Bench Tests

Bridge Engine Testing
Most of our testing centered around testing the ADSL chipset in the modems. We made an effort in the above tests to avoid coloring the data by stressing the bridge/router engine by choosing large packet sizes (1518 bytes). In this test however, we were interested in looking at the performance of the bridge/router engine in the ADSL modems. The modems were all tested on the bench with a short (10 foot) wire connecting the dsl ports. As with the other tests, the modems were connected to Fluke xDSL One Touches at each end. This test however used custom software from Fluke that performed a matrix test of 10 upstream speeds, 10 downstream speeds, and packet sizes from 64 to 1518 bytes. These tests take several hours to run, and aferwords the data is harvested from the One Touch into a spreadsheet.

We will only be showing two representative examples here. The first example is the Westell FlexCap2, which has a nearly perfect bridge engine.

This shows the Westell FlexCap2 working without showing any degredation in service. Notice that this test is being done at 128 bytes/packet, which is the next to smallest packet size.

This is the same test as above, butat 64 bytes/packet. Here we have hit the maximum packets per second that the Westell unit can bridge as is seen by the rolloff at the higher rates.

The Amati's performance is a bit more erratic in this test, but generally very good at larger packet sizes

When the amati is driven at 64 byte packet sizes, it's data rate drops sigificantly as is seen in this test.

Downstream Impairment Test
During our testing, we noticed that some of the modems were very sensitive to the chosen upstream speed. In particular, we found that if we overdrove the upstream channel beyond what it had negotiated, there was a significant drop in performance of the downstream rate. This test was designed to define this behavior. We again used the Fluke xDSL One Touch as described above. We fixed the downstream rate at the maximum negotiated rate. The upstream rate was driven below, at, and above the negotiated speed and throughput was recorded. Two modems in our trial exhibited this problem: Aware and Amati. As a comparison, the PairGain DMT prototype was also included in this graph to show the performance of a modem that lacks this problem.

Dave Lacey and Mike Lutz
To contact us check out the contacts page.