Best metric for "range"?

[skisteven1]

Hi Folks,

I'm new here, so be gentle. I looked around and didn't see anything that was directly applicable, so apologies if it's there and I missed it.

I live in an older building in NYC (ie. concrete, rebar, and lots of interference). I'm trying to get a signal one floor directly up. The AP from my cable company is visible upstairs, but not functional. I got some router from Best Buy that's OK, but not great.

Intended use: Planning to stream HD video to a apple TV or Roku, and maybe stream directly from another wifi device. Other than that just a couple of devices doing basic browsing. (ie. I'm not worried about how fast a file transfer will go, as long as it's fast enough for my HD streaming from device to device)

So my question is, what router will give me the best shot at getting a strong, consistent signal, with high throughput upstairs through the thick floor. I've been trying to find a way to compare routers "range" directly, but it doesn't seem that that's a real type of metric. Or at least, not one that's easily comparable with the tools on this site.

Currently it looks like the Asus RT66 might be the winner, but I'd love a cheaper option if it's anywhere near as likely to work for my circumstances.

Thoughts? Is there a thread on this I missed?

Thanks for the reco's!

 

[PrivateJoker]

IMHO there's not a good, universally applicable "range metric" because the variables of surrounding RF environment, building material influenced signal attenuation/reflection, and antenna placement & orientation (of both the router and the client devices) are so incredibly subjective in a case by case comparison.

You could have a top of the line router but if the floor you are trying to permeate with wifi is of a certain construction, there isn't a lot a router can do to overcome it. Also vertically oriented dipole antennas (picture the RT-N66's 3 antennas all in an upright position) tend to radiate in a pattern roughly similar to a 3D donut. Horizontal propagation is heavily biased over a vertically diverse signal. You could play around with antenna angles (by leaning them back) but there will always be trade offs and it's really impossible to make predictions on how stuff will work in a given environment. When you are on top of a router vertically it can sometimes have a less than graceful "sweet spot" depending on your angle to it (ie rather abrupt cut offs as you move closer or further away from it in a horizontal plane while the vertical distance is the same).

You might have very different signal strength on one band vs another at the same spot, or a strong signal when your laptop is facing east, but a crappy signal when it faces west, even if it's location in a 3D space is unchanged.

I'd say experiment with something bought from a place like Amazon w/ good return policies, and see what works for you. The N66 is a solid router, and I like mine a lot. But check the reviews on the main site, lots of good ones. Take repeated (and repeatable) measurements from different places and make a spreadsheet so you're not going just by feel. Here's one I made just to compare for my own reference how a change in antenna or firmware might affect me. http://bit.ly/PrivateJokerWifi

 

[stevech]

Please note that WiFi is a two-way medium. The signal from the client device to the WiFi router is not normally visible. A strong signal FROM the WiFi router does not at all imply that the reverse direction signal is anywhere near the same. The from-client signal strength limits the connection speed, as a rule.

Best metric: SpeedTest.net to a known-optimal nearby server. What matters is the speed after all overhead and error correction due to interference and weak signals. Using SpeedTest.net... be sure to test during the evening busy hour. Take many readings and average. If using WiFi, you may need to change channels if a neighbor streams HD video frequently.

 

[Fraoch]

Perhaps another approach might be not to even try to penetrate the ceiling/floor at all.

Can you run an Ethernet cable from one floor to the other? Most likely not, but would you consider powerline networking from one floor to the other?

If you can get some sort of wired networking from one floor to the other, keep your existing router and buy a low-cost one for use on the second floor, then convert it to an access point.

That way, instead of a super-powerful, very expensive router that may get an average signal upstairs, you have two lower-cost routers plus powerline networking (the total cost may be about the same) and get much better signal upstairs. Plus your wireless devices have a much closer AP and won't have to throw their much weaker signal through the thick floor/ceiling. Heck, you may even be able to get the wired network directly to your video playback devices, bypassing wireless entirely.

Something to think about at least!

 

[thiggins]

The metric we use for Range in the Router Ranker is throughput at a specific attenuation. Or in the old open-air method, throughput at a location where we know routers tend to have a low signal.

All you have to do is go to the Router Ranker, select the router class and select the Wireless Range Rank.

Or go to the Charts, select the 2.4 or 5 GHz up or downlink benchmarks and choose Location F for 2.4 and Location D for 5 GHz. Here is the 2.4 GHz downlink sort for Location F. Here is 5 GHz downlink, Location D.

Keep in mind that to get the full throughput shown, you need a client of the same class, i.e. if you are looking at AC1750 routers, you'd need an AC1750 client.

 

[skisteven1]

So then...

Re: wired networking. YES. I've definitely considered this. Running an ethernet cable seems like too much of a pain, but I could do powerline networking I suppose - I hadn't considered that particular option. Are those any good these days?

Re: Router Ranker / wireless range: I found that ranking, and it was quite useful, but I wasn't able to find a good way to compare the rankings via a metric. ie. Something along the lines of "This is really good, but the #3 option is 85% as good, and the #5 rank is actually only 40% as good". I couldn't find a static comparison like that (though it may not be feasible to produce one). Also, there's no good way to compare metrics like range across router classes.

@Stevech, PrivateJoker: Good points.

So, it sounds like the upshot is I should just buy something, try turning it in multiple directions, and test it from everywhere I'd want to use it before the return period is up. Yes?

 

[PrivateJoker]

@Stevech, PrivateJoker:[/B] Good points.

So, it sounds like the upshot is I should just buy something, try turning it in multiple directions, and test it from everywhere I'd want to use it before the return period is up. Yes?

Why not? The downside is nothing, you just say it didn't perform to expectations and print out a postage paid label. If you can swing it, maybe grab a pair of power line adapters at the same time. They could work great, they could sort of work ok - another thing that's hard to gauge until you actually install it.

Do you haze no connectivity needs on the floor where your router must be positioned? Or are there a few devices down there as well?

 

[Fraoch]

Re: wired networking. YES. I've definitely considered this. Running an ethernet cable seems like too much of a pain, but I could do powerline networking I suppose - I hadn't considered that particular option. Are those any good these days?

They're getting better - they've gone through several generations now.

http://www.smallnetbuilder.com/lanwan/powerline-charts/view

It is highly dependent on wiring but it may be worth a try.

 

[thiggins]

Re: Router Ranker / wireless range: I found that ranking, and it was quite useful, but I wasn't able to find a good way to compare the rankings via a metric. ie. Something along the lines of "This is really good, but the #3 option is 85% as good, and the #5 rank is actually only 40% as good". I couldn't find a static comparison like that (though it may not be feasible to produce one). Also, there's no good way to compare metrics like range across router classes.

Use the Router Charts, wireless benchmarks, as I indicated in my first post. They work across classes and let you see the exact throughput values used for the ranking.