More on wireless interoperability

03/02/2010


One of the comparisons I’ve been hearing about ISA100 is that it’s similar to adoption of 4-20 mA as a communication platform. While 4-20 mA has never completely eclipsed all other analog protocols, it certainly is dominant. Back when it was made into a standard, various companies began to look at what they could do to expand on it. There were several protocols developed to put additional digital information on top of the main analog signal, and among them HART ultimately proved the most successful and has prospered.

Among the companies that worked in that space (Foxboro, Honeywell, Emerson, etc.), they all created their own approach and did the initial development work. The same question applies now as companies contemplate what they might be able to do to develop application profiles that can be applied on top of ISA100.11a.

For example, if Profibus or Foundation Fieldbus decides to create a wireless version of its protocol, could it not be done as an application profile that can be carried by the 11a transport mechanism? Those respective organizations would undertake that development on their own and create the necessary code and compliance mechanisms to commercialize such a venture.

While some might consider this inviting anarchy of the first order with as many or even more possible protocols than exist in the wired world, it actually presents an interesting possibility. When Honeywell brought out its “ISA100 ready” wireless devices, some cynics scoffed that you can’t create something for a standard that does not yet exist. Honeywell has promised that it will be able to upgrade the devices when the standard is finalized. It says it will, for all practical purposes, be able to go into the software of the device and strip out the proprietary stack on top of the 15.4 radio. That old stack will be replaced with the 11a stack with no modification to the devices. In fact, Honeywell says that it can be done over the air. I don’t know if this has actually happened in the real world yet, but to my knowledge there has been no back paddling on that promise.

If this is possible and practical, it suggests an interesting possibility. Why couldn’t a wireless device be sold as a blank slate? The user could select from any number of profiles riding on 11a and simply load it on site. When a pressure sensor is mounted in the plant, the user turns it on, and as part of joining the network, it is configured by the control system to communicate using the appropriate profile. The larger system sets all the operating parameters, diagnostics, etc., automatically over the air. This would also offer a user the opportunity to change or upgrade as necessary as time goes on.

The instrumentation vendors would need to sell their devices with sufficient memory available to make this possible. Any additional cost would be mitigated by larger manufacturing lot sizes since there would be only one part number for any given device. This sounds like one of those things that is too practical to happen. Is there some huge technological hurdle I don’t know about?