Performance
We didn’t have a router (or an access point) based on the Top Dog chipset at the time of our tests, so we could only see how Marvell’s chipset behaved with other developers’ chipsets. To be specific, we tested the D-Link DWA-142 with routers based on chipsets from Atheros, Broadcom and Ralink. These routers have already been tested in our labs, and we know their capabilities well enough. And we’ll check out the D-Link DWA-142 with a router based on the Marvell chipset as soon as we have such an opportunity.
We used the following hardware and software for the tests:
- Category 5e Ethernet cable
- DWA-142 USB-adapter
- Linksys WRT300N v2 router (Atheros)
- ASUS WL-500W router (Broadcom)
- Edimax BR-6504n (Ralink)
- Centrino notebook
- PC with a mainboard that has an integrated Gigabit Ethernet controller
- IxChariot (High_Performance_Throughput scenario)
First we tested the speed of router-adapter (LAN-WLAN) and adapter-router (WLAN-LAN) connections using the High_Performance_Throughput scenario of the IxChariot program. The connection was established near the routers with WPA2-PSK encryption using the AES algorithm. This is the standard encryption for Draft N, and it is going to be such in the official 802.11n specification. We used the latest driver for the adapter and the latest firmware for the routers.
Below is the diagram for the TEW-624UB USB-adapter. It is not a comparison of the two devices (that’s why we publish two different diagrams) but rather an illustration of how the problem of compatibility between Wi-Fi chipsets from different brands is being steadily solved.
So, there are both positive and negative points. For example, the ASUS WL-500W defaulted from the test. The DWA-142 could establish a connection with it, but the connection didn’t work – the IxChariot test would abort with an error message. The connection status window would report a speed of 270Mbps at that. We can see the Atheros chipset is the leader in terms of performance after the driver and firmware have been improved. Currently, this chipset seems to deliver the highest quality, we guess.
Next we measured the WLAN coverage area using almost all possible combinations of routers (without the ASUS one) and wireless adapters. We measured the signal level of a LAN-WLAN connection in five points:
Point 1: Near the router
Point 2: At a distance of 4 meters without obstacles
Point 3: At a distance of 5 meters + two thin gypsum wallboards
Point 4: At a distance of 6 meters + one brick wall, about 30cm thick
Point 5: At a distance of 17 meters + one thin gypsum wallboard and one 50cm brick wall
Now we can see that the compatibility issues with Draft N chipsets are being solved, even though slowly. It would be better if we had lower but identical results for all the routers. With the current state of things, you have to be careful when choosing equipment for new-generation networks. As for the particular USB adapter, the DWA-142 is quite worthy of its price even though it does not deliver highest performance possible.
