How Fast Is It? AirPort Extreme 'N' Base Station, Part II--5GHz and 'b/g' Compatibility Modes

After posting our original review a week ago last Monday (Part I--"n-only" at 2.4GHz), we received a number of questions and requests for additional tests on Apple's new AirPort Extreme 802.11n Base Station. Some asked why we used 802.11n(2.4GHz) when the 5GHz mode with wide channel capability is supposed to be faster. Others asked how fast the new Base Station would be in the mode that is compatible older 802.11b and 802.11g devices with and without any being on the network. So we completed several additional runs. We also ran the tests with the network protected by WPA2 security and without security since some testing elsewhere indicated that, perhaps, WPA2 security significantly slowed throughput.

Instead of the MacBook Pro Core 2 Duo used in Part I, we used our 24" iMac Core 2 Duo which is on the second floor of our building and about 12' from our new AirPort Extreme N Base Station located in the room below on the first floor. We tested the AirPort Extreme N Base Station's speed by using our iMac C2D via the Extreme N Base Station to read from and write to an AirPort Disk attached to the Extreme N Base Station using Intech's QuickBench, a utility in Intech's SpeedTools software package. The Extreme N Base Station is set up as a bridge from our router, i.e. not using a second layer of NAT.

Here are the somewhat surprising results:

Speed (Sustained Read/Write in MBps) via AirPort Extreme N Base Station in various wireless modes*
Wireless Mode
WPA2 On
WPA2 Off
n-only 2.4GHz
4.4/2.0*
4.4/1.9
n-only 5GHz
4.2/1.9
4.2/1.9
n/b/g
4.4/2.0
4.6/2.1
n/b/g w/"g" Device
3.9/1.9
4.4/2.1
*Speed in MegaBytes per second.
Average of three tests--Extended Test mean of 50-100MB transfers

When the AirPort Extreme N Base Station was set up in "n-only" mode, the 5GHz mode with "Use wide channels" enabled was slightly slower (about 5% or less) than the 2.4GHz mode. This is exactly the opposite from what we had expected and we have no explanation for it. Turning off WPA2 security essentially had no effect, again not what we had expected. These results were confirmed with our MacBook Pro Core 2 Duo--it got even poorer 5GHz performance at 4.1MB/sec sustained read. We also disconnected our 5.8GHz telephone system in case it was interfering but doing so had no effect on the "n-only" 5GHz speed results.

When the Extreme N Base Station was set up to be compatible with older "b" and "g" Wi-Fi devices ("n/b/g" mode) with WPA2 security on, the speed was identical to the "n-only" 2.4GHz mode. This result was expected because no "b" or "g" devices were attached to the network during that test. With WPA2 security turned off, we got the expected increase in network speed, even though at about 5 percent the increase was small.

With a "g" device (a 12" PowerBook G4 with an AirPort Extreme G card) added wirelessly to the network in "n/b/g" mode, network speed slowed down as expected with a performance drop in sustained read of over 11 percent. With WPA2 security turned off, the network speed increased by about 10 percent but was still slower than without the PowerBook G4 attached. We do not recommend setting up a Wi-Fi network without using WPA2 security.

In summary, the performance of Apple's new AirPort Extreme N Base Station in 5Ghz mode with wide channels enabled in our tests did not meet the expectation of higher speeds with only "n" deices on the network. In fact, it tested slower than using the 2.4GHz band. Perhaps, Apple has not yet fully enabled the 5GHz part of the draft 802.11n specification. In fully compatible mode (n/b/g), there was no loss of performance over "n-only" mode at 2.4GHz with only "n" devices on the network. When a "g" device was added, the performance drop was on the order of ten percent.

Subsequent to our 2/12/07 review, other speed tests of Apple's new AirPort Extreme N Base Station have appeared on the Web with somewhat different results (e.g. Macworld and Bare Feats). Unfortunately, none of them have posted sufficient details to be able to replicate them to determine why there might be differences, usually caused by procedural/setting differences .

See Part I--"n-only" at 2.4GHz. [Bill Fox]

Saw something? Send a tip

The archive ran on reader tips. What did you see, where, and do you want the credit?

Read by the editor. Never published without your say.

More in Reviews · This month in the archive