Interview - January 2001
[Home Page]

[an error occurred while processing this directive]
(Click Message to Learn More)

EMAIL INTERVIEW  Tom Hartman & Ken Sinclair 

Tom HartmanThe Hartman Company was founded in 1972 as a high technology engineering firm, specializing in applying computer technology to commercial and industrial building control and energy management. Hartman has played an important role in pioneering the use of advanced computer based energy management control strategies. He continues to place a strong emphasis on the use of modeling for evaluating potential improvements, and has developed a number of in-house programs to model a variety of energy and financial improvement scenarios. Today THC is utilizing dynamic control concepts with networks, TRAV and advanced chiller design. 

"Open" Building Control Systems

Sinclair: I read with interest your firm's December Newsletter on practical network integration. I think all of us on the design/operations end of the industry want to move toward more open control systems, and many of us are frustrated at how slowly our industry seems to be moving in that direction. Would you bring us up to date on your view of the current situation.

[an error occurred while processing this directive]

Hartman: There is a lot of frustration over claims concerning "open" system architecture. We all have a basic understanding of what makes a system "open," and how our clients wish to benefit from incorporating open direct digital control (DDC) system architecture in their building plans. But it can be enormously difficult to attain the degree of openness desired, and systems intended to be open are falling woefully short of expectations. The reality is that many manufacturers who wave an open architecture banner produce systems that aren't open at all. The problem arises in part because whether or not a system is open depends on how one defines the term. In my mind to be open, a system must meet three important criteria. First, it has to be built on generally recognized standard hardware/software components. Second, those system elements must be widely available and cost competitive. Third, the system has to be capable of easily integrating a wide variety of other hardware and software elements that originate from other manufacturers. At present there is no system I know of that even comes close to these requirements.

Sinclair: Well, we now have BACnet and Echelon, two communication protocols that were intended to result in open systems, why haven't they, and is it a matter of time before they will?

Hartman: A decade ago, when serious work on these standards began, it was felt that the main hurdle preventing open systems was the widespread use of proprietary communication protocols that prohibited one system from connecting to another. We now know that communications is just one element required to achieve a realistically open control system. Other elements include standard database, programming and operator interfacing features. Here's an example. If I have an Echelon or BACnet based system and I buy a "compatible" component from another manufacturer that can connect to the network without disrupting it, does that demonstrate an open system?

[an error occurred while processing this directive]I think we would agree that connectivity alone is not sufficient to demonstrate an open system. Most would require that this new component also show that it can operate along with the existing components, and allow the operator to interact with it over the network before we would accept such this as an open system. However, if, to attain this level of interoperability, I have to buy along with that component one or more expensive programming or configuring tools and/or train my support staff on new nonstandard database configuring and programming techniques, is that an open system? Again, I think most would agree it is not, because it doesn't meet my stated criteria for openness, but many manufacturers claim that it is open. That's where we are today. Manufacturers' claim their systems are "technically" open while engineers and operators have come to understand that they remain "practically" un-open. So, a standard communication protocol is not sufficient to make a system open, and unless that is recognized, the current situation is unlikely to change much over the next few years.

Sinclair: That sound pessimistic, but in your firm's December Newsletter, I read the following statement. "I encourage designers and operators to keep their visions of truly open building networks alive, as it is a realistic goal. Already, it is practical to connect an increasing number of building system devices directly to building control networks rather than connecting them only as discrete points as has been done in the past." This sounds rather optimistic. Help me reconcile this with what you have just said.

Hartman: I am glad you mention that. I don't want my reality check to be discouraging to your readers. We as an industry simply need to understand that we have bitten off a much bigger chunk of aspirations than most thought when we undertook our quest for open systems. We have moved down the path, but we have a long way to go and it is not yet time to stop and claim victory as some seem to be willing to do. Before truly open systems are widely available, there must be a huge restructuring of the controls industry, and that will take time, but I do encourage designers and operators to take advantage of the movement we have made so far by integrating devices such as chillers, variable speed drives, actuators and other devices and equipment via network connectivity when such capacity is available and of reasonable cost.

Sinclair: Advocates of BACnet or Echelon networks tend to discount gateways as an inefficient and sometimes ineffective method of system integration. You seem to encourage their use. Why?

Hartman: Gateways do seem to have a bad rap among the communications protocol purists. But I very much encourage their use because it opens systems to so many tried and tested communication protocols that are simple and easy to implement. Keep in mind that the jury is still out on the viability of both the BACnet and Echelon protocols in the longer term. Each is relatively new and each has serious weaknesses. Furthermore, both are supported by control system manufacturers who at this point in time represent the most serious roadblock to the needed change in our industry.

Sinclair: Why do you think control system manufacturers are preventing necessary change?

Hartman: Let me be very clear that this is not some kind of conspiracy nor a result of corporate greed. These companies simply believe they are acting as any company should -- developing and maintaining their competitive edge in the market. In the ideal world, designers and operators would like control products to be interchangeable. But how can a manufacturer develop a competitive edge in a world of interchangeable products? The answer that ALL manufacturers have accepted is to keep much of their system proprietary while making their system sufficiently open such that they can connect to one or the other of the standard networks that have been developed. As long as they embrace this approach to being competitive, our industry will remain pretty much where it is today. They don't realize there is another path available to them. One with huge growth and profit opportunities.

Sinclair: There is another approach they could take that would be more helpful to the industry?

Hartman: Yes, absolutely. The industry needs to recognize that there are two distinct elements in any computer system -- hardware and software. Hardware can be developed, manufactured distributed and supported in the same way other commercial and industrial products have been for many years, but software is a very different type of product that needs to be treated differently at all levels. The idea of trying to develop products that integrate hardware and software together to the extent done in the controls industry is a source of enormous waste and inefficiency. I believe it is the lack of vision by the upper management of the control system manufacturers has led to these firms' failure to recognize and exploit the opportunities this difference provides.

Sinclair: What would the industry look like if this hardware/software separation you speak of actually existed.

Hartman: It would look much more like the PC industry in which a number of hardware manufacturers compete to make hardware platforms that are based on certain standards. Meanwhile, separate software manufacturers compete to develop and sell software products compatible with the various hardware platforms. There are also a lot of integrated hardware/software products in the PC industry, but each such product is focused on a specific purpose. In our industry we will have manufacturers that concentrate on almost entirely on hardware, others that focus on software, and still others that sell devices and special application products. Building operators will be able to buy four by four (four inputs and four outputs) and eight by eight controllers. Some may have actuators and/or other sensors directly incorporated for special purpose applications like terminal unit control. Plugs and sockets will permit the addition of memory, point expansion, or special gateway connections. These hardware components will be much less costly than the prices we currently see in the industry, but they cannot make a complete system in and of themselves because they lack essential software. To complete a controller as we know it today, the user must purchase applications software that is used to communicate, manage the system database, and implement and execute specific application programs for each node. Control subcontractors (the so-called system integrators) will finally be free distribute a wide variety of products, so that no user will be tied to a single source of support for the products incorporated in their system. We are now beginning to see new firms making inroads into our industry using this model. When you look at the economics and lack of user satisfaction with our current building controls industry, it becomes clear that a change from the current model toward this one is only a matter of time.

Sinclair: Well, although there may be some issues about accountability in such a model, I think there is likely widespread support for such a future among designers and users. How can we as participants in the industry work to achieve these goals?

Hartman: All of us need to keep our eyes on the ball. Too often I see specifications that simply dictate the communications requirement of the system with the idea of achieving an open system. Those folks are swinging with their eyes closed. They'll never connect that way, and they often look silly doing it. Instead, we need to really work with our clients to find out what they want in terms of future expansion, operation, and independence from single source features. Then we need to work toward achieving those goals. The more we push manufacturers to realize user needs, the more likely their upper management will begin to understand the shortcomings of their current products. I was talking with a colleague a short time ago who did a survey of chiller plant controls. An overwhelming majority of digital control systems for chiller plants he surveyed have been disconnected and the plants are operated manually because the controls never really worked well.

We need to avoid participating in the current protocol wars that are much ado about very little, and focus on what will really help make the system you are designing open for the needs of your client. If the system is intended for a single facility then the required openness may best be achieved by selecting the system that appears to fit best and open it up with locked-in long term unit pricing and equipment warranty agreements -- directly with the manufacturer. This may not permit other equipment in the door, but it does control future costs which is a good first step. If your client desires an open system for a large growing complex, then you should consider newer approaches that employ the model I outlined above. That approach is now possible, but it is unconventional and it takes a great deal of support throughout the design, construction and startup processes.

Sinclair: Well, this perspective will likely develop a great deal of interest and be a source of further discussion. But before we end, I do have one final question. It seems to me that more intelligence is being built into end devices like sensors and actuators, that by their nature are application specific and not easy to connect to networks. Is this complicating the present situation?

Hartman: These development are really bringing to light the weaknesses of out industry's current controls configuration practices. Manufacturers of microprocessor based actuators, sensors and other devices understand that their products can easily incorporate additional application software, programming space, and network communications for a very small marginal cost, but they are reluctant to build on these features because they are seen to be competing with the control manufacturers who are their customers. But in reality, many device and equipment manufacturers have become controls manufacturers too. I have just completed the points list for an ultra-efficient all-variable speed chiller plant. It has about 120 points and nearly all of them are chillers, VFDs, large actuators, or flow meters that can be connected via network connections. If and when these device manufacturers realize they are also controls manufacturers, such a system will not require any additional control system. We will simply connect this equipment together and put the programs in the devices themselves. Consider the savings that could bring to implementing a project of this type! More and more designers and operators are speaking up about these kinds of inefficiencies in our industry and sooner or later, someone is going to start taking advantage of them.

Sinclair: What I think I hear is that your optimism for the future is really based on the inevitability of change to a more effective controls industry.

Hartman: Yes. And I encourage all who are active in the industry to contribute however they can. Becoming more educated on digital controls helps designers become more effective advocates of change, and pushing to represent more than a single manufacturer will help quality controls subcontractors gain influence with their customers as well as their suppliers. There is a role for all of us to help speed this necessary change.

Sinclair: Thanks for your insights

Hartman: You're welcome. If we work together, we'll see a lot of change this year!

Readers may find more information about technologies discussed in this interview, or sign up to receive The Hartman Company e-mail Newsletter at  Comments and questions about the article may be addressed to Mr. Hartman at 

More Interviews

[an error occurred while processing this directive]
[Click Banner To Learn More]

[Home Page]  [The Automator]  [About]  [Subscribe ]  [Contact Us]


Want Ads

Our Sponsors