Zenith Demos VSB at Consumer Electronics Show

By Tom Butts

Hoping to present "the other side of the story" in the ongoing COFDM/VSB debate, Zenith officials conducted side-by-side comparison demos of the two modulation formats at the Consumer Electronics Show recently.

Due to the fact that nobody in Las Vegas was transmitting an 8MHz digital signal, Zenith had to conduct the demos in a controlled lab-type set-up at a Las Vegas hotel room. Company officials noted that they have been unsuccessful at obtaining 6MHz COFDM equipment—which was used by Sinclair during the public test last summer in Baltimore and are currently being used for the Brazil DTV trials.

Zenith used two current generation COFDM-based DVB receivers—a Nokia consumer decoder and an NDS professional unit. VSB units included a first generation Panasonic HDTV set top box (the same model used by Sinclair in its Baltimore tests), a second generation Zenith set, and a prototype third generation HDTV box using Motorola's new MCT2100 VSB demodulator, which Motorola says solves the multipath problem for VSB reception (see Motorola Claims New Chip Will End the DTV Modulation Debate).

The tests were conducted in three areas:

• Straight coverage—measured by the distance from the transmitter before signal decreases—the so-called "cliff effect";
  
• Multipath, where the signal was delayed one microsecond and an ensemble of five ghosts; and
  
• Impulse noise, measuring the effect household appliances (or other source of impulse noise such as ignition noise, power lines, etc.) would have on the signal.

For straight coverage, rather than lower the signal power, Zenith engineers raised the noise level in the signal for comparison. The Nokia COFDM receiver began breaking up at 21 dB signal to noise ratio, and at 20 dB, the NDS receiver broke up, while the first generation VSB receiver did not break up until 18 dB. By 15 dB, all COFDM receivers had completely cut out, as well as the first generation VSB receiver with picture freeze on the prototype Motorola VSB decoder. The second generation Zenith worked fine at 15 dB.

In the multipath tests, engineers simulated a ghost effect, reducing the effect from moderate to lower levels by raising the white noise in the signal. A 50% ghost, with 0.1% noise, generated occasional errors on the first generation VSB receiver. With 0.3% noise, the COFDM receivers continued to operate and the first generation VSB receiver broke up. The demonstration illustrated what Zenith engineers explained gave COFDM the advantage over the receiver chosen by Sinclair for its Baltimore demonstrations.

"The 2 dB advantage of the first generation VSB boxes had been washed out by a strong ghost," says Gary Sgrignoli, staff consulting engineer for Zenith. He pointed out that the second and third generation VSB receivers have a 5 dB advantage, compared with only 2 dB possible in the first generation receiver.

The engineers used a hair dryer to demonstrate impulse noise. Running the dryer while lowering the signal, the COFDM signal showed errors with no degradation on the VSB receivers. At 20 dB, reduced signal level 100 times lower than the generated signal, both COFDM pictures cut out with occasional errors observed on the VSB picture. At 23 dB, 200 times lower the generated signal, there was no observable COFDM picture, while the VSB signal showed occasional blocking errors.

The Zenith demonstration allows VSB proponents some leeway in their arguments that the technology is improving. However, the absence of reports from field tests is likely to leave many still unconvinced. Zenith officials rebut this argument, saying that lab evaluation is preferable because it is based on repeatable impairments, accurate measurements of trade-offs, and capability of objective comparisons.

Nontheless, the tests were an impressive showcase for VSB, particularly for the newest generation of demodulators. And if the name calling and political posturing can ever die down enough for the two opposing camps to work together to demonstrate tests with an unbiased observer (e.g. FCC), then maybe the issue can finally be resolved.