cancel
Showing results for 
Search instead for 
Did you mean: 

Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

I have been having issues with my signal levels since signing for Rogers a month ago. Tech support seemed to agree with me (most of the time) that I shouldn't be having such low signals and a hundred thousand uncorrectables, so they sent techs out. Unfortunately, it seems like they haven't been able to fix the issue, despite replacing things and then declaring to me that it's fixed. 

 

I'm posting here now as I feel I'm slowly being driven insane by contacting first-line tech support - it seems to be random when they can see the issue or not. The agent I spoke with yesterday on live chat took the unprecedented step of blaming me and my equipment for everything (which had never happened before in my 10+ calls to tech support), then said I wasn't responding and terminated the chat due to "inactivity" (despite the chat transcript that was sent to my email later clearly showing all my messages went through down to the minute that the chat was terminated) 

 

One of the things the last tech to visit suggested was to periodically monitor my modem stats. So, I modified a script that some Comcast users have written to periodically scrape the modem stats and generate graphs to work with my XB7. It seems to confirm pretty clearly that there is still an issue, as SNR fluctuates over a wide range of 35-42dB (with some channels sometimes going down to as low as 30dB)

 

2022-08-19_18-52.png

 

How stats look for individual channels:
2022-08-19_18-54.png

 

Downstream power levels (low-power channels are ch 1 and 33):

 

2022-08-19_18-55.png

 

Upstream power-levels (5th OFDMA channel excluded as it just shows a value of 435000.0 dBmV):

 

2022-08-19_18-57.png

 

And my most recent stats copied straight from modem:
 
IndexLock StatusFrequencySNRPower LevelModulation
Downstream
Channel Bonding Value
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
24
25
26
27
28
29
30
31
32
33
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
Locked
693 MHz
303 MHz
849 MHz
855 MHz
861 MHz
579 MHz
585 MHz
591 MHz
597 MHz
603 MHz
609 MHz
615 MHz
621 MHz
633 MHz
639 MHz
645 MHz
651 MHz
657 MHz
663 MHz
669 MHz
675 MHz
681 MHz
687 MHz
699 MHz
705 MHz
711 MHz
717 MHz
723 MHz
825 MHz
831 MHz
837 MHz
843 MHz
350000000
39.8 dB
38.2 dB
37.8 dB
36.8 dB
37.8 dB
38.2 dB
39.1 dB
38.6 dB
38.2 dB
38.7 dB
39.2 dB
40.0 dB
38.7 dB
38.3 dB
40.5 dB
39.0 dB
38.4 dB
39.7 dB
39.5 dB
39.9 dB
39.0 dB
38.5 dB
39.5 dB
41.1 dB
38.9 dB
41.1 dB
41.0 dB
40.2 dB
38.7 dB
39.2 dB
37.9 dB
39.7 dB
40.5 dB
5.9 dBmV
1.6 dBmV
6.2 dBmV
5.8 dBmV
5.9 dBmV
4.7 dBmV
4.8 dBmV
4.9 dBmV
5.0 dBmV
5.0 dBmV
5.5 dBmV
5.5 dBmV
5.0 dBmV
5.3 dBmV
5.3 dBmV
5.3 dBmV
5.1 dBmV
5.2 dBmV
5.3 dBmV
5.4 dBmV
5.8 dBmV
5.9 dBmV
5.9 dBmV
6.1 dBmV
6.0 dBmV
6.3 dBmV
6.3 dBmV
5.8 dBmV
6.9 dBmV
7.2 dBmV
6.9 dBmV
6.6 dBmV
2.3 dBmV
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
256 QAM
OFDM
IndexLock StatusFrequencySymbol RatePower LevelModulationChannel Type
Upstream
Channel Bonding Value
1
2
3
4
5
Locked
Locked
Locked
Locked
Locked
21 MHz
25 MHz
32 MHz
38 MHz
4 MHz
2560
5120
5120
5120
0
45.5 dBmV
47.0 dBmV
47.0 dBmV
47.3 dBmV
435000.0 dBmV
QAM
QAM
QAM
QAM
OFDMA
TDMA_AND_ATDMA
ATDMA
ATDMA
ATDMA
TDMA
IndexUnerrored CodewordsCorrectable CodewordsUncorrectable Codewords
CM Error Codewords
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
2941127777
4259313928
4259330913
4259345703
4259355167
4259359940
4259373389
4259371776
4259379917
4259387137
4259391486
4259403402
4259399590
4259398071
4259410045
4259418570
4259424292
4259434361
4259443450
4259453509
4259455292
4259464761
4259463457
4252880535
4254225566
4259479562
4259492585
4259497763
4257374221
4253613213
4259029911
4259520345
2941127777
1719920946
0
0
0
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
75624
215196
2161
1189
0
24091
36554
11570
0
1719920946
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
215253
549942
3093
4
0
69674
50069
23430
0
0
 
 

 

Any ideas for what I should do and whether this should be fixed or not? The first (contractor) tech that came out for this issue blamed the drop cable (which was 30 years old), so he replaced it with a temp line and claimed my issues would be solved once a new buried line is put in. Signals did improve slightly, but didn't appear to be fully fixed (still got thousands of uncorrectables, even after line was buried). Second (contractor) tech blamed high downstream power and some F connectors; he added an attenuator at the demarc box and replaced the connectors, but clearly that hasn't seem to have fixed the issue either. In fact, I'm getting more uncorrectables now than before he replaced the connectors.

 

I make an issue about this because I've been noticing some packet loss spikes of 1-2% every few hours on my OpnSense router that monitors my connection (contrast this with my Bell backup connection, which it monitors to always have 0% packet loss), which I'm not sure if it's related to the signals/uncorrectables. Though, I've been told signals/uncorrectables is a problem as well independent of whether I get packet loss or not (as the signals/uncorrectables may cause intermittent problems in the future)

 

Should I escalate to have a senior tech (employed by Rogers and not contractors) be sent out and take a look?

 

The contractors don't seem to able to figure this one out (in fact, most of these issues should have been caught by the first contractor who came out to do the install. He didn't seem to bother to do any signal tests of the lines, as he should have noticed my drop and household lines were quite old and in need of replacement. He even redirected a line clearly labelled RG59 through my vents to connect to the modem - and also managed to tangle the RG59 with some power cables, which caused issues until I had that replaced with new RG6)

18 REPLIES 18

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

MitchellS
I'm here a lot

I suppose a first step is to determine whether these fluctuating signal levels are normal, and more than that, what the actual impact on your download/upload speeds or Internet connectivity are. For example, you could get someone else on Rogers (and who has good Internet access) to run those scripts. Or you could run PingPlotter or periodically do a speed test to show what the actual variation in your Internet access is (and if it corresponds to the low signal levels shown by the scripts).

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

It's definitely not normal - even without fancy scripts, if I just sit and refresh the modem page every few seconds, I see SNR fluctuating between 35db and 39db for quite a few channels (even within a span of a minute). I've heard from others that they mostly see constant 42-43dB (in the best case) or constant 38-39dB (worst case that I've heard), no matter how many times they refresh the page. The forum help post (https://communityforums.rogers.com/t5/Internet/Troubleshooting-a-Slow-or-Intermittent-Connection-Wir...) says that Rogers expects to see an SNR above 38dB (there's a unit typo in the link) which is clearly not what I'm seeing most of the time. (The stats I posted before was one of the better ones; if I refreshed 10 seconds later, most of it would go down)

 

The point of graphing it was just so I didn't need to sit around refreshing the modem page (and having to mentally check against what numbers I saw last time). The script I used was a modification of this GitHub repo with code from another repo supporting the XB7. The original version of the script here came with screenshots (of which I presume is their Comcast connection), which shows mostly constant SNR over the displayed 4 hour window:

 

68747470733a2f2f692e696d6775722e636f6d2f70705a6a6e6b502e706e67.png

 

Since I've heard Comcast infrastructure is generally worse (in terms of maintenance and upgrades) compared to Rogers, I expect Rogers to do just as well if not better. The above is a far cry from what I see over a 4 hour window:

 

2022-08-20_21-51.png

 

I can try running PingPlotter to the CMTS later; the 1-2% packet loss spikes that I mentioned is monitoring Google's DNS servers by pinging it every second (my Bell backup connection is monitored the same way - though, it's of course not a true comparison since both routes to Google are different. The graph looks like this (for IPv4): 

2022-08-20_21-58.png

 

While the packet loss spikes seem to somewhat correlate with the fluctuating signals (though it could also be ICMP being dropped somewhere in the Rogers network - I do need to test CMTS later, but I've heard those deprioritize  ICMP requests), for whatever reason, there doesn't always seem to be packet loss when the signal fluctuates to low values. But it does seem like whenever there's packet loss, there's also low SNR values. And zooming in to observe just the latency, I can somewhat confirm that the latency spikes seem to correlate with SNR fluctuation to low values (I checked a couple timestamps manually) - though, I would need to export .csv from both the script and the router to be sure. Then again, the stddev numbers fluctuate around ~1-3ms, which I presume is normal for cable technology.

 

For contrast, this is what my Bell DSL connection looks like (zero loss; the gap is due to me restarting my router):

 

 2022-08-20_22-22.png

 

What is certainly not normal is getting 200,000-500,000 uncorrectables on multiple channels within 24 hours of doing a restart of the modem. Every technician or tech support agent at least seems to agree that I should be getting zero, or very few (<1000) within a few hours, even if they don't agree on whether I have signal variation or not. (Again, for whatever reason, every person I talk to about this seems to either agree or disagree with me on a random basis)

 

Periodic speedtests (say every 15 mins) I'm bit hesitant to do  as those do cause severe latency spikes (40ms+) while running. When I do run them manually, I see some fluctuation around 900-1170mbps (wired, to the same speedtest server), which doesn't really indicate whether it's a neighbourhood load problem or the fluctuating signals. (Though, I've heard from techs that there aren't many Rogers customers in my neighbourhood. This is a residential household, not an apartment or condo unit. The tap and node also seem to be in the Rogers box right on my lawn)

 

I mean, this isn't the most pressing issue for Rogers to fix, which is maybe why they only send contractors out (and some support agents being clueless as to what I'm talking about). But I do wish that after 2 tech visits at least something was improved significantly, not just marginally.

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

Somehow, the first two images aren't showing up when I refresh the page. Here they are uploaded elsewhere:

https://camo.githubusercontent.com/dddfdcce7c1fa7a400dfc65666deeb2b10ab9784346b7be8c80104a5b240f8fd/...

 

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

Any of the Resident Experts have thoughts about this? (@-G-@Datalink?) I will be running a 24h PingPlotter test soon, but I think it should be easy to tell whether this is normal or not (i.e. seeing SNR fluctuating between 35-39dB simply by refreshing the modem stats page + 200,000/500,000 uncorrectables on two channels, with thousands more on other channels within 24 hours).

 

I have also messaged @CommunityHelps to see if they're able to assist me further than the phone/live-chat agents, but so far haven't received a response (actually, I'm not sure how the PM system here works; you can't seem to see if the message was sent or not?)

 

I did see some 20% packet loss at the CMTS in a brief ping test recently, which may have been false packet loss, but I believe there was still 5-6% loss reported at the destination - so, some of it may be real.

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

57
Resident Expert
Resident Expert

@t_kwan wrote:

 

...I'm not sure how the PM system here works; you can't seem to see if the message was sent or not?)


At the top right of a browser window is your avatar/username.  To the left of it is an envelope. Click it and you can see your messages (inbox and sent in the dropdown).

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

@t_kwan fwiw, if you want to run a ping test to the CMTS, run that using the Windows ping command for a first run.  The moderators @CommunityHelps will only look at data from a normal ping test.  That's unfortunate as it doesn't include time stamps and the upper data will scroll up over the limit of the popup that contains the command prompt.  However, as a gauge to determine the overall packet loss, it can be a useful test.  

 

Run a ping test via ethernet to the CMTS: 

1.  Run a trace to anywhere, google for example:   tracert www.google.com

2.  If the modem is running in its default Gateway mode, the first IP in the trace should be the modem, the second IP will be the Cable Modem Termination System (CMTS).  The neighbourhood node, which is the real target of test does not show up in the trace. 

3.  Ping the CMTS IP address.  Run a long test.  Minimum test time is an hour, I recommend a 24 hour test.

ping -n 86400 xxx.xxx.xxx     where xxx.xxx.xxx is the IP address of the CMTS from hop #2 in the trace.

 

That test will run an automatically terminate at the 24 hour point.  Post the bottom rest results which will show the packet loss to the CMTS over the entire day.  The path to the CMTS is as follows:

 

1.  PC to modem via ethernet

2.  Modem to external demarcation point (Network Interface Device) via RG6 cable;

3.  NID to Local Tap via RG6

4.  Local Tap to Neighbourhood node via Hard Cable

5.  Neighbourhood node to CMTS via Fibre

 

Anywhere along the path, from items 1 to 4 could cause packet loss.  The usual culprit is the external cable that runs from the external NID to the Local Tap.  That cable is the Achilles heal of cable systems.  It doesn't last forever and requires periodic replacement.  But, there could be issues at the Local Tap or beyond, up to and including the neighbouhood node.  A problem such as the one that you're seeing requires patience to push Rogers techs to address the situation.  The starting point is to quantify the packet loss that you see from the ping test to the CMTS. 

 

Fwiw, Pingplotter is useful for one purpose, hunting for packet loss.  The lost packets will show up in red, on the bottom display area, so, its easy to see when that packet loss occurs.  

 

One problem with Pingplotter is that it will display multiple data points on the bottom plot when you either decrease the ping interval down to or below 1 second, or you switch to a high time frame display, such as 24 hours, 48, 72 hours.  There are combinations of low pint intervals and selected higher timeframes, less than 24 hours where a single plotted data point will contain more than one captured data point.  If you hover your mouse over the lower display, it should show the data point for that particular horizontal pixel.  As you either decrease the ping interval or increase the time covered by the display, at some point that horizontal pixel will use more than one data point and it will do so without warning.  You will see that on the popup display when your mouse is sitting in the lower display area. The data points will be displayed.  

 

So, the greater impact of multiple data points per individual horizontal pixel is that Pingplotter averages the data points instead of using the higher or lower time values.  So, the plot will flatten out and you will lose any high time points from the lower display.  The upper data display, if the setting is on auto, will show the correct high and low values for the lower plot.  So, if there are high time responses, you will know that they are somewhere in the plot as you can see the times in the upper data display.  You won't know where those points are unless you scale down in terms of the plot time and scroll thru the data, looking for those high time values.  

 

I haven't looked at this recently as I use Wireshark to capture and display the data.  Wireshark doesn't average the data unless you plot the average times.  Pingplotter probably hasn't changed this issue, but, I don't have a recent up to data version loaded.  

 

Fwiw, I wouldn't expect to see large variations in the signal levels or signal to noise ratios after a modem reboot.  If there are large variations, that tells me that there's an issue with the external cable, or its connectors or possibly an issue with the local tap that connects you and your immediate neighbours to the hard cable that runs from the neighbouhood node.  Here's what a typical tap looks like.  Essentially its an overgrown splitter with specs to match the RG6 and Hard Cable impedance:

 

http://www.acmetronix.com/at/index.php?m=content&c=index&a=show&catid=15&id=53

 

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

-G-
Resident Expert
Resident Expert

@t_kwan  Just a few additional thoughts/comments:

 

That you are seeing a high number of uncorrectable codewords errors in your stats is indicative of problems, especially on your downstream D3.1 (OFDM) channel since that's the one that is really carrying your traffic.

 

You should not be seeing much variation in the SNR, especially over a short period of time.  (Over the course of 15 minutes, I saw my SNR stats, across all channels, vary up and down by only 0.1 dB, and a max of 0.4 dB on one channel.  I also got only 3 uncorrectable codewords errors on my D3.1 channel over the last 12 hours.)

 

When the SNR fluctuates, since this literally is the signal-to-noise ratio, one or both of those components is fluctuating, and I'd bet that noise is the culprit.  When you call into Rogers, I don't think that the first-level tech vas visibility into these stats in your local node but a higher-level tech will, and can also check to see whether there are any "neighbourhood issues" that could be impacting you.

 

Within your home, what does the coax path look like from the time it enters your home to where it reaches your modem/gateway?  Does the incoming Rogers cable feed connect directly to the coax connecter on the XB7 or do you use an F81 barrel connector to connect a coax cable that leads to a wall plate upstairs, and then to another coax patch cable that connects to your modem?  If it's the latter, do you have any other wiring that runs on the same path as your coax cables or that shares the same wall plate?  Each coax coupler on the path (including the one in the wall plate) is a point where you could get noise ingress.

 

Can you also confirm whether or not the coax coming into your home has been grounded?

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

@t_kwan I looked at your first post again.  I see that you're running a router and that you have a signal attenuator installed.  

 

I'm assuming that your modem is running in Bridge mode.  In that case, the first IP address in the trace will be the router, the next IP address will be the CMTS.  Ping the CMTS.  

 

It would be odd to run a router in full router mode with the modem running in Gateway mode, however, it seems that more Ignite systems users are going that route due to the idiosyncracies of the XB6 or XB7.  If thats the case, the first IP address is the router, the second IP address is the modem, the third IP address is the CMTS.  Ping the CMTS. 

 

For the attenuator, since the introduction of the Ignite system, I've noticed more users with high downstream signal levels.  Don't know why that is, but, in some cases its a real problem.  Installing an attenuator will drop the downstream levels, but, it will also push up the upstream levels.  There is a real balancing act that occurs when you try to drop the downstream levels without pushing up the upstream levels to the point of failure, where the modem drops upstream channels due to the upstream power limitation of the modem.  The DOCSIS spec calls for 51 dBmV max for three or four channel upstream ops.  Rogers uses 52 dBmV for some reason.  By then I'd expect to see the modem drop one or two upstream channels.  There is a power extension up to 56 or 57 dBmV within the DOCSIS spec, but, I don't know if Rogers actually uses it.  

 

I suspect that Rogers still uses a normal attenuator for situations like this, which affects the signal levels in both directions.  There are Forward path attenuators which will drop the signal level for the downstream direction only.  Here's an example:

 

https://www.amazon.ca/line-Signal-Forward-Attenuator-FPA6-54/dp/B07882H96R

 

There are other signal attenuator levels, usually, 3 or 6 or 9 dB attenuators.  

 

I agree with @-G- .  Have a look at the signal path from the modem out to the external demarcation point.  There shouldn't be any splitter in that path.  It should be a single path, where the internal house modem cable and inbound external cable are connected via F-81 connector. 

 

You would have to be able to open the external Network Interface Device to check the ground situation.  That  requires a tool to open the NID.  You can however ensure that there is a ground cable exiting the NID and that the cable is connected to the external power box which contains the electrical meter for your house.  That ground cable is usually clamped to the meter box, which is then grounded to a house ground or to the inbound water pipe which serves as a ground point. 

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

Glad to hear that it's not just me thinking that the variation is abnormal. (In a 15-min window, my min variation was at least 0.5dB, and a max of 3-4 dB over most channels. And it still follows a zig-zag pattern, with me scraping the stats every 20 seconds)

 

Do you have any ideas on how I can reach a higher-level tech? I keep getting sent away by first-level support by 'Your signals look fine' or that they'll send out a (contractor) technician (who still have not been able to fix this after 3 visits: 1 installer + 2 techs to solve this issue. Signal tests have been done on my inside lines, which they claim are fine. Drop cable was replaced and has already been reburied. Though, I noticed the SNR variation is mostly the same - whether it was the old line, temp line, or new buried line - but I am getting increased power as much as 10dBmV and fewer uncorrectables than the millions I getting before (within 24 hours) with the new line)

 

The coax path looks like this: XB7 -> brand-new RG6Q (quad shielded) through gap between vents and floorboard -> F81 barrel connector in unfinished basement -> RG6 (at least 20+ years-old, seems to be dual shielded Belden) from basement through the garage into the demarc box -> ground coupler (copper ground attaches to a clip that is clamped to the hydrometer) -> 6dB normal attenuator installed by Rogers in last visit -> buried drop RG6 -> tap in Rogers green box (on my lawn).

 

Some comments on the route:

  • The neighbourhood node seems to be in the same green box as well, as I believe I saw a thin, coloured (probably fiber) cable going into a metal box (probably node) within the same box. If I'm correct, then the run from the node to the tap should be quite short i.e. narrows down the issue to either inside the pedestal itself or the lines inside my house. I assume the new drop cable hasn't been damaged somehow so quickly. (And, for what it's worth, the inside of the pedestal box seems to have an overgrowth of weeds as tall as the box itself and cobwebs covering most of the equipment. Not sure if this is normal for outdoor Rogers green boxes, or an indication of a maintenance problem)
  • There are no wall-plates in the room where I keep my modem, hence the vent/floorboard trick (that the installer did to run the RG59, which was replaced). The gap seems to be large enough and there is no noticeable pressure on the cable if I move it in/out of the gap. The vent cover is not squishing the cable either.
  • The garage-basement cable does enter the basement right next to the central breaker box, with the barrel connector being placed right below the breaker box - it's not attached to it, just floating in the air and being held up by the two cables. I was worried that this may have been the issue before, but after looking at the graphs, I'm a bit skeptical this is the problem. My signals actually improve during the day, which is when there is the highest amount of power usage in my home - if you notice, there's an 1dB arch in the minimum SNR being reported during the day. At night is when I have the least power usage: A/C is off, no heaters or appliances running, etc.)
  • Just to confirm, no amps or splitters anywhere along the path (between my modem and the tap).
  • Coax coming in has been grounded (as mentioned in detailed path above)

(EDIT: Forgot to mention, the connectors on both ends of the garage-basement cable have been replaced; they are no longer the 20+ year old connectors from cable TV days)

 

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

First, thank you for all your advice in both replies - they are quite helpful.

 

The modem is indeed bridged, so I am pinging the second address from the trace. (I've read some previous posts of yours to figure out which one was the CMTS). Though, I may switch it back, attach a computer to it directly and leave it to run ping tests - just to avoid the hassle of being accused that it's my own equipment that is at fault (which I'm 99.9% sure is not the case, as it also monitors Bell (without problems) and I have already tried multiple checks/settings-swap to ensure it's not the router. And besides, the stats (aside from ping) that indicate the problem come from the modem itself)

 

For me, the increase of downstream signal occurs when I replaced the drop cable (which was quite old), with a +6-10dBmV increase across the channels. (Channels 2-32 used to have power levels ranging from +13-15dBmV, with +-7-8dBmV on Channels 1 and 33)

 

As to the FPA (forward path attenuator), I actually did look at that on Amazon and considered getting one, but I was concerned it would cause problems as it attenuates from 54 onwards (which I believe would catch the 54-85 MHz upstream for mid-split?). I did end up buying a regular 6dB attenuator from Amazon (Holland Electronics) - didn't notice much of a difference after installing it at the back of the modem so I ended up taking it off.

 

I was hoping the tech who came to fix it professionally would solve the problem earlier than the drop, but all he did was install a 6dB attenuator at the demarc box (or NID as you call it?). It does seem much higher quality than the Amazon one I got, but honestly haven't noticed much of a difference with/without it (just like the Amazon one). It is a normal attenuator, as my upstream power levels were previously at 39-41dBmV.

 

He did install new weather boots on all the connection points in the demarc box - i.e. boots on all ends of the ground coupler and attenuator. I was concerned about humidity within the box previously, so I guess that can be ruled out now).

 

See my reply to @-G- for more details about the path from the modem to the tap.

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

-G-
Resident Expert
Resident Expert

@t_kwan  All the graphs and troubleshooting that you are doing is great... but I think that you need to change focus a bit with regard to how you get Rogers engaged to fix your signal issues.

 

The graphs, that you posted here, highlight that your SNR is dancing around and across a wide margin... definitely NOT normal... but as long as the signal levels are within Rogers' acceptable norms, they won't escalate tickets to get this fixed.  You need Rogers to see that there is a problem that they think needs to be fixed, and it will help to have all of your equipment is running in a supported, as-installed configuration.  Keep things simple, be patient, and let Rogers follow their processes.

 

First, disable Bridge Mode on your Ignite gateway.   Remove your own router, if possible, and ensure that everything is in an as-installed configuration.  Next, power-cycle the gateway to reboot it and reset the hardware, and then just leave it running.  DO NOT reboot it.  Let the Uncorrectable Codewords errors accumulate.

 

As I said before, if your signal levels are within acceptable norms, the Rogers techs cannot escalate your support ticket.  Instead, focus on your error stats, especially on the downstream DOCSIS 3.1 OFDM channel.  It's totally normal to have a huge number of correctable codewords on the D3.1 channel but the uncorrectables should be zero, or closer to zero.  If you are getting a ton of uncorrectable codeword errors on the D3.1 channel, and on the D3.0 channels as well, that's a problem.  That is packet loss.  You can then point to the dancing SNR values as further evidence that you have might have a serious noise issue that needs to be addressed.... and all that combined will be enough to get a support ticket escalated and a qualified tech (and possibly a maintenance) crew to get dispatched.

 

If you don't feel up to talking with Tech Support on the phone all over again, try getting the @CommunityHelps  team engaged. Keep Rogers focused on investigating the health of the cable plant in your area -- noise, stats on your local node/CMTS, and checking to see whether any other customers in your immediate area are also seeing a high number of errors.

 

Regarding other testing, it's good to perform speed tests using a computer connected to the Ignite gateway via Ethernet.  You need to confirm whether or not you are getting your advertised speeds.  If you are, then that does not necessarily mean that you do not have a problem that needs to be fixed.  If you are not getting your advertised speeds, then this is something else that Rogers needs to investigate.

 

You don't need to run ping tests to confirm packet loss if you are seeing a high number of uncorrectable codewords errors in your stats.

 

Best of luck with getting your connection stabilized!

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

@-G- So, I followed your advice and managed to reach the second-level phone support. (Unfortunately, in the process, also had to endure one call with a very unhelpful first-level agent who claimed there was no such thing as second-level, that they never heard of such a thing, and the only person I could speak to was them)

 

They still couldn't see anything on their end, but dispatched a Rogers in-house technician to come and take a look. He couldn't find anything either, but referred my case to a senior in-house technician who could do a reverse sweep to the first active and call out a line maintenance tech if necessary. He forwarded my contact details to the senior tech and claimed I would be contacted soon.

 

Unfortunately, that was exactly one week ago (claimed I would receive a callback either last Friday or on Tuesday). Strangely, I also received a robocall last Monday (before the in-house tech visited) that my issue was "fixed" (it wasn't), repeated a case number, and said the case was now closed. I can only assume that somehow my case has been marked closed now in their internal system, and I will have to escalate all over again to get them to reach the same level again.

 

My issue is definitely not fixed, still seeing 3-5dB swings on several channels within 5-min time intervals (usually at night, when signals are worse). If anything, it has gotten worse as I managed to log the following just this afternoon - multiple swings down to 25-30dB on channel 31 (837MHz) with uncorrectables spikes on most surrounding channels (which has never happened before in the last 2 months that I've observed/logged; does not appear to be maintenance work as that usually impacts all channels).

 

 

In the meanwhile, I did manage to speak to a cable engineer (not from Rogers) who confirmed that any fluctuation beyond 0.5dB is indeed abnormal, let alone 3-5dB swings (at least at the ISP they worked at). They claimed my issue is temperature related, given signals are worse at night and the decline correlates with outside temperature change, and likely related to something being wrong with the AGC (automatic gain control) at the first amplifier behind the tap. (Also, I was wrong. There is no fiber at the pedestal and it is indeed a amplifier + tap located there; managed to get a closer look at the internals when the in-house tech visited)

 

I'm not so sure what to do now, other than wait for this issue to cause the entire neighbourhood problems, given it is near impossible to escalate this through frontline support (and I don't have the time/energy to repeat what I did last time). I've messaged @CommunityHelps  already, but either my message didn't get through or no one wants to take a look at my case (received no reply).

 

Any suggestions, or should I just resign myself to wait for whenever a maintenance tech has to adjust the amplifier due to the seasonal changes (if Rogers even does that for each neighbourhood these days) or whenever this causes catastrophic neighbourhood-wide issues? ( @-G- , @Datalink , @57 ?)

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

MitchellS
I'm here a lot

To the OP, while you've posted excellent signal strength details, I've only seen one reference to your actual download speed of 900 Mbits/s or more, which would be very very good. So what are your actual upload and download speeds at some different times of the day.

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

t_kwan
I plan to stick around

I'm not sure why I keep being asked about speed tests, as that's not the problem here. The problem is stability.

 

Turns out I was right about fall/winter being a problem, hence why I tried to get this fixed early while it was still in the summer (and this seems to confirm theories I've heard from cable techs/engineers about it being related to a faulty line amplifier adjusting to temperature changes).

 

I've now completely lost connectivity multiple times in the last few weeks, up to 3 times per day; so far, it's 2-3 mins of outage once it goes down, but it will likely get worse. The modem would abruptly start ranging - during those same times when I usually get uncorrectable spikes (~13:00, 18:00, 04:00, 09:00) - the status LED would flash orange for several seconds, followed by green and then repeating until it turned white again with connectivity restored.

 

It's not a modem restart as the uptime and CW stats remain unchanged after connectivity is restored, and I'm quite sure it's not a firmware upgrade, as that shouldn't be happening multiple times per day (just today it went down at ~1:23pm and then again at 6:41pm) and multiple times within a week.

 

Here's the SNR fluctuation graph from today (plotted are the raw difference between last sampled value and the next value). The fluctuation has only gotten worse as the weeks go by, and not better: 

 

Does anyone have any further advice @-G-@Datalink@57? And does anyone at Rogers care about my issue @CommunityHelps@RogersCorey?

 

I've gotten nothing back from that senior in-house Rogers tech who was supposed to contact me, and no reply from CommunityHelps. Calling into support only leads to me being told that contractor techs need to be sent out to inspect my modem/in-house wiring - starting again from the bottom of the chain - which is ridiculous given 3 contractors + 1 in-house Rogers tech have inspected them multiple times in a span of less than two months.

 

There's clearly a need for line/maintenance techs to be involved - that was where my case was going before it abruptly and randomly got closed - yet no one I talk to seems to have the ability to escalate it up back there without starting from ground zero.

 

Does my internet need to go out completely for Rogers to care, or do I need to escalate this to the CCTS?

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

MitchellS
I'm here a lot

What I was getting at is I don't think you have described what the end-user impact of these signal-level problems is – does the Internet stop working periodically throughout the day or just get slower. If so, then how often and for what duration and can you correlate that to times-of-day, or temperatures or weather. Also, it would be very helpful to know who else in your neighbourhood has these problems (since those "end user" symptoms are all they would know/report). Being able to determine the geographic area of these problems would help Rogers decide what component is defective, and Rogers might have record of what unresolved complaints others in the neighbourhood have (and where are they relative to you, as connected to their cable network).

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

57
Resident Expert
Resident Expert

@t_kwan : You asked for my comments, but I'm not sure if it'll help:

 

1. In 2020 I had intermittent issues with my Internet where it went down for a minute or two every once in a while. Occasionally it went down for longer - say 20 minutes. This happened 50 times over a period of 3 months.  (mainly August to October 2020). I contacted @CommunityHelps and they monitored my internet and the problem went away. I assumed that something in the line to our neighbourhood got fixed. I asked, but no-one knew what the fix actually was.

 

2. In 2021, about a year later, the same thing started happening. I again had about 50 brief outages over a 3 month period. (mainly August to October 2021).   Again, @CommunityHelps monitored my connection, but they could find nothing wrong this time.  A (good) technician was dispatched to my house and he found some issues with the main splitter in the grey box outside my home. One side of the splitter went to my modem and one went to another splitter feeding my TV boxes and Home Phone modem. At the time, I was still on Legacy Digital Cable.  He replaced that splitter and also replaced some cable ends.  The problem has not reoccurred and everything has been fine since October 2021.  The interesting thing is that that same splitter had just been replaced a year or two earlier when I was having SDV channel issues.  I switched to IgniteTV in late March 2022. 

 

3. In summary, it appears I've had several bad two-way splitters.  Perhaps because that splitter was outside in the grey box, caused premature failure.  When I switched to IgniteTV, I believe that splitter was removed since I only need one connection to my Gateway.

 

Good Luck.

 

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

-G-
Resident Expert
Resident Expert
@t_kwan wrote:

I've gotten nothing back from that senior in-house Rogers tech who was supposed to contact me, and no reply from CommunityHelps. Calling into support only leads to me being told that contractor techs need to be sent out to inspect my modem/in-house wiring - starting again from the bottom of the chain - which is ridiculous given 3 contractors + 1 in-house Rogers tech have inspected them multiple times in a span of less than two months.

 

There's clearly a need for line/maintenance techs to be involved - that was where my case was going before it abruptly and randomly got closed - yet no one I talk to seems to have the ability to escalate it up back there without starting from ground zero.

 

Does my internet need to go out completely for Rogers to care, or do I need to escalate this to the CCTS?



I just had some signal issues of my own; things started to degrade for me back in June and then, slowly, started getting worse and worse.  However, if the signal levels are within Rogers' acceptable norms and the issue is not service-impacting, it won't be a high priority for them to fix.  It wasn't until my power levels and SNR reached the bottom end of (and dipped below) acceptable levels that I could get the matter investigated.  By that time, the @CommunityHelps team was also able to confirm that my neighbours were also impacted and this was grounds to open a ticket.  Unfortunately, that ticket got closed without any action being taken... but we escalated again, and my area maintenance tech then finally got dispatched and fixed the problem the next morning.  The process took only a few days, and everything got fixed before it seriously impacted service.  My neighbours had no clue that there were even any problems in the area.

 

 

When you look at your modem's stats, you only see symptoms.  We don't even get access to any useful logs on the Ignite gateways... but Rogers has access to a ton of stats and they can very quickly localize problems.  HFC infrastructure is tricky to maintain, and if tech support recommends sending another Rogers field tech to your home, then DO THAT.  Even if your modem is dropping off the Rogers network and re-registering, it could be completely unrelated to your fluctuating SNR.  Tech support will be able to see why.  The problem may even be due to noise ingress in your own home, and totally unrelated to any anecdotes about other neighbourhood issues that you may have heard about.

 

As a next step, I would take the Ignite gateway and connect it at the point-of-entry.  Plug the gateway's power adapter directly into a wall outlet, not into a power bar.  Do not connect anything else to the gateway other than a computer, then check your stats.  If your SNR is still dancing around, call tech support again or send a PM to @CommunityHelps  and have them investigate.  However, if everything is normal in this standalone, isolated configuration, then you have an in-home issue that you need to troubleshoot.

 

The @CommunityHelps  team should also be able to see whether the SNR on your neighbours' modems are also fluctuating, or if there are any problems with their signal levels as well.  However, if Rogers does not see any other affected customers in your area, then the problem is localized to you home and they will not be able to escalate to a maintenance crew.  If Rogers does not see any problems at the tap in your pedestal, then they won't be able to escalate to a maintenance crew either.

 

Lastly, when dealing with tech support, don't fixate on the fluctuating SNR.  Don't show them graphs, pingplotter results, or any other tools that Rogers does not endorse or trust.  If the signal levels remain within acceptable norms and your modem is not disconnecting due to line conditions, and Speed Tests show that you are attaining your advertised speeds, then Rogers will not take any corrective action.  However, if your in-home network is in a Rogers-supported configuration and line conditions drop to a point that is service-impacting, then Rogers will work quickly to remedy the situation... and if Rogers needs to dispatch another tech, then work with them and work within (and do not try to shortcut) their processes.

Re: Terrible modem signal variation (and uncorrectables) that techs can't seem to fix

Hello @t_kwan!

 

I'd like to run some tests of my own and look into the documented history of this issue on your account. If you've had 4 techs out in the last two months, you may now qualify for a senior tech to come out and continue the investigation.

 

Feel free to send a private message to @CommunityHelps so we can assist you further. For more information on how our Private Messaging system works, you can find out more here.  

 

Regards,

RogersCorey

Topic Stats
  • 18 replies
  • 3045 views
  • 9 Likes
  • 6 in conversation