dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
1018
RandomBlue
join:2012-10-30
Rogers, AR

1 edit

RandomBlue

Member

[AR] Outages, slowdowns and question about UPS usage.

Click for full size
Modem Status
Click for full size
Modem Signal
Click for full size
Modem Config
Click for full size
Modem Log
I've been experience service outages and 2-3%+ PL for the last week and had a Cox tech out here today. He said my signal levels looked fine and noticed my Motorola SB6120 had some T3/T4 errors in the logs and said those might have caused the issue. He also stated that running the SB6120 through a power strip or UPS could cause those to happen. It seems counter-intuitive to me that they're recommending I avoid using a surge surpressor or UPS, especially in our area where thunderstorms are common. Is what the tech said true?

He did install a video filter on my modem and that seems to have helped with the packet loss. It was running with 0% PL for the last 12 minutes whereas before we couldn't go 1-2 minutes without PL but as I typed this message it just dropped 3 consecutive packets at the first hop.

*sigh*

We have been experiencing neighborhood level events (outages/slowdowns) recently as well but it appears they self resolved and Cox is avoiding looking further into the issue. Saturday night when I called in an outage they said 98% of our neighborhood (a few hundred homes I believe) were down. Sunday afternoon we had a similar but shorter event and when I called it in the rep said they were seeing something at the neighborhood level but wanted to send the tech out today to check things on my end.

Here are my modem stats and they've been pretty much the same every time I look at them:

Rcv signal: 38/38/39/39
Power level: 1-2dBmV
Upstream power level: 42-43 dBmV

The last time a T3 timeout is mentioned in the logs is yesterday and we had an outage/slowdown this morning for a short period.
signcarver
join:2005-03-20
Phoenix, AZ

signcarver

Member

I think the no surge suppressor thing was generally for the cable itself as some can block or interfere with signals but if you have no problems using it (compared to not using it) I would probably want to err on the side of protecting my equipment vs an occasional packet loss as the 2nd doesn't cost me money but I'm out $100 if the modem is fried.

Though my feeling is more on the rf side of not using the power strip, the issue could also be electrical. One of my homes is in an area with poor ground (new homes now require a whole house grounding system rather than just a couple of grounding rods) and as a result, in the older houses (though none are older than 15 years), electronics in that area go out quickly and I've seen the following problem occur even on single level houses from one end to the other. It used to be in networking we were taught not to run cat5 vertically and use fiber between floors (and buildings) as it increased the likelihood that the ground differential (potential ?) was different between one floor (or building) and the other... as transmission speeds increased this became more of a problem since the voltage levels (bits) aren't true on/off but differences in ups and downs so they don't have to wait for the voltage to drop or rise completely. In theory surge protectors/ups's can also disrupt this but I would say generally there would not be a problem.

So it would not surprise me if a cable tech who has had some training would pick up on the idea of not using UPS's and surge protectors on the networking equipment as they are a possible problem... I just don't think they are that likely to cause a problem in most homes.
RandomBlue
join:2012-10-30
Rogers, AR

RandomBlue

Member

said by signcarver:

I think the no surge suppressor thing was generally for the cable itself as some can block or interfere with signals but if you have no problems using it (compared to not using it) I would probably want to err on the side of protecting my equipment vs an occasional packet loss as the 2nd doesn't cost me money but I'm out $100 if the modem is fried.

Yeah, he was talking solely about running the power through surge protectors and UPSes. I had mine hooked up to a decent size APC UPS and I'm going to switch it back. I never run RF through them.

I swapped out the modem to a Cisco DPC3010 and both modems are now seeing about 1% packet loss that happens in bursts of 3-4 lost packets every 12-15 minutes or so, which equates to 3-4 seconds of lost connectivity.

Both modems display the same issue with about the same frequency, though with the Cisco I get better signal and I'm bonding 5 channels instead of 4. Guess I'll have to set up yet another service call.

Here's the Cisco's signal info:

Model: Cisco DPC3010
Vendor: Cisco
Hardware Revision: 1.0
Bootloader Revision: 2.3.0_R1
Current Software Revision: DPC3010-v302r12901-110714a-CHR
Firmware Name: dpc3010-v302r12901-110714a-CHR.bin
Firmware Build Time: Jul 14 10:56:53 2011
Cable Modem Status: Operational

Downstream Channels
Power Level: Signal to Noise Ratio:
Channel 1: 0.3 dBmV 41.5 dB
Channel 2: 0.0 dBmV 41.1 dB
Channel 3: 0.0 dBmV 41.9 dB
Channel 4: 0.0 dBmV 41.5 dB
Channel 5: -1.1 dBmV 41.4 dB
Channel 6: 0.0 dBmV 0.0 dB
Channel 7: 0.0 dBmV 0.0 dB
Channel 8: 0.0 dBmV 0.0 dB

Upstream Channels
Power Level:
Channel 1: 44.5 dBmV
Channel 2: 0.0 dBmV
Channel 3: 0.0 dBmV
Channel 4: 0.0 dBmV
RandomBlue

RandomBlue

Member

Click for full size
Last 6 minute event
Here's current info based on what was requested in the forum sticky:

Symptoms: 3-4 second dropouts/lost connectivity every now and then throughout the day and longer periods with high PL and slow or no connectivity, sometimes lasting 5-10+ minutes.

Modem: Cisco DPC3010
Router: Asus RT-N16 running TomatoUSB 1.28 (it's been running fine for several months)
Speed Tier: Ultimate
Signal Levels: 41.0-41.9 for all 5 channels, has been in that range every time I've looked.
Splitters: None, plugged directly into the wall except for the video filter/trap added by the tech earlier today.
Speed tests: All are fine when the line is actually working now.
Pingplotter results will be attached.
Tweaked? - No, Windows 7 and Windows 8
Wired/Wireless: Wired - gigabit ethernet
Bypassed router: No, but I have rebooted it during long outages with no change. The cable modem also has trouble reconnecting during the long outages.
TV Affected: No

Modem signal info during last medium duration event (6 minutes):

Power Level: Signal to Noise Ratio:
Channel 1: 0.0 dBmV 41.8 dB
Channel 2: -0.3 dBmV 41.1 dB
Channel 3: -0.3 dBmV 41.8 dB
Channel 4: -0.3 dBmV 42.0 dB
Channel 5: -1.4 dBmV 41.4 dB
Channel 6: 0.0 dBmV 0.0 dB
Channel 7: 0.0 dBmV 0.0 dB
Channel 8: 0.0 dBmV 0.0 dB

Upstream Channels
Power Level:
Channel 1: 44.0 dBmV
Channel 2: 0.0 dBmV
Channel 3: 0.0 dBmV
Channel 4: 0.0 dBmV

I'm attaching a screenshot of the PingPlotter tracking of the last 6 minute event with 17% PL to the first hop and no ability to open any websites.
RandomBlue

1 edit

RandomBlue

Member

Click for full size
Part of the ongoing packet loss today
Currently going through a pretty long event that's been going on for a little over 30 minutes. 20% packet loss during that timeframe. I managed to get Cox support on the phone while it was happening and it took a few tests for them to finally see the packet loss but they did finally see 60% PL on one of their tests even though the modem's signal and power levels look fine. They're running a 24 hour monitor on the line and sending a tech out tomorrow.

Due to the high signal and power levels on the modem and the recent neighborhood level events/outages I'm thinking this is occurring at the neighborhood cox hub and I just happen to be the one reporting it since I work from home and I'm online pretty much all day every day.

Also found out they're charging me for the tech that came out yesterday simply because I did a self-install on my cable-card last week (even though it took 3 cox techs 6 hours to actually get that working) because apparently there's a policy where calls within 30 days of a self install have to be charged. I'll definitely be arguing this one with billing though since the tech didn't fix the problem and the problem is not with my wiring or equipment.

The cable-card install was awesome as well. 3 hours before they figured out no-one had removed the trap/filter outside and another 3 before they gave up and I found out the blinking yellow light on the tuning adapter meant it hadn't been activated. One call to support line to get the automated system to send an activation to the adapter and it was up and running.
RandomBlue

RandomBlue

Member

Click for full size
Tonight's first long event.
Today they replaced the connectors on the line going to my cable-modem and the connectors outside in the box and ran a new drop to the post. Things looked good at that point but I wasn't having consistent packet loss at that point anyway so it was hard to tell. Unfortunately it didn't fix the issue, tonight from 5:40-7:15 I had 12% packetloss with spikes much higher than that at the start. See the attached pingplotter graph showing the first hour.
RandomBlue

RandomBlue

Member

FYI, here are my current levels after running the new drop:

Power Level: Signal to Noise Ratio:
Channel 1: 5.2 dBmV 41.6 dB
Channel 2: 4.6 dBmV 41.3 dB
Channel 3: 5.1 dBmV 41.8 dB
Channel 4: 4.9 dBmV 42.3 dB
Channel 5: 4.5 dBmV 41.5 dB
Channel 6: 0.0 dBmV 0.0 dB
Channel 7: 0.0 dBmV 0.0 dB
Channel 8: 0.0 dBmV 0.0 dB

Upstream Channels
Power Level:
Channel 1: 42.5 dBmV
Channel 2: 0.0 dBmV
Channel 3: 0.0 dBmV
Channel 4: 0.0 dBmV
RandomBlue

RandomBlue

Member

cox.net.2.pp2.zip
663,309 bytes
Attaching one of my recent pingplotter data sets.
RandomBlue

RandomBlue

Member

Click for full size
So another packetloss storm tonight, see the attached graph. Called in and got to Tier 2 after only 20 minutes and they saw some packetloss on my line. They also said another neighbor reported packetloss tonight, so hopefully they'll start looking at a broader issue than just my line. They've done practically all they can do with just my house and I've been pretty certain since the first couple of neighborhood issues that this wasn't solely my house.