speedtest result over wi-fi

[darth_adversor]

Hey all,

So I just got AT&T fiber internet installed today, 300 mbps up and down. I set the gateway up for IP passthrough, so I could use my TP-Link Archer C5 (with OpenWRT) instead of the unit AT&T provided.

My wireless throughput over the LAN averages between 45-50 MB/s (360-400 mbps), yet when I run a speedtest, I only get around 170 mbps. On the same machine over ethernet, I get the full 300 mbps.

I understand ethernet will be faster than wireless, but even at the low end of 360 mbps via the wireless, I should be able to realize the full bandwidth from my ISP.

Thoughts?

 

[L&LD]

Be sure you are not too close to the router. Between 8 to 20 feet and in-line of sight is ideal.

Assuming you are on 5GHz, correct? Use one of the lower numbered channels and test them all (36 to 48). Do not use the Auto Channel mode. Disable Universal beamforming.

Make sure the client device is in Maximum Performance mode when testing via Wireless.

Make sure the antennae are straight up.

Test all options for Channel Width.

Test with a new 8 character SSID and 16 character password. With no spaces, punctuation, special characters and no smiley faces.

 

[darth_adversor]

All great ideas, thank you. I’ll give all of that a try and report back.

 

[MichaelCG]

What type of WiFi features/specs does your client support?

 

[sfx2000]

So I just got AT&T fiber internet installed today, 300 mbps up and down. I set the gateway up for IP passthrough, so I could use my TP-Link Archer C5 (with OpenWRT) instead of the unit AT&T provided.

Depends on the unit that ATT dropped into your place...

Putting it into bridge mode can impact performance - there's magic sauce in the RNG/Gateway that can optimize performance that a consumer device can't.

Since you're on OpenWRT - take a look at Cake SQM, and throttle things back to 90 percent, and you'll have a sweet connection

 

[sfx2000]

My wireless throughput over the LAN averages between 45-50 MB/s (360-400 mbps), yet when I run a speedtest, I only get around 170 mbps. On the same machine over ethernet, I get the full 300 mbps.

What's the WiFi adapter you have?

Also with the Archer C5, depends on what version you have - V1 is QCA, V2 is Broadcom - the V2 is pretty spooky, the QCA is better...

 

[RogerSC]

Also on AT&T fiber, 100/100 tier provisioned at about 123/123. Using an eero mesh system, same IP Passthrough connection with the BGW210. I see full speed on a wireless speed test when the remote nodes are connected via MoCA. When the MoCA is disconnected, I see about 5% lower speed on a wireless speed test. The wireless speed tests are usually done with an iPad Pro 10.5".

By the way, when I started looking at AT&T fiber I was thinking Gb. fiber, wouldn't that be cool *smile*? However, decided to try 100/100 to see if that would work for us, and it's turned out to be enough for our needs. We've "cut the cable TV", and do a lot of streaming with both of us streaming different TV shows at the same time, along with mobile device use, computers and laptops, etc. Anyways, we find that the 100/100 tier is fine for us, and saves us $30-$40/month. If we need more speed, though, there are two more speed tiers available. Data cap isn't an issue at this point, the OTA antenna helps with that, but if that should change, we'll most likely go to 1Gb., since there's no data cap on that speed tier. Very happy with the "symmetric" fiber, consistent speed and reliable. No more problems with internet speed going down by 50% or so at prime time. Things are working well here at the moment.

 

[mjc775]

A lot depends on the WiFi channel width. On 2.4 GHz the standard channel width is 20 MHz wide - I’ve never seen over 100 Mbps. On 5 GHz the standard channel width is 80 MHz wide - I’ve seen 400+.


Sent from my iPhone using Tapatalk

 

[darth_adversor]

Hey all, sorry, I've had a busy few days, and I didn't realize I had so many additional responses.

Things I've tried and thoughts:

1. I did a channel scan to make sure I wasn't getting interference from a neighbor. I settled on channel 40. I just noticed, however, that even when I manually select a channel, sometimes OpenWRT reports a different channel being used. Strange. Anyone ever experience that?
2. I couldn't find a "universal beamforming" option in the settings.
3. Channel width didn't make much of a difference.
4. I've confirmed that my wireless chip, in power options (Win 10), is set to "maximum performance."
5. Antennae are straight up.
6. I have to admit that I did not try changing the SSID or password. That sounds like voo-doo.

 

[darth_adversor]

What's the WiFi adapter you have?

Also with the Archer C5, depends on what version you have - V1 is QCA, V2 is Broadcom - the V2 is pretty spooky, the QCA is better...

My adapter is the Intel 8260 2x2. My Archer C5 is V1.

I haven't heard of Cake SQM, I'll check that out. My next step is to reset the gateway back to default, disable my Archer, and try running a series of tests using solely their equipment. Honestly, I think their provided gateway is actually superior to my 4 year-old TP-Link, I just don't like the idea of using a device with standard, probably-not-as-secure firmware.

 

[darth_adversor]

Also on AT&T fiber, 100/100 tier provisioned at about 123/123. Using an eero mesh system, same IP Passthrough connection with the BGW210. I see full speed on a wireless speed test when the remote nodes are connected via MoCA. When the MoCA is disconnected, I see about 5% lower speed on a wireless speed test. The wireless speed tests are usually done with an iPad Pro 10.5".

By the way, when I started looking at AT&T fiber I was thinking Gb. fiber, wouldn't that be cool *smile*? However, decided to try 100/100 to see if that would work for us, and it's turned out to be enough for our needs. We've "cut the cable TV", and do a lot of streaming with both of us streaming different TV shows at the same time, along with mobile device use, computers and laptops, etc. Anyways, we find that the 100/100 tier is fine for us, and saves us $30-$40/month. If we need more speed, though, there are two more speed tiers available. Data cap isn't an issue at this point, the OTA antenna helps with that, but if that should change, we'll most likely go to 1Gb., since there's no data cap on that speed tier. Very happy with the "symmetric" fiber, consistent speed and reliable. No more problems with internet speed going down by 50% or so at prime time. Things are working well here at the moment.

I went with the 300 mbps package to avoid the data cap. I agree with you, otherwise. 100 is probably sufficient for most people. Also yes, I'm really pleased with the service. I was getting 5-6 mbps upload with Comcast. When the sales rep knocked on my door and said "300 up and down," I did a double-take.

You lost me on MoCA, though, I don't know what that means. Anyway, a 5% difference still wouldn't explain why wireless file transfers between computers run at 350-450 mbps, while my speedtest tops out at 170 mbps.

 

[L&LD]

Hey all, sorry, I've had a busy few days, and I didn't realize I had so many additional responses.

Things I've tried and thoughts:

1. I did a channel scan to make sure I wasn't getting interference from a neighbor. I settled on channel 40. I just noticed, however, that even when I manually select a channel, sometimes OpenWRT reports a different channel being used. Strange. Anyone ever experience that?
2. I couldn't find a "universal beamforming" option in the settings.
3. Channel width didn't make much of a difference.
4. I've confirmed that my wireless chip, in power options (Win 10), is set to "maximum performance."
5. Antennae are straight up.
6. I have to admit that I did not try changing the SSID or password. That sounds like voo-doo.

6. No, that isn't voodoo.

What you're effectively doing is allowing your clients to re-negotiate with the current best wireless settings between the router and themselves and save that microcode for that specific network for future use.

When you're using old SSID's, that old saved microcode may or may not be optimal.

 

[RogerSC]

Yes, the 5% that I mentioned would not explain your wireless speeds. What I was doing was presenting an alternative home network setup that's working a lot better with essentially the same internet connection hardware. I have no control over channels or how many SSID's are used, etc., but the mesh system really works well for wireless speed and coverage.

Also, MoCA (Multimedia over Coax Alliance) uses the coax cable already in your house's or apartment's walls for cable TV as a "wired-equivalent" to putting ethernet cable in your walls. Our house is on a slab foundation without a full attic, so adding ethernet cabling to the walls isn't going to happen. There is coax in the walls, though, for cable TV, which is very useful. The MoCA adapter at each end point of each run of coax adds a little bit of latency, couple of ms., but makes the cable act as a full speed, wired-equivalent ethernet link. So you end up with wired connection speeds with just a little added latency.

Very useful, and interesting to set up. You do need to have access to cable splitters and what not, though, so that they can be upgraded to MoCA compatible ones. Ours is MoCA 2.0, fast enough for our needs.

 

[ACwifiguy]

With a 2x2 AC client you should be able to easily get 300 mbps in line of sight with 5ghz and 80Mhz channels. I have a similar set up with beamforming and DD-WRT. I was able to max out my ISP internet connection easily over WiFi.

If you are not getting that throughput it is either the settings, OpenWRT limitation, or a hardware issue. I’d go in that order when problem solving it. Send us a screenshot of your openwrt wireless settings.

 

[mjc775]

Here’s a speed test comparison I did last year comparing 2.4 GHz vs 5 GHz WiFi on 400/20 Spectrum service using their Sagemcom router.

Sent from my iPhone using Tapatalk

 

[sfx2000]

Speedtest on a local hosted thing... from my Chromebook

 

[darth_adversor]

Hey guys, thanks for the additional responses. I appreciate it. So, nothing I've tried (I even tried the above recommendation to change the SSID and password) makes my speedtest go any faster than about 170 mbps. Wireless local network transfers continue to run in the 350-450 mbps range. I finally disabled my router, reset the AT&T gateway back to stock, and then configured the wifi on the gateway to more or less the same settings I'm using through OpenWRT. Runs perfectly, the wifi easily hits the ceiling imposed by the ISP.

I'd say it's my router (either the hardware or the firmware), EXCEPT...how can that be, when wireless transfers over the LAN run at full speed. I think sfx2000 was correct when he said there's magic sauce when it comes to using their gateway.

I think what I really need is a fiber modem/gateway that is capable of true bridge mode, rather than this "IP passthrough" stuff.