Adequate fee and selective focus can help you walk the time/effort line.

We are fortunate in that the focus of many of our projects is making the controls work (and surprise - a superior control system is more easily achievable when it is the project’s main goal). However, most building controls are installed as part of general construction projects (e.g., a “spec” office building) in which making the controls work does not seem to be at the top of the priority list.

Recently, we provided services on a fairly typical new-construction office building project which provided some great insight into what is needed for making controls work in this environment. What we found added some interesting counterpoints to our September 2010 column, “Why Is It That Control Systems Don’t Work?”


Making a control system work starts with getting a sufficient design fee for the hours needed to “do it right.” Without pointing fingers, current A/E business practices for typical commercial projects usually yields engineering fees that are not commensurate with the design effort required (which in turn leaves little to no fee for the controls design). If you want your control systems to work then you first need to learn what fee is needed and then pick and choose your projects based on the ability to meet this goal. If that fee cannot be obtained, then the quality of the control system will be mostly left to luck (or, worse still, the fee will go into the red, finishing the design through the RFI and/or commissioning processes).

Even with a reasonable fee, the amount of engineering time available for controls design on general construction projects will still be fairly limited. Unfortunately, current A/E practice involves many design iterations that seem to go well-beyond the traditional SD/DD/CD phases. Interim/90% review and/or GMP design submissions add to the effort, along with the near daily battle to keep up with the architect’s design changes. The controls design effort needs to practice restraint concerning when to expend the fee.

The best approach is to keep design efforts to a minimum except for the most important design phase deliverables. Better yet, leave as much of the controls design to the latter part of the project when it is clear what the control system needs to control in the first place (in earlier stages deliverables should remain at a clearly delineated “outline” or even narrative level). If not, the controls design effort will just be needlessly chasing all of the other disciplines’ design iterations.


Many engineers substitute quantity for quality when developing controls design deliverables. A typical commercial project doesn’t need as much information as many might think. The following provides some specific guidance on where to put your design efforts:
  • The sequence is of the greatest importance. Nearly everything about a controls system - the points, architecture, integration, etc. - comes from the requirements defined in the sequence of operation.

  • A point list is useful but don’t over-embellish it. Are large complicated tables (with columns for “trends,” “setpoint,” “alarm,” etc.) really worth the effort, especially if points are already shown on the drawings?

  • Are control drawings really necessary? Does a control drawing for a VAV AHU or box add any useful information?

  • Keep the controls specification simple. Once the allowable manufacturers are listed in the specification, 90% of the system products/design becomes fixed. Use the specification for the 10% you can influence, like the few field devices that involve real quality/cost choices (e.g., flowmeters), some installation issues (e.g., where should IP controllers be used, who provides the IP communications), etc.

  • Don’t forget some key integration issues. A little time spent on reviewing/editing all of the other specification sections for equipment to be integrated (i.e., to specify the protocol to be used and the communicated data) will save huge amounts of time during construction.


You might be able to guess where we’re going with this but it will have to wait for another column.ES