Then And Now: Two Decades Of DMX512


Let's face it. The subject of industry-standard protocols might be a real snoozer for those of you who don't really care how they are created, modified, approved, tweaked again, reissued, etc. — you get the picture. Maybe you just want to know what a standard can do to make your job easier and your show go together faster.

Well, the history of DMX512 (and that's 20 years of history and counting) is the story of making your jobs easier, and all this time after the idea was hatched, it's still, quite literally, the industry standard. With all the advances in technology over two decades, it remains the protocol that made interoperability among manufacturers' products a reality, and whether or not you know it, it made the task of putting on a production a whole lot easier.

Without going into a detailed history of who said what on which date, who cried out against it, and who gnashed teeth, let's take a look at where we were as an industry 20 years ago, as well as the impact DMX512 has had, and continues to have, on our industry. Many of you remember this, and some of you don't: there was a time, not so long ago, before a standard protocol existed to help gear communicate. “Prior to DMX, each of the lighting companies had its own communication protocol. The most common was an analog voltage sent to each dimmer,” says Gordon Pearlman, now director of technology for Genlyte Controls, then president/owner of Entertainment Technology (then a consulting product design firm). “Even that was not at all standardized; Kliegl used 0 to +10 Volts DC, Strand used 0 to -10 Volts DC, and EDI used 2 to +7.6 Volts DC, for example.”

The field of multiplex protocols was even worse. Each manufacturer spoke its own language, and complex external black boxes were needed in order to speak to other brands. This was becoming a costly issue as shows got larger.

Bill Florac, now senior technical product specialist at ETC and then a product manager at LMI involved with development of the company's existing digital protocols, had made a number of digital attempts to compete with the Strand analog multiplex scheme. According to Florac, a major advancement came with the standardization of the RS422 drivers. “These provided the robustness needed,” he says. “But we were not the only ones to utilize these. This was good, but everyone had done it differently.”

At the time, LMI internally had at least two “standard” protocols and developed a number of boxes to convert one manufacturer's protocol to another. Even within different manufacturers' standard protocols, there were variants that would make them incompatible. “I know of a few sites where one of our boxes was used to convert Colortran to Colortran,” he says. “This seems odd, but our receiving logic was very flexible, so we were able to receive the Colortran variant from some odd console and put it out in a more ‘normal’ fashion for the dimmers.”

As the shows grew, so did the issue. So, some innovative folks — namely Pearlman, Mitch Hefter (then USITT engineering commissioner and in charge of programming engineering sessions for the 1986 USITT Annual Conference in Oakland, CA and applications engineering manager at Strand Lighting), Steve Terry, Al Pfeiffer, Steve Carlson, Bill Florac, Bob Goddard, and Fred Foster, among others — set out to create a standard by which all dimmer and console manufacturers would play well together in the production sandbox.

At USITT that year, a session was held to discuss the possibility of really making it happen. Representatives from major rental houses at the time, such as Production Arts, BASH Theatrical, and Vanco, as well as most of the major lighting manufacturers — including Colortran, Electronics Diversified Inc. (EDI), Entertainment Technology, ETC, Kliegl, Strand Lighting, LMI, and Teatronics, among others — were present. After some drama ensued, Pearlman whipped out a one-page paper he had prepared with his idea.

“Prior to that meeting, I sat down with my partner, Steve Carlson, and we discussed what we thought would be best and what we thought would be acceptable,” recalls Pearlman. “We had looked very hard at the protocol that we had developed for Kliegl that had many features that would have been desirable like error checking and addressable output devices but finally eliminated it because it required a CPU on both ends and, at that time, many products did not contain CPUs.”

The most likely choice, therefore, was a variation on the digital protocol then in use by Colortran, with an increase in the number of channels and the baud rate. “I have often thought that it was accepted only because it was the only one that was written down,” jokes Pearlman.

Once there was agreement that it was a realistic goal to create a standard, a consensus had to be reached on how it would be structured. “No manufacturer would get a competitive edge because they already had the protocol,” says Hefter, currently senior project engineer at Entertainment Technology. “Everyone had to share some pain.”

A committee was established to work out the standard, most of the details of which were hammered out by Pearlman, Carlson, and Terry, all at a time with no fax or email. Hefter did the CAD for the DMX512 timing diagram and the timing diagrams for AMX192, which also came out of that meeting, and he then took it to the USITT Board of Directors for approval and publication.

Later that year, Patricia MacKay, then publisher of Theatre Crafts (TCI) and Lighting Dimensions and founder of LDI, published the entire standard in both magazines. “It's important to put the problem in context: the consumer electronics industry had been stalled by the controversy over Beta versus VHS standards,” she says. “The personal computer industry had just emerged from the nightmare of battling operating systems. Both sides of the entertainment technology industry — designers/users and manufacturers — only stood to gain by everyone getting together and agreeing on a methodology by which all the equipment in the loop could communicate, without regard to who the original manufacturer was.”

And it worked. Well, DMX512 didn't work really well at first, but it would work. Some of the parameters were intentionally written with some flexibility so that everyone could adopt it and so it didn't favor any manufacturer. There was even a blatant error in the original timing parameters. These issues had to be worked out in years to come.

“It was arguably a bit technically inept when it came out, but that turned out to just not matter,” says Terry, currently vice president of research and development at ETC, and back then, executive vice president at Production Arts. Terry took a leadership role pretty early on in the process of implementing DMX512. “Lots of people thought it was a lousy standard that would never work. Refinements came in 1990 and a bigger set 14 years later, but the power of the concept of interoperability overrode all those technical issues. It hit that sweet spot of a protocol that was instantly understandable by a wide audience, was implementable for not too much money, and did about 80% of what you wanted it to do with that kind of standard.”

That's not to say there was no resistance. Until then, manufacturers were really in a position to “own entire segments of the market, and in order to own those segments, keep interlopers out,” continues Terry. “There were no kids in garages getting into these market spaces, because manufacturers were building walls around their products.” In addition, many consultants were not on board right away, because they were concerned about losing control over the quality of the systems they were designing. Since interoperability would make everyone's products more specifiable, consultants and owners would have to be responsible for the integrated systems. Would it really work?

But DMX512 would bypass the severe technical scrutiny that today's standards undergo because interoperability made it suddenly in demand by those who were buying volumes of gear: rental companies. Once people started talking about the possibilities, another transition came surprisingly fast. The original plan for DMX512 to simply connect consoles and dimmers rapidly gave way to the addition of automated lights, color scrollers, slide changers, fog machines — just about any piece of entertainment lighting gear that needed to be controlled. So while DMX512 was meant to be the lowest common denominator protocol for communication between consoles and dimmers, it instead became the only protocol, extending well beyond dimmers and consoles. And while the standard did not necessitate manufacturers give up their own proprietary protocols, those languages eventually faded away.

By 1992, there was no question that DMX512 had permeated the industry. The market had spoken and exploded, and manufacturers were suddenly able to push products into segments of the market never before imagined, all because of this concept of interoperability. It had worked.

“It has been said that, more than any other single event, the use of DMX512 is responsible for the huge growth of the automated lighting industry — if true, a somewhat amazing result from a little ‘guerilla’ standards effort,” says Terry. “Today, if you went into a standards-setting environment like we currently have in our industry, you'd get 15 analyses of why the original DMX512 was not going to be good enough. It was not error-checking; it's not really very fast; it only deals with 512 dimmers, etc. Then you'd basically engineer something that wouldn't, arguably, have hit the relatively unsophisticated sweet spot that turned out to be just right for our industry in 1986.”

Somehow, DMX512 came along at the right time to hit that spot on the mark and has served its purpose quite well, arguably better than its originators ever imagined. “I think it went well beyond our expectations,” says Florac. “I never thought it would last this long. In fact I, like others, thought of it as a temporary gap until we had a ‘better’ protocol.”

With two major revisions to DMX512 (it's now part of the ESTA Technical Standards Program and became ANSI E1.11-2004 — USITT DMX512-A, and it's 63 pages, while the original was about four), the standard doesn't seem to be going away any time soon, and interestingly, it turned out to scale up pretty well from that 512-channel “high end” limit.

“DMX512 will continue for many years to come,” says Hefter. “Although the overall communication structure of lighting control systems will migrate to ACN (ANSI E1.17), many of the end devices will still operate with DMX512, and various devices will take care of the translation between the protocols.”

Hefter adds that, in addition to the undeniable advances interoperability has brought to our industry, another benefit has been felt. “One of the most amazing aspects of the standards work in the entertainment industry over the past 26 years — the work on improving the National Electrical Code® began in 1980 — is how it has also formed strong relationships among the participants,” he says. “We may be competitors, but there is a lot of mutual respect and great friendships among us, and we all understand the benefits that standards bring to the industry.”

Here's to another 20 years of collaboration and technical achievement to make all our jobs easier.