[BayMoWe335]

Originally Posted by M_Six

Are you running 5GHz or 2.4GHz? Our microwave oven used to screw up our 2.4GHz wifi, so I switched to 5GHz. That should be a very noticeable thing, though, so it doesn't sound like your problem.

It is dual band, so I run both.

[BayMoWe335]

Originally Posted by zx10guy

Microwaves, baby monitors, cordless phones, etc run over 2.4 GHz. Switching to 5GHz is one solution to get around interference from these devices.

Neighboring WiFi networks can also cause interference on 2.4 GHz, but also on 5 GHz. It's more likely to happen on 2.4 GHz than 5 but can still happen over 5 GHz.

A fix to help with neighboring RF interference on both 2.4 and 5 is to locate a channel on either which is not being heavily used to run your WiFi network on. You can figure this out by utilizing a WiFi tool which sniffs the RF space and gathers information on what devices are out there and what channel they're operating on. I have this free app I run on my Android phone called WiFi Analyzer which I use to do a crude site survey. Some wireless network devices will have that feature available for the network admin to use. The wireless system I have running in my house from Aruba Networks gives me the ability to do a full RF spectrum analysis by designating one of my wireless access points as the RF analyzer. As a feature of unified wireless systems is the ability for these systems to do an auto channel selection on both 2.4 and 5. These systems have intelligence in them where they monitor the RF space and do an automatic channel selection to avoid using noisy channels. I have not seen personally SOHO routers with this feature. One of the reasons why 5 GHz is preferred is because of more non overlapping channels available for one to use to avoid neighboring interference. 2.4 only has 3 non overlapping channels in the US to choose from (1, 6, and 11). So you can see how one can run into issues quickly using 2.4 GHz.

For 5 GHz, there are a ton more channels to choose from depending on the channel width (bonding) being used for the specific top end wireless speed you're trying to achieve which gets into a whole other rat hole. I'm attaching a chart which illustrates the number of available channels based on channel width.

Attachment 2211724

I should also note there is a friend in my area with the same fiber internet from the same ISP (ATT) and they have the SAME issue. It seems like it's a problem with the ATT equipment "main box" or wherever our signals ping when they leave our houses. We live 2 minutes apart.

[BayMoWe335]

Originally Posted by zx10guy

Run a continuous ping test where you're pinging google.com and the internal IP of your router at the same time. See if you're getting any dropped packets at the same time on both ping tests.

I was having this issue with Mediacrap at my vacation home. Couldn't stream anything. VPN connectivity when I worked remotely was crap. Even web pages were starting to load slowly. Mediacrap didn't even argue with me when I finally had the time to work with an onsite tech. The rep could see all the errors on their end. A tech was sent out. Found the TV block they put to prevent me from stealing basic cable was sopping wet. The block was removed which made the line signal better but it wasn't the sole issue.

For what ever reason, the coax line from the exterior DMARC of my home to the structured wiring cabinet went bad. The tech tested that run and the signal to noise ratio was unacceptable even though when I had service first established everything was fine. No work on my house that could have caused that issue. Fortunately, the sub that ran the structured wiring did the smart thing and ran an extra run of RG6. We tested that run and everything test fine. So I switched over to that line and it's been good since...except the typical Mediacrap performance issues.

I did a continuous ping and had 0 packet loss, but the ms ping time is widely variable.

I don't know how to ping the internal IP of my router and google at the same time.

edit: I did it by opening 2 different command prompts. When I ping my internal router, it's 3-4ms a lot but sometimes pver 30. When I ping Google, it's widely variable.

Internal IP:

Packets: Sent = 66, Received = 66, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 38ms, Average = 5ms

Google.com
Packets: Sent = 139, Received = 139, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 23ms, Maximum = 79ms, Average = 27ms

[pennsiveguy]

Originally Posted by BayMoWe335

I did a continuous ping and had 0 packet loss, but the ms ping time is widely variable.

I don't know how to ping the internal IP of my router and google at the same time.

edit: I did it by opening 2 different command prompts. When I ping my internal router, it's 3-4ms a lot but sometimes pver 30. When I ping Google, it's widely variable.

Internal IP:

Packets: Sent = 66, Received = 66, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 38ms, Average = 5ms

Google.com
Packets: Sent = 139, Received = 139, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 23ms, Maximum = 79ms, Average = 27ms

For whatever it's worth, here are my times to my firewall/router:

rtt min/avg/max/mdev = 0.128/0.202/0.251/0.046 ms

...and to Google:

rtt min/avg/max/mdev = 11.131/11.885/12.565/0.391 ms

Fairly fast, fairly consistent. No lost packets when running them simultaneously.

[jimbobiggens]

the ping time to your internal router should be almost instantaneous. Here is my wired ping results.


Here is my 2.4 GHz wireless ping results.

[pennsiveguy]

Originally Posted by BayMoWe335

I did a continuous ping and had 0 packet loss, but the ms ping time is widely variable.

I don't know how to ping the internal IP of my router and google at the same time.

edit: I did it by opening 2 different command prompts. When I ping my internal router, it's 3-4ms a lot but sometimes pver 30. When I ping Google, it's widely variable.

Internal IP:

Packets: Sent = 66, Received = 66, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 38ms, Average = 5ms

Google.com
Packets: Sent = 139, Received = 139, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 23ms, Maximum = 79ms, Average = 27ms

Can you take your router out of the device chain and connect directly to the modem's LAN ethernet port, and run some ping tests from there? While you're at it, you could run some QoS tests on various internet sites.

Somehow I suspect your router may be flaky. Many SoHo routers are marginally adequate at slow internet speeds but become flaky when you push them into the bandwidth territory you're in.

[zx10guy]

Originally Posted by BayMoWe335

I did a continuous ping and had 0 packet loss, but the ms ping time is widely variable.

I don't know how to ping the internal IP of my router and google at the same time.

edit: I did it by opening 2 different command prompts. When I ping my internal router, it's 3-4ms a lot but sometimes pver 30. When I ping Google, it's widely variable.

Internal IP:

Packets: Sent = 66, Received = 66, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 38ms, Average = 5ms

Google.com
Packets: Sent = 139, Received = 139, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 23ms, Maximum = 79ms, Average = 27ms

If you could put up a similar output of what jimbobiggens did with his continuous ping output, that would be helpful. You don't even need to do a screen shot. Just save the output off to a file. I do agree that you shouldn't have latency times more than a millisecond though. You should be getting what jimbobiggens has where it says less than 1 ms.

Per what pennsiveguy said, let's try plugging directly to the Ethernet hand off from the ONT. Hopefully AT&T isn't doing a MAC address registration which cable companies typically do. If so, you may have to spoof the MAC address of your router with your laptop/PC to get connectivity.

ETA: I just did a continuous ping test over wireless and I am getting an average 4ms latency. I did a continuous ping test to google.com and have an average latency of 15ms. I'm doing this over my Comcast service at my primary home.

[Drv4fun]

We see this a lot (I own an IT co). Usually it’s routers that do not have enough processing power, a lot of times it’s poor cabling (assuming the ISP is not having the same results upstream of your connection). First, we have to establish the issue is on wired connections. Assume all further suggestions are on wired only connections. Try a new certified Cat6 patch cable from your router to the modem (that would be store bought, not self made/crimped). Did you try the modem directly wired to a computer as suggested by others earlier? If your neighbor is having the same issue its likely the router upstream (ISP) is busy and may delay in icmp (ping) reply. Ping is not an ‘only’ tool. Its just one trouble shooting step, but a good one nonetheless.

[sethwas]

Originally Posted by BayMoWe335

Long story short, I have ATT fiber internet (300 Mbps down and 300Mbps Up) that tests those speeds all day long.

Problem is, my internet sucks. It's not that it isn't fast. It's that it constantly drops for 1-2 seconds so it's always hanging on simple Google searches.

I see this thing professionally a lot, and unfortunately there isn't a single cause.
I read further replies and I need some more info to find out where the point of failure could be.

First, if you connect your PC/laptop directly via a wire to your gateway provided by the ISP does this occur?

It wasn't clear if you were connecting with a wired or wireless device, and/or how many intermediate devices there are in this setup.

the ISP will tell you nothing is wrong if they can communicate over their advertised "speeds" to their gateway inside your home. Their responsibility ends there.

So in summary, for this test, there should be no network, just the device provided by the utility to your testing machine via a wire, and the machine should have no programs open only a speedtest app (not in the browser as this adds a layer of complexity) or command prompt.

Seth

[cncmastr]

https://blog.daveasprey.com/dangers-...h-mercola-424/

[BayMoWe335]

I cancelled ATT and went with spectrum. Problem solved, blazing fast speeds.

[BayMoWe335]

Originally Posted by sethwas

I see this thing professionally a lot, and unfortunately there isn't a single cause.
I read further replies and I need some more info to find out where the point of failure could be.

First, if you connect your PC/laptop directly via a wire to your gateway provided by the ISP does this occur?

It wasn't clear if you were connecting with a wired or wireless device, and/or how many intermediate devices there are in this setup.

the ISP will tell you nothing is wrong if they can communicate over their advertised "speeds" to their gateway inside your home. Their responsibility ends there.

So in summary, for this test, there should be no network, just the device provided by the utility to your testing machine via a wire, and the machine should have no programs open only a speedtest app (not in the browser as this adds a layer of complexity) or command prompt.

Seth

It was definitely an ATT network issue in my area. People either too oblivious to complain or they just lived with it. Spectrum is perfect now.

[2000cs]

Originally Posted by Drv4fun

We see this a lot (I own an IT co). Usually it’s routers that do not have enough processing power, a lot of times it’s poor cabling (assuming the ISP is not having the same results upstream of your connection). First, we have to establish the issue is on wired connections. Assume all further suggestions are on wired only connections. Try a new certified Cat6 patch cable from your router to the modem (that would be store bought, not self made/crimped). Did you try the modem directly wired to a computer as suggested by others earlier? If your neighbor is having the same issue its likely the router upstream (ISP) is busy and may delay in icmp (ping) reply. Ping is not an ‘only’ tool. Its just one trouble shooting step, but a good one nonetheless.

Routers seem to be a weak link in all of these conversations. Do you have a recommendation for a better household router (wi-fi) that has enough processing power and can reliably work at 400MBPS or 1GB? Or are they all just crap?

[BayMoWe335]

Originally Posted by 2000cs

Routers seem to be a weak link in all of these conversations. Do you have a recommendation for a better household router (wi-fi) that has enough processing power and can reliably work at 400MBPS or 1GB? Or are they all just crap?

I have a Netgear r7800 and it’s perfect for my house. I have 20 devices connected and no issues on a 4000mbps connection.

[Drv4fun]

Originally Posted by 2000cs

Routers seem to be a weak link in all of these conversations. Do you have a recommendation for a better household router (wi-fi) that has enough processing power and can reliably work at 400MBPS or 1GB? Or are they all just crap?

Retail consumer routers are just that, made for the home. It has taken a bit of time to get better hardware in them to be able to sustain the higher throughput not to mention if the firewall has decent protection (the more firewall inspection the more processing is required). High end retail consumer units will likely work at 1Gbps but I doubt they are designed with constant 1Gbps throughout in mind. Most times no one is saturating a connection at high bandwidth anyway except for speedtests and those are not full duplex.

I don't have any recommendations. My own router has a 250Mbps limitation (even though it has a physical 1Gbps link) and my connection is 300. I don't have any issues.

[Drv4fun]

Originally Posted by BayMoWe335

I have a Netgear r7800 and it’s perfect for my house. I have 20 devices connected and no issues on a 4000mbps connection.

Was this the same router when you had ATT?

[BayMoWe335]

Originally Posted by Drv4fun

Was this the same router when you had ATT?

Yes, exactly the same setup. ATT service was just garbage.

Funny enough, ATT recently sent out several flyers saying they have "improved our service in your area."

I guarantee a switch station, line, or something was screwed up and they finally got enough complaints.