Letters

Where Analog Tools Stand In response to the letter "Interest In Analog" \[Sept. 17, p. 54\], analog HDLs/VHDL-AMS have been under development for many years. But they haven't achieved the level of adoption of logic HDLs, which, in turn, continue to evolve upwards.

IEEE 1076.1 was established in 1999 to add continuous-time modeling to the event-driven (digital-like) capabilities of the 1993 IEEE 1076 standard. It's a high-level modeling language covering electrical, thermal, and physical modeling in the time and frequency domains. It's comparable to high-level, mixed-signal behavioral mod- eling capability built into electrical simulators, like Spectre, ELDO, and Saber, and other tools, such as Matlab and MathCad.

The basic problem for analog HDLs compared to logic HDLs is the lack of a low-level universal building block to deliver independence from process technology. This, coupled with the fact that "analog" covers such a variety of disciplines, implies a major support problem: the number of variables, entities, and libraries associated with analog HDL can be huge.

The other key difference is that for analog HDLs, there isn't an automated synthesis flow from HDL to silicon: digital HDLs are used for design entry and verification, while analog HDLs are used for modeling.

Anadigm's current focus is on providing rapid, inexpensive, and easy methods to achieve analog and mixed-signal integrated solutions using field-programmable analog techniques. On the other hand, tools such as VHDL-AMS are targeted at ASIC systems for which NRE is large, and for which investment in such tools is cost-effective and necessary. So right now, there's a mismatch.

Two key considerations arise:

1. IP modules (the circuit building blocks supported within the Anadigm-Designer tools) are self-contained parameterised elements. Each maps straight onto the concept of an analog-HDL entity. But each also has a corresponding reproductive, process-independent physical implementation.
2. The emerging programmable systems market won't only require the ability to verify interaction of large system components, but also need to model changes in the system as components are reconfigured in real-time. For this complex simulation problem, VHDL-AMS and others are well suited.

Therefore, we can fully expect to see a greater adoption of this level of tools as programmable analog matures in the programmable systems market.
Ian Macbeth
Chief Technology Officer
Anadigm

It's Not The Tool That Creates Quality I enjoyed "Bad Drives Out The Good" \[Sept. 3, p. 75\]. Over the last 40 years, I too have noticed this phenomenon and attribute it to the business models wherein things be made cheaper and faster—but not better. Because high technology has become more widely accepted and used over the same period, the logical conclusion would be that this business model works.

People complain about the lack of quality in products, but aren't willing to pay or wait for it. Fortunately, not everyone thinks this way. I'm pleased that someone else has noticed this degradation in the quality of electronic designs and the influence that high-level programming languages have had. Somewhere along the line, the idea that it's not the tool that creates quality but those using it was forgotten.
Dick Suchter

Content Is King"Digital Audio Broadcasting: A New Frontier" \[Sept. 17, p. 50\] was perceptive and right on. The answer to "Will consumers pay a fee for this improved digital radio broadcast service..." has little to do with the digital nature of the medium, and everything to do with content.

Remember 4 Channel Stereo and AM stereo, just to recall two technologies that amounted to repackaging the same stuff that we were already getting in 2 Channel Stereo and AM High-Fi? These technologies failed because they didn't give consumers anything really different for their money. The current interest in Dolby 5.1 surroundsound? It's succeeding due to the content placed in video/audio services to exploit its abilities. So, bottom line, the answer is a question of content, not of the container.

Terrestrial commercial broadcast operators who essentially play the satellite programming and sell advertising are quite fearful of DAB, and rightfully so. But terrestrial broadcast operators who provide a targeted, truly meaningful service to their communities that can't be replaced by a nationwide program service feel little threat.

Consumers will pay a fee for an improved digital radio service if, and only if, what comes out of their speakers is more gratifying to them and is sufficiently more appealing than what they're getting today. Simply offering audiences more channels of what they're getting for free today in technically better digital sound tomorrow won't generate DAB subscribers! Content is king! Niche formats will succeed well. Special programming, similar to what HBO is to television, will drive subscriber sales. Listeners will subscribe and abandon "broad reach" stations that don't really serve anybody while attempting to serve everybody. That has many terrestrial broadcasters deeply concerned!
Mike Langner
Chief Engineer
Citadel Communications