mine) but they were unable to tell me if it's localized to just the homes connected to my CMTS or if it's more widespread.

Honestly, this smells like a CMTS problem, especially since I still get IPv6 packet loss (but not IPv4), which could also point to a CMTS problem.  I'm not saying that the two issues are related for sure, but I don't think it's just coincidence.

I'm remaining optimistic that Rogers will find root cause and fix this quickly but I'm posting here to see if there are any other members of this community who experienced the same issue, and who can potentially shed some light on what they think the problem is?

Thanks!

[-G-]

---

@toolcubed wrote:

No worries @-G-   I might do that or I might just reinstall the grounding block myself...as I've mentioned in an earlier post, I've done this stuff before 🙂

---

Again, your risk exposure is pretty low so I would leave it to Rogers to fix.  Do you have a high-performance/3GHz-rated grounding block installed, with blue-coloured insulator on the F connector?  If you have an older block, with a white centre, I would leave it disconnected.  I would contact the @CommunityHelps  team for guidance.

[Argee]

Just as an aside and for info purposes, I'm forwarding a link to the SCTE.org site where they have a slew of info on CATV specs, etc. This particular page is on grounding and while many of the webinars have finished you may still be able to glean some helpful info. One page in particular from PPC appears to still be active.

https://www.scte.org/search/?page=1&query=grounding

[toolcubed]

Thank you both.

@-G- yes, the grounding block has the blue insulation inside the coax coupler.  I also have a brand new one in my toolbox if I should decide to replace it myself, which also has the blue insulation and which I know for a fact is rated for 3GHz.  I'll contact community helps for guidance.  Thanks again.

[toolcubed]

One more technical question regarding the upstream power level boost that the senior tech adjusted at the node yesterday. Does anyone know exactly how that would’ve been done?

My understanding is that a lower upstream power level is actually a good thing because it means downstream signal from the cable modem is very good. And only if downstream power levels drop should the upstream levels increase. I didn’t think that 34db on the upstream (what my modem registered before the tech made the change) was considered “too low” or a bad thing as the tech implied.

After the tech increased the upstream level at the node (from 34 to 38 according to him), my downstream levels didn’t change at all and my SNR increased slightly. This seems rather odd because an increase in upstream power (especially an increase of 4db) should’ve dropped my downstream levels, but instead, those levels remained the same. I’d like to understand how this was achieved. Also, despite the tech saying that he increased the upstream at the node to 38db, my modem registers 39.8db on one upstream channel and 41db on the other three channels.

[Datalink]

@toolcubed:

1.  Yes your cable ground block should be grounded, preferably to a suitable electrical ground point.  Did Bell install fibre in your neighourhood any time in the recent past?  If so, I'd bet that the contracted locators who were responsible for locating the underground services disconnected the cable grounds in order to locate the Rogers cables.  And if thats the case, I'd also bet that the same crew didn't come back to connect the ground cables.  That's what happened in our neighbourhood last summer, and, as far as I know, all of the neighbourhood cable grounds are still disconnected.  

2.  As for the signal levels, if there is an amplifier between your local tap and the neighbourhood node, perhaps at the neighbourhood node itself, then the tech would simply adjust the outbound and inbound signal levels.  If the inbound signal levels at the node were adjusted downwards, and the CMTS wasn't happy with the result, it would command all of the connected modems on your line to run at a higher level.  The reverse would happen if the tech adjusted the amp inbound levels higher.  In that case the CMTS would command the modems to run at a lower output signal level.  

[-G-]

---

@toolcubed wrote:
One more technical question regarding the upstream power level boost that the senior tech adjusted at the node yesterday. Does anyone know exactly how that would’ve been done?

I’d like to understand how this was achieved. Also, despite the tech saying that he increased the upstream at the node to 38db, my modem registers 39.8db on one upstream channel and 41db on the other three channels.

---

How the adjustment is done depends on the node.  (You basically adjust a resistive attenuator pad either physically or electronically.)  However, the tech absolutely did the right thing in bringing your modem's upstream power levels to approximately 40 dBmV.  That's ideal.  34 is too low; 50 is too high.

[toolcubed]

Thanks Datalink.

There are no amps around here. I confirmed that with Rogers a couple of years ago when they did the network uplift and removed all amps.

When you say “inbound” signal level at the node, I’m assuming you mean upstream? If that’s the case, then the tech told me that he increased those levels at the node (i.e. adjusted upstream higher - from 34db to 38db). But that resulted in my modem’s upstream levels increasing, not decreasing. After the tech adjusted the level at the node, my modem went from ~34db to ~41db. My downlink levels remained the same but my SNR levels increased from between 38 and 40 to between 40 to 41.

[toolcubed]

And yes, we did recently have Bell fibre installed. They wrapped up construction in May. When would they have disconnected the grounding block though? I know for a fact that the block was connected in early June because I watched the Ignite tech check my signal and reconnect the coax to the block. At that time, Bell was already finished their FTTH construction.

Edit:

Just saw your other response. Thanks for the details and for confirming that the increased upstream levels registered at my modem are now ideal.

[Datalink]

I wonder if there's a misunderstanding between what the tech said and how you interpret his remarks.  Perhaps he meant that the modem upstream levels would increase, which would be caused by cranking down any inbound gain value at the node? 

If you saw the tech connect the cable to the cable ground at some point in the recent past, and now its completely bypassed, I'd say that someone has done that deliberately.  It doesn't make any sense from my perspective, but, from what you're indicating, that is what happened.  I wonder if that has also occurred with your neighbours?   If thats the case, then the question should be posed to the area manager.

When I indicated that the cable grounds were disconnected, the RG-6 cables were left connected to the ground block, the ground cable was left connected to the ground block, but the the other end of the ground cable was disconnected from the clamp that was attached to the house electrical meter.    Seems that years ago, that was the accepted practice, to use a clamp to connect to the electrical meter box.  

[toolcubed]

Ah. Yes, there’s certainly a possibility that I misinterpreted what the tech said. Your explanation makes much more sense. I already reconnected the coax cables to the grounding block. In my case, it was different from the way yours was disconnected. My grounding block was still physically present inside the grey box (with ground wire attached), but there weren’t any coax cables connected to it. The cables were joined together with a standard coax coupler.

My ground wire runs from the grounding block into my basement where it’s clamped to a bare copper wire, which runs into my electrical panel and then clamped to my main water shutoff.

[toolcubed]

Quick update.  The issue has persisted and occurred again this month, just as I knew it would.  This time, I called the area network technician directly because the Rogers escalation team gave me his direct contact details (rather than having me call into tech support and go thru the same process again).

The network tech from the maintenance crew showed up this morning.  Upon checking and running some tests on one of the nodes, he suspects a faulty power supply.  When he disconnected the main power, instead of the node transparently cutting over to backup battery power, it went down completely and took down service for the entire area being served by that node.  He also confirmed that each month, these nodes go into automatic self-tests, and unfortunately, this particular node fails that self test each and every time because it goes down completely for 10 mins instead of cutting over to the batteries...and voila...it causes my monthly 10 min disconnects 🙂

He said he's going to replace the power supply after midnight tonight but also wants to run a few more diagnostics to make sure there's nothing else going wrong with this node.

FYI #1 - this is a new node with all new equipment (new power supply, new batteries, etc), which was installed three years ago as part of the "network uplift" in my area.

FYI #2 - the power supply being used in this node is an XM3-HP model (link below) for those of you interested.  The technician is also going to make note of this in case there's a vendor defect with this particular model.

XM3-HP Series (alpha.com)

Fingers crossed that once he replaces the unit, these monthly disconnects will stop occurring.

[toolcubed]

Update:

The UPS did not need to be replaced.  It wasn't defective.  The area network engineer called me back this morning.  After he did more research on the problem, he found that the power output wasn't configured properly.  Once he corrected the power output level, he performed another two tests - 1.) he initiated the self-test and it finished successfully with no drop in service; and 2.) he manually disconnected the power, and it successfully cut-over to battery backup with no drop in service.  So I think my monthly disconnect issue has finally been resolved.  Obviously I'll keep an eye on it going forward.

Another thing that the technician mentioned is that there appears to be similar cases reported in other areas (different neighbourhoods), so he suspects that power levels might also not be configured correctly in those other areas.  I'm wondering if other Rogers customers who are reporting similar outages are being impacted by the same issue with a misconfiguration of their local node's UPS? 🤔

Tagging @RogersIan .  This should be circulated internally so that Rogers network engineers can proactively check on this across the cable footprint.