The focus of this blog is construction-related topics. The purpose is discussion, so please feel free to comment! See Specific thoughts for thoughts from the daily life of a specifier.

21 February 2010

Finish Mock-Ups

I recently received an e-mail from one of my favorite architects, asking for help with an unsatisfactory drywall installation. Using the Gypsum Association's GA-216 - Application and Finishing of Gypsum Panel Products as a reference standard, he had specified a Level 5 finish, yet he still was having a difference of opinion with the installer.

I was a bit surprised this was an issue. We specify Level 4 for almost everything and have had few problems; it’s hard to imagine not getting a good Level 5 finish when specified. The exceptions, as you might expect, have been walls at an angle to large windows, or walls with down lighting, and the higher the sheen of the paint, the worse the problem.

The main effect of the skim coat required by a Level 5 finish is to give the same texture to the entire surface. If you have done your own do-it-yourself taping and sanding, you probably noticed that after sanding, the joint compound is smoother than the paper face of the drywall. Depending on the gloss of the paint and the way the surface is lit, this difference in texture can be quite pronounced. A Level 5 finish addresses this problem by requiring a skim coat of joint compound over the entire surface to minimize variations in texture.

The Master Painters Institute, which has defined seven levels of sheen - flat, velvet-like, eggshell, satin, semi-gloss, gloss, and high gloss - recognizes the Gypsum Association recommendations and standards, and states that a Level 5 finish is required for eggshell or higher sheen. That's a nice theory, but in practice, a Level 5 finish often is considered too expensive, and is used only for surfaces that will receive a semi-gloss or higher finish.

Unfortunately – and this bothers me - many terms we use in specifications are poorly defined, or not defined at all, making them essentially impossible to enforce. One person’s expectation may differ substantially from another’s, but it’s often the case that neither can accurately define what they mean. Specifications regularly require that installation be true to line, plumb, level, securely fastened, free of defects, smooth, and so on, usually without a reference standard. The terms "plumb" and "level" have definitions, but they are absolutes, as is "free of defects", and I'm sure most people think of "smooth" in the same way. The problem is that perfection is unattainable, so tolerances must be included or the specification is not literally enforceable.

Many industry organizations have gone to great lengths to help you specify what you want in a meaningful way, but even they can go only so far. For example, the Architectural Woodwork Institute (AWI) has three quality grades, each of which dictates how wide a piece of wood or veneer may be, how many knots it can have in a given area, how visible machine marks are, and so on.

For masonry construction, the Brick Institute of America’s Technical Notes 9A sets limits for dimensional tolerances and chips, and ASTM C216 - Standard Specification for Facing Brick states:
Other than chips, the face or faces shall be free of cracks or other imperfections detracting from the appearance of the designated sample when viewed from a distance of 15 ft (4.6 m) for Type FBX and a distance of 20 ft (6.1 m) for Types FBS and FBA.*
That sounds good, but what I can see at fifteen feet may not be the same as what you see, and we may disagree on what an imperfection is.

For drywall finishes, we could create something like the flatness and levelness numbers used to measure concrete slabs, add limits for the size and spacing of pinholes, and specify the sheen of the sanded surface, but at some point the definitions become unwieldy, and the cost of verification unreasonable.

To get what you want you must be able to define it. That means you must understand the limits of the materials and installation processes, detail and specify precisely, require mock-ups for critical finishes, monitor the work, and notify the contractor immediately when the quality strays from the approved mock-up. This won’t eliminate differences of opinion, but at least you have a basis for discussing what can be done. If you still can’t agree, you may be able to get a representative from an industry organization to help define what constitutes locally acceptable practice. And don't forget manufacturers’ representatives, who usually are willing to offer advice, especially if their products are being used. My experience with both industry and manufacturers’ reps has been very good.

In practice, mock-ups sometimes work against installers, as they try to do a really good job on the mock-up, only later realizing that maintaining the approved level of quality is too expensive. 
Example: Here is a picture of a mock-up for an exterior stone wall:
and here is a picture of the wall as first constructed:

It appears that the masons responsible for production were not the same ones who prepared the mock-up. While the production work in this example may be unacceptable, the mock-up created an unrealistic expectation. Even though perfection is the ideal, design professionals should have a realistic understanding of what can be done. (The wall in the example was rebuilt, not quite as well as the mock-up, but much better than the first week's production.)

  • Architectural Woodwork Standards (joint publication of AWI, AWMAC, WI)
  • ASTM C216 - Standard Specification for Facing Brick
  • ASTM C840 - Standard Specification for Application and Finishing of Gypsum Board 
  • Brick Institute of America’s Technical Notes 9A - Specifications for and Classification of Brick
  • GA-214 - Recommended Levels of Gypsum Board Finish
  • GA-216 - Application and Finishing of Gypsum Panel Products
  • MPI Architectural Painting Specification Manual
*Reprinted with permission, from ASTM C216-07a Standard Specification for Facing Brick (Solid Masonry Units Made from Clay or Shale), copyright ASTM International, 100 Barr Harbor Drive, West Conshohocken PA 19428


  1. Sheldon:

    Why did the mockup create an unrealistic expectation? The stonemason demonstrated a standard of workmanship that he could achieve, that was accepted, and that therefore was to be expected for the project. If we require that the mockup is to be constructed by the same people who will do the work, then there shouldn't be any issue of unrealistic expectations.

  2. Sheldon -

    Craig here. Good article. Another valid point that we ran into at Gehry was that a level 5 finish does nothing for a curved or slightly faceted surface. The GA definition calls for an even veneer over the entire surface for level 5. If the surface has any faceting at all, it will simple telegraph that through the veneer. And curved surfaces are basically never true to radius.

    We ended up creating a level 6 finish that called for a full veneer over the entire surface that must correct any surface defects or irregularities in geometry and requires multiple coats as necessary to achieve it. It basically is creating a gypsum veneer type of finish, but using standard taping compounds. (Gyp veneer specs have the same problem with irregular surfaces). It is an expensive surface, but the problem was actually raised by a subcontractor on a project. Once we inserted the level 6 finish in our specs, we very seldom had any serious arguments from the subs, most of them seemed grateful to understand what was expected from them from the start.

  3. Start from mid-point where somebody else has done the work would definitely make difference and I think this was the case with finish mock ups.