Optimal network architecture (to replace equipment that has no stability)?

riotubes

n00b
Joined
Dec 4, 2010
Messages
25
Hello,
I'm at my wits' end and need some helpful advise from an experienced networking pro.
I have Spectrum 200 mbps service and there are no other viable alternatives here at the house at which we've lived for 8 years. Nothing but constant problems. I've traded out modems 4 times, tried countless routers, had Spectrum field tech visit at least 8-10 times.
Until 5 days ago, my equipment was: Netgear CM600 >> TP-Link Safestream TL-R600VPN (not using a vpn; not a wifi router) >> hardwired to Orbi RBR50 Mesh (main and satellite) used as WAP only. I have the TV hardwired into the TP-Link, and run ethernet from another TP-link port to the Orbi Mesh, and run one ethernet cable from Orbi into the main PC. See image below. While I have ethernet wired through the attic (runs between router and main Orbi), in order to eliminate the ethernet cable as the culprit, today I replaced the attic run with a brand new CAT7 cable that runs across the floor. Still no wifi, although the main pc has internet through its hardwire connection to the Orbi.

Wifi devices on our LAN: 4 ipads, 5 mobile phones, a gaming pc, 2 laptops, a sonos soundbar and a network audio receiver. I have 3 devices hardwired into the main Orbi appliance (main PC, VOIP phone, and Sonos Connect appliance (to enable playbar to work).

I experienced 235 mbps rock-solid performance for 6 months, the longest continuous "good" stretch since we've been here. Then speeds dropped to 40 mbps across the board for past 3 days. After speaking with Spectrum, I picked up one of their modems (only) and replaced the CM600. Now constant wifi dropouts. The pc hardwired to the TP-link router is fairly stable at much slower speeds (but with intermittent drops). Prior to this hardware config (modem>>hardwire router>>WAP) I used various routers with integrated wifi all of which produced horrible performance (constant intermittent dropouts).

The numerous spectrum field techs have proposed different solutions; some want to insert splitters, others remove them, some want to insert signal attenuators on the modem, others take them out. And, unfortunately, I don't know what I'm doing either. I simply want stable, 100+ mbps speeds. I'm thinking of replacing the router and Orbi with Cisco or Ubiquiti gear (router and WAPs) however how I do rule out this is a Spectrum problem? The wifi is basically non-existent at this point. I would say it could be the Orbi but the TP-Link router is also dropping out intermittently (although less so). Ideas on a more permanent fix, or how I birddog this?

Slide1.JPG
 
Last edited:
Joined
Dec 1, 2011
Messages
937
Does the problem occur if you pull everything off of the cable modem, and directly wire a PC into it? Have you tried pulling status reports from the cable modem, both when it's seemingly acting normal and when not? Are there any particular patterns to the issue (time of day, weather, usage, etc.)? Any signs of physical damage or possible water ingress to the coax?
 

SamirD

2[H]4U
Joined
Mar 22, 2015
Messages
3,988
Does the problem occur if you pull everything off of the cable modem, and directly wire a PC into it? Have you tried pulling status reports from the cable modem, both when it's seemingly acting normal and when not? Are there any particular patterns to the issue (time of day, weather, usage, etc.)? Any signs of physical damage or possible water ingress to the coax?
This is where I would start too. I've got a couple of accounts through spectrum and while they do have outages, they usually hold up better than the other isps. If you have problems straight at the modem, it's definitely the service.

Once you're hooked up direct, use packetlosstest.com and the dslreports.com and fast.com speedtests to see what you're getting (I've found these both to be more reliable than anything else).
 

riotubes

n00b
Joined
Dec 4, 2010
Messages
25
Thank you both for your prompt reply. I am concerned about water ingress because we had a rainstorm and the coax from telephone pole terminates into a splitter inside a plastic box affixed to the side of the home before this coax run terminates inside the garage where the modem is located. I worked on this all day yesterday. After 6 hours, I could not get the Orbi to work in WAP mode with the new Spectrum supplied modem positioned after the TP-Link router, so I gave up on that hardware configuration. I removed the TP-link router and used the Orbi in router mode and plugged it directly into the Spectrum modem. Per BlueLineSwinger's and SamirD's suggestions, I used packetlosstest, Fast and NPerf to measure many different configuration scenarios in which I traded out various equipment and ethernet cable to try to isolate the problem. I found a free program called Star Trinity Continuous Speed Test that also measures packet loss and jitter. The output can be exported as a csv file on which I ran some stats (see below). Because there are over 3,500 rows of measurement, I've only included a few dozen to give you a sense for the output from this diagnostic tool. But the other summary metrics are based on all 3,565 rows of measurements. This tool strikes me as much more helpful than others (FAST, NPERF, Speedtest.net) that only take a single or short series of measurements. What do you make of these results?

Networking Stats.jpg
 
Last edited:
Joined
Dec 1, 2011
Messages
937
I have no idea what that test is running across. If it's operating as a typical internet speed test, it's mostly useless for diagnosing what may be an issue with your cable modem and/or ISP connection.

Have you pulled status reports from the cable modem as I suggested above? It varies by make/model, but often you can get into the cable modem by accessing 192.168.100.1 in a browser. For example, this is what mine looks like:
Code:
Startup Procedure
Procedure     Status     Comment
Acquire Downstream Channel     651000000 Hz     Locked
Connectivity State     OK     Operational
Boot State     OK     Operational
Security     Enabled     BPI+
IP Provisioning Mode     Honor MDD     honorMdd(4)
 
Downstream Bonded Channels
Channel     Lock Status     Modulation     Channel ID     Frequency     Power     SNR     Correctables     Uncorrectables
1    Locked    QAM256    44    651000000 Hz    -6.2 dBmV    39.4 dB    11164    47
2    Locked    QAM256    17    483000000 Hz    -6 dBmV    40 dB    20635    409
3    Locked    QAM256    18    489000000 Hz    -6.1 dBmV    39.9 dB    20693    755
4    Locked    QAM256    19    495000000 Hz    -5.7 dBmV    40.1 dB    20307    762
5    Locked    QAM256    20    507000000 Hz    -5.4 dBmV    40 dB    19213    984
6    Locked    QAM256    21    513000000 Hz    -5.6 dBmV    40 dB    18738    746
7    Locked    QAM256    22    519000000 Hz    -5.7 dBmV    40 dB    18024    914
8    Locked    QAM256    23    525000000 Hz    -6.1 dBmV    39.8 dB    17555    909
9    Locked    QAM256    24    531000000 Hz    -5.9 dBmV    40 dB    15928    582
10    Locked    QAM256    25    537000000 Hz    -6.1 dBmV    40 dB    15403    722
11    Locked    QAM256    26    543000000 Hz    -6 dBmV    39.9 dB    14384    106
12    Locked    QAM256    27    549000000 Hz    -6.3 dBmV    39.8 dB    14756    816
13    Locked    QAM256    28    555000000 Hz    -6.4 dBmV    39.7 dB    14913    844
14    Locked    QAM256    29    561000000 Hz    -6.4 dBmV    39.6 dB    15318    815
15    Locked    QAM256    30    567000000 Hz    -6.6 dBmV    39.6 dB    14855    807
16    Locked    QAM256    31    573000000 Hz    -6.3 dBmV    39.7 dB    14799    826
17    Locked    QAM256    32    579000000 Hz    -6.3 dBmV    39.6 dB    15199    912
18    Locked    QAM256    33    585000000 Hz    -6.3 dBmV    39.6 dB    15064    941
19    Locked    QAM256    34    591000000 Hz    -6.4 dBmV    39.5 dB    15254    969
20    Locked    QAM256    35    597000000 Hz    -6.5 dBmV    39.5 dB    14847    661
21    Locked    QAM256    36    603000000 Hz    -6.7 dBmV    39.4 dB    14467    576
22    Locked    QAM256    37    609000000 Hz    -6.8 dBmV    39.4 dB    13722    926
23    Locked    QAM256    38    615000000 Hz    -6.7 dBmV    39.4 dB    13049    873
24    Locked    QAM256    39    621000000 Hz    -6.7 dBmV    39.3 dB    12450    798
25    Locked    QAM256    40    627000000 Hz    -6.7 dBmV    39.4 dB    12186    873
26    Locked    QAM256    41    633000000 Hz    -6.6 dBmV    39.4 dB    11866    793
27    Locked    QAM256    42    639000000 Hz    -6.4 dBmV    39.3 dB    11133    840
28    Locked    QAM256    43    645000000 Hz    -6.2 dBmV    39.4 dB    10840    858
29    Locked    QAM256    45    657000000 Hz    -6.2 dBmV    39.3 dB    11638    889
30    Locked    QAM256    46    663000000 Hz    -6 dBmV    39.2 dB    11881    836
31    Locked    QAM256    47    669000000 Hz    -5.9 dBmV    39.4 dB    11518    824
32    Not Locked    Unknown    0    0 Hz    0.0 dBmV    0.0 dB    0    0
 
Upstream Bonded Channels
Channel     Lock Status     US Channel Type     Channel ID     Symbol Rate     Frequency     Power
1    Locked    ATDMA    1    5120 Ksym/sec    16400000 Hz    45 dBmV
2    Locked    ATDMA    2    5120 Ksym/sec    22800000 Hz    44.3 dBmV
3    Locked    ATDMA    3    5120 Ksym/sec    29200000 Hz    44.3 dBmV
4    Locked    ATDMA    4    5120 Ksym/sec    35600000 Hz    44.5 dBmV
5    Not Locked    Unknown    0    0 Ksym/sec    0 Hz    0.0 dBmV
6    Not Locked    Unknown    0    0 Ksym/sec    0 Hz    0.0 dBmV
7    Not Locked    Unknown    0    0 Ksym/sec    0 Hz    0.0 dBmV
8    Not Locked    Unknown    0    0 Ksym/sec    0 Hz    0.0 dBmV
 
Downstream OFDM Channels
Channel     Lock
Status     Modulation /
Profile ID     Channel
ID     Frequency     Power     SNR /
MER     Active Subcarrier
Number Range     Unerrored
Codewords     Correctable
Codewords     Uncorrectable
Codewords
1    Locked    0 ,1 ,2 ,3    48    850000000 Hz    -10.4 dBmV    37.7 dB    1108 ~ 2987    53327762116    50319980500    9941
2    Not Locked    0    0    0 Hz    0 dBmV    0.0 dB    0 ~ 4095    0    0    0
 
Upstream OFDMA Channels
Channel     Lock Status     Modulation / Profile ID     Channel ID     Frequency     Power
1    Not Locked    0    0    0 Hz    0 dBmV
2    Not Locked    0    0    0 Hz    0 dBmV
 
Current System Time:Sun Jan 03 13:24:47 2021
System Up Time:59 days 12:38:31
 

SamirD

2[H]4U
Joined
Mar 22, 2015
Messages
3,988
Judging by just the TX Loss column, there's a problem with your service. It's not much loss, but it is probably getting worse at certain times leading to the problems with RX and others.

If you want continuous monitoring like this, check out the thinkbroadband free monitors. I use them to keep tabs on all of my isp accounts continuously, and the graphs are really easy to read to spot problems:
cf44d9b76e6c018cabefc9c2d265ab0f138e170e.png
382b0bb8ed0f52d92889475cacb7c6016493ff34.png
1447cb89d4a66d170c457abf262f40b5c85000a4.png

Like this micro outage the other day:

Or explains the sluggishness and long ping times this day:
 

riotubes

n00b
Joined
Dec 4, 2010
Messages
25
BlueLineSwinger, I had meant to note above that, as strange as it sounds, the Spectrum supplied modems do not allow access to event logs (via login to IP address). This is pretty frustrating and renders one even more reliant on Spectrum's customer service and response times. I should have been more descriptive, but the column headers in the stats I published above are Download Mbps, Upload Mbps, RX Loss %, TX Loss %, and I presume the jitter is measured in ms.
SamirD, thanks for your feedback. I think it was water ingress as they did not cap one of the coax female connectors not in use and we had a big rainstorm.
While my speeds are much slower, at least the service is up. Spectrum can't get out here until Jan 13 so hopefully the service will remain up until then.
Thanks for your help.
 

SamirD

2[H]4U
Joined
Mar 22, 2015
Messages
3,988
Ah yes, water can do that. I've also had issues when the temp changes and there are line issues further up the circuit. Spectrum is usually good at fixing stuff once they're on top of it, so hopefully you're on the road to a solution. I would still keep monitors and logs though--just in case, and also to validate that the problem is truly fixed.
 
Top