Those of you who aren't fans of the Azimuth-based test procedure I use for testing WLAN products will be happy to see that I used a walk-around type of test for the AP 200. This is mainly due to the nature of the product, and the fact that it doesn't have an bridged Ethernet interface.
The approach I took was to set up my trusty Linksys WRT54G V1 in my downstairs office and a test notebook with a Linksys WPC54G V1 card. I first took baseline throughput measurements using IxChariot with the WRT54G, AP 200 and notebook in the same room. Those results are shown in Figure 11.
Figure 11: Baseline Throughput results
Up and downlink tests were run separately with the test notebook directly connected to the WRT54G (tests with wrt54g prefix) and with it connected via the AP 200 (tests with pepwave_surf_ap_200 prefix). You can that, due to the single radio doing double duty, throughput via the AP 200 is about 60% lower than when directly connected to the WRT54G.
I also checked throughput with WEP 128 and WPA2 enabled on the AP (the one the notebook was connected to) and measured no appreciable throughput loss.
I next moved the notebook to Test Location 4 ("the black hole of Wi-Fi") located in my kitchen. This location is one floor up and at the diagonal opposite end of my home. Check the old test method description for more details about the distances involved.
Finally, I moved the AP 200 to the five different locations shown in Figure 12 and reran the throughput tests in the downlink direction. (I spot-checked uplink performance and didn't find much difference.)
The test rationale was that the throughput delivered to the notebook depends on both the throughput from the WRT54G to the AP 200 and from the AP 200 to the notebook. And that is going to depend on how each of the three radios is affected by the RF obstacles in my home's construction.
One thing that the radios didn't have to contend with were strong neighboring wireless LANs. Even though the AP 200 picked up my closest neighbors' network (which is at least 500 ft. away through plenty of trees (leafless at this point), the signal level was very low.
Figure 12: Test Locations
The results, summarized in Table 1, show that you'll need to experiment to determine the right place to park the AP 200. They also show that the best place might not necessarily be the place you think it will be! The Table also includes the Baseline measurement made in my office via the AP 200 and a Direct measurement made with the notebook located in the Kitchen connecting directly to the WRT54G.
For example, Location 5, which was one floor below the test notebook, delivered very good throughput, even though there were plenty of sheetrock walls between it and the WRT54G.
|Location||Average Downlink Throughput
1 minute average (Mbps)
Table 1: Throughput Test Summary
The other notable example is Location 2. I thought that moving the AP 200 closer to the notebook would raise throughput from that of Location 1. But I found the opposite effect, perhaps due to the tile backsplash that was directly between the AP 200 and notebook in Location 1.
The other lesson learned is that it helps to think three-dimensionally when determining the best AP 200 location and to keep visualizing the direct line-of-sight paths between the three pieces to the puzzle.