I attended the ATC meeting of the NDIA earlier this week. That translates to the Automated Testing Committee of the National Defense Industrial Association. This industry association is composed of test equipment manufacturers, system integrators, prime contractors, and Department of Defense participants. The participants meet periodically face to face, and weekly via teleconference.
The ATC meeting in Orlando included a recap of programs and activities that span the various Services, and the work of the ATC itself. There is a large modular flavor to the work of the ATC.
Everyone knows that modular instrumentation is a great match with maintenance systems that keep vehicles and weapon systems alive. The small size allows it to be deployed in depots, in hangars, on vehicles and onboard ships. VXI has been the workhorse architecture for years, complemented by MMS (HP/Agilent’s Modular Measurement System, which was widely deployed for CASS) or high performance boxes where additional performance is needed.
Open modular standards, and VXIbus in particular, has enabled the commercial industry to leverage its considerable measurement expertise to the US armed forces and its allies. This has produced a win-win situation for the equipment vendors, system integrators, armed forces and the taxpayers.
It wasn’t pre-ordained to turn out that way. In early 1987, the US Air Force set to approve and publish its own “IAC” (Instrument on a Card) standard, based on the VMEbus. Frustrated by the lack of commercial standards and equipment, but needing downsized instrumentation, it felt it had no choice and was in the final phases of releasing a specification through the MATE (Modular Automatic Test Equipment) program. Formal approval was to occur at the next MATE meeting, scheduled in July 1987 in San Antonio.
While the proposed IAC standard would have satisfied the current Air Force requirements, it wasn’t suitable for commercial instrumentation, and would bifurcate the market. The products would not be dual purpose. This fear of splitting the market forever caused Hewlett Packard in April 1987 to call Tektronix to see if there was anyway to work together on an alternate standard that could satisfy military and commercial requirements. I know this, because I placed that phone call.
Tektronix was very receptive, and we invited three other vendors: Racal Dana, Wavetek, and Colorado Data Systems. What transpired over the next several weeks was the equivalent of the Manhattan Project for modular instruments. The July MATE meeting was an unmovable deadline, and a frantic set of rotating meetings was scheduled at each company’s location. Each company brought its best architects, and best practices, to create a modular instrument standard based on VME, in record time. How this was managed to move so quickly is a story in its own right, and became the subject of several news articles the following year. Its official title was VMEbus Extensions for Instrumentation. By June, the five companies believed they had enough for a standard, and to go public. Agreeing on a technical standard may be daunting, but having five competitors agree on a joint new release, and a name, is yet another story. It almost postponed the announcement. Deadlocked on names and trademark issues, I received a call from my Tek counterpart proposing a clever alternative: call it the VXIbus, where the X is for “eXtensions”. Bingo! VXI was named.
Besides the press event scheduled for the last week of June, there was still the issue of the MATE IAC standard the following week. A short polite letter was authored that said a new instrumentation standard called “VXIbus” was being created, that we believed it could satisfy military and commercial requirements, and that we requested that the vote to approve the MATE IAC standard be postponed long enough to allow a presentation about the new specification. It was signed by five companies:
- Colorado Data Systems
- Hewlett Packard
- Racal Dana
Our request was granted. Not just to present to the IAC subcommittee, but to the entire MATE general assembly. One after another, each of the five companies gave a portion of the presentation to a surprised audience. At the end of the presentations there was dead silence. I remember thinking to myself while looking at the audience from the lit panel on the stage, did we overstep our boundaries? Is it too much to take in? None of this occurred through the proper channels.
Then the auditorium erupted into spontaneous applause.
The rest is history. The Air Force embraced VXI, and that ignited a wave of adoption across the Services, and then to MODs of other countries as well. You can browse through the home page of the DOD Automatic Test Systems Executive Directorate to see many ATS families that have adopted VXI. Proving its versatility, VXI has been deployed outside of aerospace and defense to industries ranging from automotive test to semiconductor characterization. The VXIbus Consortium set a new model for how new standards are created and managed in the instrumentation industry, with consortiums formed for IVI, PXI, LXI, and now AXIe.
This brings us to the present, and the NDIA ATC. Some of these systems, like the author, are getting a bit old in tooth. New weapon systems demand new capability. In the days of VXI, there was just one choice to be made: which of the 4 sizes (A, B, C, or D) to deploy? Even then, C-size quickly became the de-facto standard. Now there are VXI, PXI, LXI, and AXIe standards. And many of these standards have choices within them:
VXI, with the largest modular installed base, has just added a high speed PCIe serial upgrade path through VXI 4.0.
PXI, arguably the prevalent commercial modular standard today, has two module sizes and has been updated to include PCIe fabric with PXIe.
LXI is the accepted standard for traditional box instrumentation, and has the largest overall product offering.
AXIe is an emerging standard that is best described as the “big brother” to PXI, and brings new breakthrough capabilities to the modular domain.
IVI has a number of standards related to I/O and drivers, and supports the above four standards.
When do you deploy which architecture where? This is where the ATC steps in. Giving perspective and guidance is the objective of a report being developed by the Platform Working Group within the ATC. It will be used as a reference source when developing new automatic testing solutions for vehicles and weapon systems.
For the above standards it will summarize:
- Application and use of each standard or approach throughout the automatic test industry
- The IEEE and industry standards that each implements
- Advantages and weaknesses of each
- Features of each including scalability, modularity, bus speeds, frequency ranges, power, and so forth
- Relative cost of implementation
- Physical dimensions of chassis and modules; number of slots available
- Electrical and mechanical requirements
- Availability of products that use the standard or approach
- Market trends (growing or declining market shares); outlook for the future of the standard
No single report can substitute for good judgment and pragmatic decisions. But this report will be a concise consolidation of the different aspects of these standards and serve as a starting point for new ATE designs. Experts on each of the architectures are working to present a consistent overview of the standards.
Open system modular instrumentation has come along way since July 1987 when it was merely a promise: a paper specification drafted on a UNIX word processor. Kudos go not only to the dozens of original developers of the VXIbus who started this, but also to the US Air Force who made a bold decision that forever changed the course of aerospace/defense ATE.
Subsequently, vendor, system integrators and the Services have embraced the modular bandwagon. Countless new vendors have added capability nobody imagined, in VXI and in the subsequent architectures. Now, there is no argument whether to adopt modular architectures or not. It’s clear that modular instrumentation will remain the primary form factor of defense ATE systems going forward.