Part I: Computerization Begins
The arrival of the first computers 50 years ago signaled that truss design had come of age. And they came just in time. We urgently needed them to meet the growing demand for truss designs. But they didn’t come with experienced operators – in the late 1960s, no one in our company had ever written a computer program before. In fact, colleges had just begun teaching programming.
Before the 1964 launch of the IBM System/360, computers had been largely unaffordable and incompatible. However, this machine so revolutionized the computer industry that over one thousand 360s were sold in the four months after its debut. But at $2500 per month, 360s were still out of reach for our industry. One did land at the University of Missouri in 1966, and I was fortunate enough to be among the first to ‘confound’ it with my FORTRAN programs. Later I gained a little more programming proficiency in my engineering classes.
By the late 1960s, IBM offered us a more affordable version of its System/360, the IBM 1130 (shown here), for about $700 per month, or half of the monthly salary of an engineer! Yet our engineers had to do double duty, designing trusses manually, and also learning how to program the 1130. And they did a commendable job, creating a program that initially calculated truss member forces and moments.
Prior to the computer, a truss designer’s main tools had been slide rules and gargantuan mechanical calculators. We generally received blue prints of truss jobs via mail from our customers. After doing a take-off, we drew the trusses to scale, and produced a graphical representation of member forces called a “stress diagram” (see illustration). After the forces were determined as shown, the truss joints were enlarged to half their actual size, and a transparent template was laid on top of these joint details to determine the size of the truss plate. On the template, the nails of the truss plate were depicted. We then counted the nails on each member and compared the counts to the number required by dividing the force in each member by the capacity of each nail. After our initial computer programming, some of the work, namely the drawing of the stress diagram, was eliminated. However, a few building departments still required it, and it had to be hand drawn.
This laborious design process was performed in a large open space, a “bull pen,” with rows of large drafting boards attended by our well-dressed draftsmen/designers. As is the case today, numerous phone communications were necessary to clarify design details. After considerable manual work, we delivered finished drawings to our PEs in cubicles behind us for review. When acceptable, our PEs sealed the drawings, returned them to us, and we ran copies with the blue print machine (not shown) in front of the bullpen. On a good day, our entire engineering department could design only a few dozen trusses. Even if those few designs enabled the fabrication of hundreds of trusses, the design costs of each truss were excessive, and had to be recouped from the ultimate buyer of the trusses. Thus, the need was urgent for the computer to automate more of the design tasks, and it was quickly integrated into truss engineering groups across the country.
Next Month:
Leaps Forward in Computerization