Can Glass Distortion be Measured?
by Jeffrey C.F. Ng

As glass technology has evolved and matured, user demands have never been higher for more transparency, clarity and distortion-free glass products. 

Glass manufacturers have met many of these challenges by developing visually exciting and energy-efficient products. However, the demand for clarity and minimal visual distortion continues to be a persistent issue, particularly with tempered or heat-strengthened glass. While these products are safer and stronger than typical float or annealed glass, the process that heat-strengthened and tempered glass undergoes introduces roll-wave distortion—a sinusoidal relief in the glass surface, often with regular peaks and troughs, which can result in visual distortion. 

Glass distortion has been reviewed, researched, studied and discussed among many design professionals, glass manufacturers, glass associations and national standards bodies. It does not take much to understand why. A tour of new glass structures reveals waviness, “pillowing,” reflected distorted images of the horizon or adjacent structures and other visual distortions. End users are concerned about the distortion-free quality of the glass assemblies and glass manufacturers understand that if they do not voluntarily develop improved standards and quality control, the market and governing agencies may impose criteria for visual distortion for glass.

What is Distortion?
Currently there are no national and objective standards for evaluating visual distortion in glass assemblies. Establishing consensus among design professionals, manufacturers and governing agencies has been particularly challenging. However, there are existing methodologies for evaluating visual distortion in glass assemblies and specification criteria for ensuring the flatness of glass panels. Additionally, there are good practices for installing glass assemblies to minimize the potential for visual distortion. Recent research by the Glass Association of North America (GANA) has begun to understand that visual distortion in glass is a form of lens distortion. This promising approach aligns the study of visual distortion in glass with well-established principles and standards of the science of optics. It may finally be possible to define measurable and objective criteria for glass distortion based on its optical distortion.

In the field of optics, distortion is a form of optical aberration, which deforms the image of an object as a whole as the object is viewed through a glass assembly. There are many other types of aberrations, including chromatic, spherical and astigmatic. These other aberrations refer respectively to color fringing or dispersion, an increase in blurring of images with distance from the optical center and selective blurring of parts of the images as viewed through the glass. Distortion is not a color or a sharpness issue, but an aberration that renders straight lines in the object to appear curved when the object is viewed through the glass assembly. Hence the term “curvilinear distortion” often is used. There are two main forms of curvilinear distortions, barrel and pin-cushion. In barrel distortion, straight lines appear to bend outward or “bulge” and, in pin-cushion distortion, they bend inward. 

A more detailed definition of distortion also considers three other variables regarding the geometry between the viewer, the glass surface and the object or reflection viewed: 

• The distance between the glass, the viewer and the object or reflection viewed. Distortion of the reflection of a building far away is often greater than reflection of a building that is near. Also, if the viewer is right up against the glass, distortion may not be evident, whereas it can be obvious if the viewer is a great distance from the glass.

• Whether the object is viewed through the center or periphery of the glass. As in optics, distortion often increases toward the periphery of glass.

• Angle of view of the object. Distortion often is more evident when an object or reflection is viewed obliquely through glass rather than at or near 90 degrees to the glass surface.

Evaluating Distortion
There are three main methodologies for evaluating distortion: visual, comparative and quantitative.

Visual evaluation involves a viewer looking at an object or its reflection through an individual glass assembly or representative mock-up. Approving a full-size assembly or mock-up helps establish a standard for a project, and a point from which all parties agree on an acceptable range of deviation. Because reflected images tend to accentuate the level of visual distortion, distortion can be evaluated by using a planar grid or a zebra board, which consists of alternating diagonal bands of black and white stripes, placed in front of the glass assembly. Comparing the linearity of the reflected image with the lines on the planar grid or zebra board can readily confirm if the glass assembly is distorted.

Comparative evaluation involves several designated viewers looking at several similar glass assemblies or representative mock-ups. The advantage of viewing several similar assemblies in this manner is that consensus can be developed among viewers on the acceptable range or variation from the approved mock-up. Future installations then would be compared to verify whether the level of visual distortion falls within this accepted range or deviation from the approved project mock-up or standard. 

Quantitative evaluation involves measuring various physical and optical characteristics of the glass assembly and its components. One of the most common and important measurements of heat-treated glass is measuring the peak-to-valley depth of the typical sine curve or roll-wave surface of the glass. Roller wave can be measured by using either a flat-bottom gauge or a three-point gauge.

It is important to recognize that the leading and following edge of the heat-treated glass panels will tend to have a higher peak-to-valley dimension, or wave amplitude, than the rest of the glass. Within the glass industry, criteria for roll wave varies significantly. The amplitude can vary from as little at 0.001 inch to more than 0.010 inch, depending on the manufacturer and the specifications for the project. Many glass manufacturing plants are set up with in-line sensors that can identify and reject glass products that exceed plant standards. 

While roll-wave distortion of the glass surface is a well-known contributor to visual distortion, it does not take into account the wavelength or peak-to-peak dimension. The GANA roll wave subcommittee has concluded that peak-to-peak dimension is an important factor in visual distortion. That is, a small peak-to-valley depth does not necessarily ensure distortion-free glass. A small peak-to-valley depth in conjunction with a relatively small peak-to-peak dimension still may result in unacceptable visual distortion. Therefore, it is important not just to specify the proper amplitude but also to develop a chart that correlates specific maximum amplitude with minimum wavelength. For a given amplitude, the longer the wavelength, the less visible distortion.

In heat-treated glass, the relationship between wave amplitude and wavelength essentially treats each roller wave of the curved glass surface as a lens that has a focal length. The lens power of the curved glass surface can be defined as the inverse of the focal length and can be measured in millidiopters (mdpt). This measurement is called the roll-wave factor (RW). It is possible to take all the roller waves of a glass panel and develop an average lens or optical power for the entire glass panel. Comparative evaluations of visual distortions can then be correlated with measurements of comparative optical power of glass assemblies.

The lens or optical power can be calculated based on measurement of the wavelength and amplitude of the roll-wave profile of the glass panel. It also can be measured directly off the glass surface, using a planar grid, a digital camera and commercially available software that meets ASTM C 11652/C 1652M-06 (Standard Test Method for Measuring Optical Distortion in Flat Glass Products Using Digital Photography of Grids). With this type of software, it is possible to calculate the average and standard deviation of the composite cylindrical lens power as well as the spherical lens power of a given glass surface.

Practical Tips 
Design professionals need to establish clear, definable standards for distortion-free glass assemblies. Agreement on such standards is a key to project success.

Several full-size mock-ups of glass assemblies should be specified and built prior to full production and final installation. Ideally, the design professionals should view the assemblies in their final project locations or similar project conditions and under various light conditions and times of day to establish a range of acceptable deviation from the approved project standard. 

They should coordinate with glass manufacturers to verify the feasibility and affordability of specifying particular criteria for roll-wave amplitude to ensure optimal flatness in the glass assemblies. 

The roll wave (axis of wave crest) should be oriented parallel to the horizontal window sill. The applicability of this criterion should be verified in a full-size mock-up. 

If distortion-free glass assemblies are critical to project success, design professionals should consider specifying a maximum acceptable lens criterion of RW factor for glass assemblies and coordinate with glass manufacturers on feasibility and cost. Bear in mind that generally the lower the RW factor, the higher the cost. 

Jeffrey C.F. Ng is a vice president at Thornton Tomasetti, which provides building engineering services and does curtainwall consulting and exterior wall rehabilitation.

Architects' Guide to Glass & Metal
© 2008 Copyright Key Communications Inc. All rights reserved. 
No reproduction of any type without expressed written permission.