Rubber components used in voting machines.

Punch card voting machines are used in many election districts throughout the United States because of their relative low cost, ease of adaptation to long or multilingual ballots, and generation of more or less machine-readable permanent records. In the aftermath of the last presidential election, those used in certain districts of Florida led to a reconsideration of their appropriateness. In addition to other factors, the durability of rubber components was called into question (ref. 1).

Two types of devices were in use. In both, a metal-tipped stylus is placed through a circular hole in a clear molded polycarbonate sheet over a position on a punch card ballot, presumably corresponding to the voter's choice, driven through a partially incised rectangle in the card, carrying this effluvium, the "chad," through a slot formed by strips of rubber lying directly beneath the punch card. The spacing between the rubber strips is smaller than the chad (or the stylus). Thus, the machine traps the chads; in some areas, these are actually recycled.

The force needed to drive the stylus is rather variable (25 to 80 grams) as there are several types of stylus, and variation in the punch cards.

The Votomatic machine, based on models developed many years ago by IBM, uses molded T-strips of vulcanized rubber in a parallel arrangement, held in place by a flattened area at each end that is clamped into the plastic (ABS) frame of the device. Unused T-strips and those from a machine made in 1975, used and stored under ambient conditions, were obtained (ref. 2). Deconstruction of the compound suggested the following:

SBR 1605          112.5
Natural Rubber     25
FEF or similar     40
Process oil        10
Coumarone resin    10
Zinc oxide          5

plus minor amounts of stearic acid, phenolic antioxidant and a sulfur-sulfenamide cure system. It is essentially a 65 durometer A platen or roll cover, with tensile strength over 17 MPa at optimum cure and ultimate elongation of about 350%. The surface of unused parts was without any evidence of exudation or bloom. The 25-year-old samples had advanced to about 70 durometer A, but showed no evidence of micro-cracks under low magnification, even in the ends that had been clamped. Accelerating aging of unused samples for 100 hours at 75 [degrees] C led to about the same increase in hardness.

The Poll Star machines use a one-piece injection molded TPU pad, the gaps being provided by slots, rather than arrangement of individual T-strips. The compound used is Bayer Texin 285, a polyester TPU. Although this compound yields 85 durometer A, the force needed to drive the stylus is about the same as with the Votomatic machines since a slightly thinner configuration is provided. An unused pad and one with four election's usage were obtained (ref. 3). The latter showed no increase in hardness compared to the unused pad (82-86 durometer A).

The above samples were examined by means of dynamic mechanical analysis (DMA) using a Rheometrics three-point bending fixture adjusted so that the amplitude of deformation approximated that caused by a stylus in a punch card voting machine. The samples were not reduced to a standard thickness, but used as they are found in typical voting machines, that is, the TPU samples are slightly thinner than the SBR/NR analogs. In figure 1, elastic modulus versus frequency of deformation is plotted for unused samples of the two strips.

[GRAPH OMITTED]

Two conclusions are evident. There is not a great deal of difference between the response of the TPU and SBR/NR compounds at low frequency. It is not known whether the TPU compound was chosen specifically to approximate the dynamic characteristics of the original SBR/NR compound, or simply as a durable material that could be readily injection molded. Secondly, the range of frequency of applied deformation needed to induce a significant increase in apparent elastic modulus is far greater than could be found in the field, even assuming usage by voters who struck rapidly without study of the ballot (not to speak of the instructions).

In figure 2, elastic modulus is plotted versus frequency for the used samples. The TPU samples are essentially unchanged. The strips of SBR/NR compound show a considerable increase in modulus, but only at frequencies far beyond what could be encountered in the field.

[GRAPH OMITTED]

Maintenance of the voting machines comprises disassembly, removal of accumulated chads, and treatment of the rubber strips with silicone spray, such as used to clean typewriter platens. Folklore has it that this treatment "rejuvenates" the rubber, the assumption being that some component is absorbed into the composition, affecting its properties (ref. 4). The used rubber strips of both types were so treated, the applied spray being allowed to dry at room temperature. Repetition of the testing described in figure 2 yielded identical results. It is likely that this maintenance procedure has instead the effect of removing paper fibers from the acting surfaces of the strips, returning coefficient of friction to its original level.

In summary, the elastomeric components of punch card voting machines appear to have satisfactory durability for the intended use. It is likely that unreadable ballots derive to a much greater extent from human error than lack of machine fitness for service.

References:

(1.) Case No. 00-2808, Gore, Lieberman et al., v. Harris, Bush, Cheney et al., Circuit Court, Leon County, Florida, Dec. 2, 2000.

(2.) Courtesy of United American Election Supply Co., Napa, California.

(3.) Courtesy of Election Data Co., Valley Center, California.

(4.) For an actual example of remediation in the field, see Rubber World, July 2000, p 20.

Richard Grossman is technical director with Halstab, a supplier of lead based and other heat stabilizers. He began his career in the rubber industry in 1957 with Anaconda Wire & Cable. He has also worked for Polysar, Synthetic Products, Cooke Color and Polymer Services.