Last month we reported the result of our first in-home tests of HomePlug powerline networking, and concluded that HomePlug could play a major role in consumer networking. We reported that we obtained substantially identical performance results from the three sets of Ethernet adapters and from the one USB adapter we tested. Although we had not completed our tests, we expected that the results from our initial tests would be representative of other products based on the same Intellon chip and that products based on the new Cogency chip would provide comparable results.
We have now completed our tests on two USB adapters made by ST&T xNetworks -- the U21 based on the same Intellon A1 chip as the units we reported on last month, and the U22 based on the new Cogency chip.
To our surprise, both ST&T units gave significantly better results than the units we tested before.
The latest tests were all run using an ST&T M51 Ethernet bridge as the "master" adapter and the U21 and U22 as "mobile" adapters. (Please see last month's issue or our web site http://www.BBHcentral.com/bbhl/homeplugprocedure.html for a description of our test procedure and the "master" and "mobile" adapters.)
We talked with several technical people to better understand the performance differences we measured and concluded that while the digital chip has a lot to do with HomePlug performance, other parts of the design can make a big difference. We were told that the design and specific components in the "analog front end" - which sits between the digital chip and the wall outlet - can affect performance, as can the design of the power supply and the layout of the circuit board. ST&T seems to have done this better than the other vendors.
In the previous issue, we reported that the ST&T Ethernet bridge performed identically with the other tested Ethernet bridges. But we had tested the ST&T unit with another vendor's bridge; ST&T says that if had tested it with a second ST&T Ethernet unit (which we didn't have) we would have gotten results similar to those we got with the two ST&T USB adapters.
We did not see any incompatibility between ST&T units based on two different chips. This is a good sign for HomePlug compatibility.
Our evaluations should not be regarded as definitive since they were performed only in one house, which we cannot say is representative of others. As more HomePlug units come to market, we'll be interested in seeing more thorough evaluations to advise consumers of the performance differences between identically-labelled products.
We have updated our website with additional information about the details of our test results: http://www.bbhcentral.com/bbhl/homeplug.html.