empirical studies of wifi antenna performance?

[pr0metheus]

as a man heavily steeped in the scientific method, i was wondering -- is anyone here aware of any groups that undertake performance tests of different wifi antennas and related geek paraphenalia?

i'd like to get a wifi antenna, but i really would like to compare the performance of different ones to establish the relationship between purported specifications and actual, real world performance. if the testing is standardised, then comparisons over time would still be valid.

on that note, let's say i was to do this myself, what would be the best metric for performance? i'm presuming something like speedtest.net, but perhaps there's an alternative that could be useful? and please... don't say number of bars

[ki2k]

Hey pr0metheus,

Antennas (other than homemade) are certified by their manufacturers as they are built to specific gain and coverage pattern values. If you are home-brewing, that's different. But you need to learn basic RF math, learn about modeling programs like EZNEK and have the ability to measure signal strength (data rates, as measured with speedtest.net, are waaaay after the fact and are usually limited by your ISP connection, which tends to be a fraction of what what modern wireless is capable of and influenced by many other factors) at the output of the transmitter, at the output of the antenna, and out in the field to "prove" performance.

Or you could use like yogi project plans and free online calculators to get a sense of what a given antenna design should do if built right.

It is fun stuff.

[pr0metheus]

nonono! that's not my point -- in fact, i think it's the antithesis of the point i'm trying to make.

at the end of the day, we're looking to get a wifi antenna so as to improve our wifi signal e.g. faster internet. even though we may choose the antenna with the highest dBi as a metric in improving our internet performance, antenna strength isn't isomorphic with download speed.

to make my point by analogy -- choosing our antenna based on dBi might make as much sense as basing our choice in basketball draft on a player's height. sure, height is strongly correlated with performance, but it isn't the same thing. as a corollary, it may be better to look at the number of points they score per game as a measure of performance. again, points per game is an attempt to operationalise the concept of "performance" but it isn't the same thing as performance.

likewise, what i wanted to know is if there is a group or resource that assesses the quality of a wifi antenna using a more tangible metric of performance e.g. download speed as measured on speedtest.net.

It is fun stuff.

diff'rent strokes n all

[GuardMoony]

Like you said yourself. Antenna are more or less to increase the range. For performance increase you need to change the transmitter ea your router/Wireless card.

The test the performance part you probaly gone need to measure. Packet/rate, loss/drop rate, delay, maybe noise ratio. But most of that stuff need special software / hardware

[ki2k]

There are no real groups that do what you describe, because you're missing a significant point. Antennas have little to do with throughput and speedtest websites are poor measures of radio performance. The increases in performance you imagine testing in relation to antennas come from 802.11 technologies in play, adapter capabilities, and a number of other parameters that have nothing to do with antanna. The best you'd be able to do is study sustaining higher data rates allowed by radio technology in use at further distances, but these rates are reported by the adapters themselves.

[Hibby]

Like you said yourself. Antenna are more or less to increase the range. For performance increase you need to change the transmitter ea your router/Wireless card.

The test the performance part you probaly gone need to measure. Packet/rate, loss/drop rate, delay, maybe noise ratio. But most of that stuff need special software / hardware

There are no real groups that do what you describe, because you're missing a significant point. Antennas have little to do with throughput and speedtest websites are poor measures of radio performance. The increases in performance you imagine testing in relation to antennas come from 802.11 technologies in play, adapter capabilities, and a number of other parameters that have nothing to do with antanna. The best you'd be able to do is study sustaining higher data rates allowed by radio technology in use at further distances, but these rates are reported by the adapters themselves.

Further to this, antennae don't exclusively increase the range of your signal, specifically what they do is focus signal akin to a reflector behind a bulb in a torch. Some types, referred to as omnis, will give you a more toroidial pattern, like a squished doghnut, as their gain (or, more accurately, directivity) increases.

Others, known as directional, will focus the beam as the reflector mentioned earlier - yagis fall into this class... They won't increase your throughput or magically give you more power (don't confuse antenna gain for amplifier gain), but they will increase your range in a single direction.

As the others have said - this range increase is manifested in a higher throughput at greater distances, yes, but it doesn't increase the maximum throughput and, vitally, it decreases performance in all other directions apart from the beam direction. The trade off for a high gain omni is vertical performance - you get a large, flat, horizontally wide beam pattern but poor vertical performance.

For measuring your RF performance, your adapter will offer you a dBm output figure - that's decibels referenced to 1mW. The closer to positive dBm (negative dBm is a fraction of 1mW -> -3dBm == 0.5mW, positive dBm is a multiple 1mW -> 3dBm == 2mW) the better your performance. This is, however, influenced by your receiver - specifically it's antenna's characteristics and receive sensitivity, so make sure you use the same one with the same physical orientation (90 degrees from horizontal/etc). So it's a bit like bars, but with numbers attached :).

To get into the proper RF analysis tools that I use in work, you're looking at thousands of USD/GBP/Local currency - it's not cheap to properly map an RF field, and not 100% accurate given the sheer number of variables. Proper systems and antennas are analysed in an Anechoic chamber, with the RF nearfield analysed with hundreds of sensors. The power is then turned down and done again to map far-field, with some clever maths done to extrapolate the full field as close as is possible. It's a good fun process that one!

In short, antenna choice depends on the application and the location - both an omnidirectional or a directional will have advantages, but these come at a cost - there's a tradeoff to consider which varies between situations.

So... Tell me more about how you want to increase your range...