Not getting full bandwith of internet connection on Wireless N

[stevech]

Just because something is 300Mbps or whatever, doesn't mean it is going to be remotely as fast as another device that is 300Mbps.

The WiFi connection speed, such as 300Mbps as discussed here, is an indicator of the instantaneous signal strength.

The REAL measure is an IP layer throughput test that moves a large quantity of data to- and from-client.
In a perfect world, the (consumer affordable) test tool would show the percentage of errored frames, delays due to channel-is-busy, etc.

A WiFi device with a distorted transmitted signal (e.g., due to overdriving the power amp), will have lower throughput due to errored frames. An "amped" device often has this, plus a receiver that's seeing added noise from the RX amp and thus gets a net decline in throughput.

stronger and quality signal? No substitute for better antennas. But don't mess with an improvement in antennas less than 6-10dBi. Even that is a small fraction of the total attenuation in a typical inter-room situation.

 

[azazel1024]

Very true, though I semi-sort of disagree with the antenna gain.

True, 1-2dB probably isn't going to make any noticable difference at all, but even just 3-4dB might.

It might not be a big difference, but that could still be the difference between connecting at 11Mbps in a far off room or 54Mbps (though not likely that dramatic a difference).

At least looking at inSSIDer, there are plenty of locations in my house that are hovering between a good and an excellent RSSI.

It isn't likely to let you connect a room further away where you couldn't connect at all before, but it could still bump your speed up a little at further locations.

A 6-9dB difference might well mean being able to connect in a location you couldn't before though.

Amusingly in testing out a router in my garage to be used as an outdoor AP (router in the garage, antennas mounted with 3ft coax extension wires through holes in the top of the garage wall under the roof overhang), unplugging the antennas still gives pretty good through put right by the router.

Walking outside the garage it is abysmal though.

-54dB no antennas attached. -25dB, antennas attached. So the signal is about 800x stronger with the antennas attached, but it is remarkable just how good wireless gear is at filtering noise and picking up signals that it makes almost no difference in the speed between -25dB and -54dB (it is about -68dB 3ft away with the exterior wall between the router and my tablet). Of course the story there is different with the antennas outside and mostly LoS to everywhere I care to connection, but I am thinking of swapping the 5dB antennas for either 7 or probably 9dB antennas just to up the signal a little. It hovers around -60dB where I want it and I'd really prefer to see it closer to the mid 50's if I can.

 

[stevech]

A high gain omni antenna in urban areas has this not well know disadvantage: Increased noise and interference from neighbors!

A panel antenna as in outdoor bridges, or a yagi/dish, mitigates a lot of this.

An "amp'd" WiFi product that has only a transmitting amp (HPA) is not suitable - because it can improve range in only one direction! The complex waveforms in 802.11n and 11ac require an extremely linear HPA. If not done, the HPA distorts the transmitted signal and the throughput falls due to more error correction overhead. Most Asian vendors don't know this "Rho" issue. The WiFi alliance (not FCC) has specs on Rho, but testing houses don't check it in the type certification for Part 15 regulations. And the WiFi alliance lacks funds to police this. (A distorted transmitted signal is not an issue with the FCC.)

Products with a receiver amp (LNA) add lots of noise to the receiver and that hurts the signal-to-noise. Nothing's free! A low noise LNA with a really good "noise figure" is too costly for consumer gear. They're used in cellular.

As we see more streaming broadband in WiFi (video) this all gets slowly worse in urban areas.

"Wireless is not a thousand times harder than Wired; It's a Million times harder!" Prof. Paulraj, Stanford Univ.

 

[azazel1024]

I have the advantage of not living in an urban error. I have the second smallest property in my neighborhood at 1.01 acres. Most are 1.5-5 acres. In my backyard, I can pick up 5 other networks nearby. The strongest signal strength by my access point is -78dB from my immediate neighbor on channel 1 and I am operating this on channel 11, which has 2 SSIDs I can pick up on that channel, at -88 and -91dB.

From where I really want to get signal, the strongest SSID is on channel 11, where I operating, but at -78dB, with mine coming through at -60dB right now. So another 2-4dB would not be amiss. Just a question of talking my wife in to 7dB antennas hanging down from the roof overhang, or 9dB antennas (no way to talk her in to 12dB antennas. 14" is probably already pushing it for 9dB antennas).

I do agree with everything else you are saying though.

 

[stevech]

Preachy: The number of and signal strength of detected neighbors' WiFi SSIDs is not important.

What is important is which of those SSIDs, if any, use a lot of air time (for streaming video, etc). And do so for long durations. There's no free tool I know of that measures that. And usage habits on those SSIDs change over time.

Best suggestion I have is use freeware pingplotter, let it run a plot to your router's gateway address, e.g., 192.168.1.1, for the evening hours, weekday, saturday, etc. On a certain WiFi channel. If the channel has a lot of traffic, you'll see increased ping times due to channel-busy delays in WiFi. Repeat for other suspect SSIDs/channels.

 

[azazel1024]

Thanks for the suggestion!

I'll probably give it a whirl at some point when I am fine tuning things.

Though, I only expect to be using my current setup for a year or maybe two. Then I plan on being able to have the time and money to invest in a better access point/router for outside and switch over to 5GHz for most things, which is completely barren, even poking around outside.

PS Though, to a degree, the number/RSSI is a bit important. I agree on how much usage they actually get, but if the RSSI is dramatically below my own and also if it is pushing the limits of sensititivity, no matter how much it is used, it shouldn't impact my own. I don't know real cut-offs, but at least in as much real world testing and experience as I've been able to get, I've never seen an impact of cochannel SSIDs which were at less than -80dB, at least not when my own RSSI was pushing >-70dB. I've done some limited testing with my own routers and access points by setting one to its own SSID and a cochannel and doing a big, long directory copy over it and seeing how it impacted things on my other router. It was only the one test, but the RSSI on the "interfering" router was -76dB from where I was sitting and it was -42dB from my test router and I saw no impact. In another test, I moved further away and the interferring router was at -81dB and my test router was at -56dB and I saw no impact, versus the interferring router not being active.

From my experience in my townhouse a couple of years ago with attempting to figure out the best channel to use (there wasn't one), it only seemed like once other SSIDs were pushing the -70dB or so range or greater OR were within about 20dB of my own RSSI that I saw bandwidth impacts during parts of the day.

 

[stevech]

From my experience in my townhouse a couple of years ago with attempting to figure out the best channel to use (there wasn't one), it only seemed like once other SSIDs were pushing the -70dB or so range or greater OR were within about 20dB of my own RSSI that I saw bandwidth impacts during parts of the day.

Depends on how heavily used a certain SSID is.
If a WiFi router or AP has zero users ever active, it will sit there sending the IEEE standard "I am here" beacons which is how survey tools detect the SSIDs. The beacons are normally 10 per second. Each beacon transmission is just a few 10's of microseconds. Miniscule transmitter on vs. off ratio.
That's why the number of/strength of SSIDs detected isn't itself meaningful, without the utilization measurement which isn't done in freeware.