5Ghz Speed Tests Giving Inaccurate Download Speeds

[Guest]

A few months ago, I obtained a new router and modem. One thing in particular that had my interest with this router and modem is the support for 5Ghz connections. Although, I noticed something very strange upon testing the speeds to ensure everything was in working order after installation. According to the test, and every test I've ran since, 5Ghz connections and direct connections report a download speed between 110-140mbps. This can't be right, because applications used to show my speeds in real time while browsing on all my devices never showed those speeds exceeding 40-45mbps downloads. Such is to be expected, as the package we're getting through our ISP is one for 50mbps download speeds.

I never really thought anything of it, until I started to have issues with the connection I was lead to believe could be my ISP throttling the connections. Over the past three weeks, speeds from my bedroom, which sits no more than 20 feet from the modem, have gone from their initial drop to 35-40mpbs which is to be expected with the distance from the modem, down to as low as 3mbps! When I tried to resolve this issue with my ISP, they had told me it appears that I'm getting roughly 110mbps connections, so they had no idea why I was complaining about low speeds. At that moment, I realized that the speeds are going to be showing as being way faster than they actually are, because for whatever reason, tests on direct connections and 5Ghz wireless connections show ridiculously inaccurate download speed results.

Since what little I know about networking was a crash course back when I was looking for a replacement modem, I have no idea why this is happening, and figured someone here would probably understand the problem far better than I would. Googling the answer also failed me. (Google seemed to think I wanted to know about the differences between 5Ghz and 2.4Ghz connections when I searched for the solution.)

Now that you know the problem, here's the information one of the pinned posts suggested I post:

Modem: Motorola Surfboard 200 series model SB 6121
Router: Linksys AC 1200 model EA6350

PC Specs (Posting the parts they ask for specifically, because I wasn't sure how redundant a full list of specs would be. I will post everything if needed): ASRock H97m Anniversary mATX motherboard, Thermaltake TR2 600w PSU, TP Link TL-WDN3800 N600 PCIe wireless adapter, running Zorin OS 9 Core 64-bit (Linux OS heavily based on Ubuntu 14.04)

ISP and Connection Type: Comcast/Xfinity broadband

Known Connected Devices: 1 Desktop (mentioned above), 2 laptops, 6 phones, 3 tablets, 1 Apple TV, 1 Roku stick, and 2 Chromecast sticks.*

*This is the total amount of devices I know of that can be connected to the internet at any time. At least half of these devices are not regularly connected and all speed tests are only conducted at times when everyone else who would be on their devices is either not home, or I can confirm that they are not actively using the device.

 

[getochkn]

Connect a PC directly to the modem and run a test from speedtest.net then connect directly to the router with an ethernet cord and test again. Then test from a WIFI device and do speedtest again.

Post all 3 results here.

That will let you know you direct speed from the ISP to you with the modem, then through the router and see if there is a big difference.

WIFI speeds at what they are "advertised" at, and what you get vary a lot by distance, by other WIFI routers in your area, lots of things.

Try on android phone to use an app called wifi analyzer. It will show you charts that show what other wifi networks in your area, the channels they are on, the best channel for your wifi, etc.

 

[bigpinkdragon286]

Just because you don't believe the numbers, doesn't mean they are inaccurate. Have you run multiple independent speed tests from different sites?

It's possible you're on one of 'those' ISPs that pulls shenanigans and lets speed tests go through an unrestricted or less restricted path to your location.

Perhaps, the site you're attempting to download from is the limiting factor? Are you using a service such as BitTorrent that can vary greatly in it's transfer speeds?

Does your ISP support a burst speed, such as Comcast, which tends to be double for the initial part of a download?

Are there other users in the house sharing your bandwidth when you're downloading?

Are there low speed radios on your network when downloading?

 

[Guest]

Just because you don't believe the numbers, doesn't mean they are inaccurate. Have you run multiple independent speed tests from different sites?

It's possible you're on one of 'those' ISPs that pulls shenanigans and lets speed tests go through an unrestricted or less restricted path to your location.

Perhaps, the site you're attempting to download from is the limiting factor? Are you using a service such as BitTorrent that can vary greatly in it's transfer speeds?

Does your ISP support a burst speed, such as Comcast, which tends to be double for the initial part of a download?

Are there other users in the house sharing your bandwidth when you're downloading?

Are there low speed radios on your network when downloading?

It's not that I don't "believe" the numbers. It's that the nunbers conflict what is showing on my system monitor(Linux equivallent of Task Manager) in real time as I browse, and what an app on my mobile devices meant to monitor the speeds in real time are showi g.

As for the test themselves, I ran tests from Speedtest.net, Xfinity's speed testing web site, and Speakeasy.net. all of the speed test results were withing a megabit of each other.

I don't torrent anything, and no one on my network torrents anything (to the best of my knowledge).

Being that my ISP is comcast, it's entirely possible this is their doing. That's why I'm here. I want to ensure as accurate a test as possible next time they try to claim I'm somehow getting double what I'm paying for when I'm actually getting as low as 1/15 what I'm paying for.

As for the bit about other users, there are. It's a decent sized house with several people living with me. That's precisely why I noted that I only test the speeds at times when I'm sure I'm the only one using the internet.

Anything I didn't answer was something I didn't really know the answer to.

 

[bigpinkdragon286]

I would start from the standpoint that, if multiple speed tests are giving results within a margin of error, they are reporting correct figures.

 

[getochkn]

Speedtest's over wifi don't give any measure of your internet, too many variables, that why I suggested what I did in my post above.

 

[bigpinkdragon286]

Speed tests over Wi-Fi are fine, so long as you understand what the numbers are telling you. Totally agree however, for a test of the speed to the home, you really need to cut out the variables by going as straight to the source as possible. Even just using an underperforming laptop can torpedo accurate speed tests.

Since you're on wireless, and you say you're 20 feet away from the router, I suggest switching from the 5Ghz to 2.4Ghz band. Drop off with 5Ghz tends to be quite a bit more severe with 5Ghz and penetration is even worse. You can easily see higher throughput using a more crowded 2.4Ghz band, if you're going through obstacles.

 

[Solandri]

According to the test, and every test I've ran since, 5Ghz connections and direct connections report a download speed between 110-140mbps. This can't be right, because applications used to show my speeds in real time while browsing on all my devices never showed those speeds exceeding 40-45mbps downloads.

This is quite common. There are two reasons for it.

1) The vast majority of websites out there are on connections slower than 100 Mbps, or won't give a single connection that much bandwidth so as not to degrade service to other customers. So you will never see 100+ Mbps from them. Steam and Microsoft are a couple which seem to have gobs of bandwidth and don't limit you. I regularly get my full Internet speed (110 Mbps) downloading from them.

2) Some ISPs cheat and give your transmissions priority whenever you're hitting a speedtest website. But put you in the congested slow lane when doing other things.

(1) is much more common than (2).

I never really thought anything of it, until I started to have issues with the connection I was lead to believe could be my ISP throttling the connections. Over the past three weeks, speeds from my bedroom, which sits no more than 20 feet from the modem, have gone from their initial drop to 35-40mpbs which is to be expected with the distance from the modem, down to as low as 3mbps!

Pull an ethernet cable from your modem to your bedroom and test the speeds. If that is also dropping to 3 Mbps, then it's your ISP's fault. If it gets the full speed, then the slowness is being caused by your wifi.

Known Connected Devices: 1 Desktop (mentioned above), 2 laptops, 6 phones, 3 tablets, 1 Apple TV, 1 Roku stick, and 2 Chromecast sticks.*

Whoa, that's a lot of devices. Bear in mind that unless you're using 802.11ac in MIMO mode (likely only your router and newer laptops support this), wireless bandwidth is shared. 802.11n will give you about 110 Mbps of actual bandwidth, 802.11g about 20 Mbps of actual bandwidth. If a device is using some of that bandwidth, other devices will have reduced bandwiidth.

Also understand that neighbors' wifi networks and things like microwave ovens will interfere with your wifi signal. The 2.4 GHz band in particular is very crowded. If you're experiencing mysterious intermittent drops in speed, then go into your router's settings and try changing the wireless channels. The "Auto" channel select tends to fail on moderately sparse networks in medium to large houses. It's too far from a neighbor's wifi to sense it and thus puts it on the same channel as the neighbor, while your device may actually be sitting closer to your neighbor's wifi than to yours. For 2.4 GHz try manually setting it to channel 1 or 11 (6 is the default, and most crowded).

The 5 GHz band is lot better off with many independent channels. I hardly ever encounter interference at 5 GHz. Its drawback is that the signal strength attenuates more quickly with distance and through walls, so its range is shorter.