Faster Internet with Wifi vs. Wired?

Oct 31, 2018
3
0
10
Hi there, I am having an interesting issue with my home network that I would love to get everyone's thoughts on. In short, I have one computer on the network that appears to be getting a faster connection than everyone else and it's driving me bananas.

Here are the details:

Internet is Xfinity residential rated for 200mb down and 10mb up. It runs into a D-Link DCM-301 modem which then runs to a Netgear Nighthawk R7000 wireless router.

The computer that is getting the fast connection with wifi has an external usb wireless receiver (a Wsky Wireless USB WiFi Adapter - AC 600M Dual Band (2.4G/150Mbps+5G/433Mbps), USB WiFi Network Dongle Adapter). I have it hooked up to the 5G and it is getting 180MB down and the full 10 up (actually around 11).

And then for the wired devices upstairs I run a wire from the Nighthawk to a switch (TP-Link TL-SG1005D 5-Port Unmanaged Gigabit Desktop Switch) and then into (1) my work laptop via a port replicator (Lenovo X series), (2) a PS4 Pro and (3) an XBOX One X. The cable that runs from the Nighthawk to the switch is very long (100 ft) as I have to route it through a couple rooms. Cable is Cat 6 Ethernet Cable 100 ft Flat White from Jadaol. For all these upstairs devices I have done A/B testing to bypass the switch and go direct from the router and the speeds are similar. And when I unhook the work laptop from the port replicator and use wireless the speeds and the same and are still the same even if I put the laptop right next to the Netgear router! These devices all get around 45 down and 6 up.

For the wired devices downstairs I run from the Nighthawk to an Actiontec bonded MoCa 2.0 Ethernet to Coax adapter and then to the same model of TP-Link switch that I have upstairs). Devices on this are (1) the home desktop computer that seems to get a faster connection with wifi and (2) a second PS4 pro. These get the similar 45 down and 6 up including whether I bypass the switch or not. The PS4 pro gets similar speeds on wifi, as well (just not this crazy desktop)! The desktop is wired from the switch with the same Jadaol 100ft flat cable.

And then there we have a few mobile phones that are also on wireless (and get even lower speeds per speedtest.net - about 12 down and 10 up) and a Nintendo Switch.

So that's my quandary--I would say everything is running somewhat as expected other than the one desktop computer that has markedly better performance on wifi! And I would say this extends beyond speedtest.net results--when I simply A/B with wifi on or wired for that desktop and watch netflix, videos load almost instantly, but when wired they need to buffer a couple seconds before loading.

I should also note that the ping appears lower on the wired vs. wireless, as well.

And I should also note that my goal with all of this is to be wired because I do game on the desktop computer downstairs and the PS4 Pro upstairs, and in particular fighting games where packet loss is the enemy. So I'm not sure if for this desktop computer I should just be happy with the greater download speed or still utilize the wired connection for the, in theory, more consistent connection with fewer packet drops. Or does the markedly faster speed imply a better connection even if it's wireless?

My next step in testing would be to bypass the long ethernet cables to see if that is causing attenuation, or perhaps buy a beefier more shielded 100 ft cable since I really need that long for my setup (at least upstairs--I could probably go 50 ft downstairs).

Also, I should note I'm springing for business internet soon (300 down and 35 up), as well, and am wondering if maybe a fatter pipe in will render all of this largely moot, but we'll see.

Any thoughts you have would be greatly appreciated, and please let me know if you need any more information. Thanks very much!
 
Solution
Do you have another cable to test from? Try moving a pc next to the router. It seems like the 100ft cable is a weak link.

With ethernet you should be able to get your LAN to 940Mbs easily. Wifi around 200Mbs with 2x2 mimo AC.

Try going from modem to a PC as well. Are you running QoS or VPN from the router? or even trying to log? Services like these could slow it down.

iperf3 is a great program to test LAN speeds. Run a server from a wired client and then connect to that from wired or wireless. Windows fw might block inbound so turn it off the server temporarily unless you're on the modem. keep the fw up when testing the modem.
 
Oct 31, 2018
3
0
10


Thanks very much for the reply! I think next step will definitely be to check the shorter cable. I'll try a shorter cable of the same brand and type first and then maybe if that's still the same try a "higher quality" cable to see if that helps. I'll also look into trying to figure out how to use Iperf though I'm not sure I'll be able to use it with my work computer being one of the two computers--are there any tools that can test your internal network that only use one computer? And my work laptop upstairs runs a VPN but I'm not running anything with my own equipment--I didn't touch hardly any of the factory settings when I set the router up (probably just ssid and password) so I shouldn't be running any QoS or VPN through it though I may poke around the settings to make sure. Thanks again!
 


cat5e, cat6, cat6a are all rated for that distance. you don't have to buy a more expensive cable. You might just have a bad one or bad terminations. the speeds you get wired are very odd. it's normally 1000/100/10 or nothing if it's the wire. Have you tried direct to modem?
 
Solution
Oct 31, 2018
3
0
10


Hi there, I wanted to give everyone an update after getting business internet and rerunning the tests, and a few more tests. In general, the lower speeds to the work laptop computer can now be explained by the VPN that my work laptop runs through. When I run speedtests with and without it on it's widely different. My business internet is rated for 300 down and 30 up and I'm getting all of that without the VPN and I'm getting 45 - 55 down and 5 - 10 up with the VPN. But it seems like the VPN does perform better with the business internet, anyway--seems like it helps just having the fatter pipe in the background that will never be the bottleneck.

The downstairs computer is still something of a mystery. I'm getting 175 down and 30 up so that's pretty good but somewhat similar to what I was getting on wifi from before. And now it gets about the same on wifi but it's downstairs so that could explain the discrepancy there for the wifi and the moca adapter may be stalling things a bit on the wired connection.

Overall, though, the connection is great. It's been better for working and it definitely feels a bit more stable when playing fighting games online, too, and every little bit helps there.
 
VPN speeds could be the server you are connecting to.

It's very cpu intensive and it's single core. AES-NI is recommended. Any new intel or ryzen chip has it. VMs generally don't have it on their emulated cpu.
Openvpn 256b maxes around 500Mbs with very high end x86 single thread using AES-NI.