DeviceGray 8 bits

Omar Blodgett (left) with co-workers at the Globe Shipbuilding Co. shipyard in 1945.


For a given strength of weld metal, the capacity of a fillet weld is in direct proportion to its length and throat (and throat is directly related to the weld leg size). Increasing the weld length permits a decrease in the weld throat as long as the other factors remain the same. For any given fillet weld leg size, there is a corresponding length that will resist the applied load. While this basic principle of weld design is well known, it is less commonly understood that the combination of leg size and length can significantly affect cost.

Consider the three welds shown in Figure 1: the first is a 3/8-in. fillet, 6 1/2-in. long; the second, a 5/16-in. fillet, 7 3/4-in. long; and the third is a 1/4-in. fillet, 9 1/2-in. long. Weld capacity is proportional to the leg size and the length of the weld; so all three of these welds have the same strength. But when we compare the weight of the weld metal for each weld and the arc time it takes to make each weld, the savings achieved with the smallest, longest fillet are dramatic, as shown in the table.

When the designer specifies longer fillet welds, the greater length adds to both strength and weight in direct proportion. If the designer increases weld size, strength is again increased, but this comes at the price of a disproportionate increase in the weld weight (that is, the weight goes up by the square of the leg size). Good weld design, therefore, adheres to this principle: For fillet welds use a longer length and smaller leg size where possible.

When I was welding superintendent at the Globe Shipbuilding Company during World War II, we applied this principle to some intermittent welds. During this era, materials were in short supply, and we were under tremendous pressure to deliver ships as rapidly as possible. By applying the above principle to intermittent fillet welds, we reduced the required amount of welding electrode, reduced welding time and increased the quality of the assemblies.

We welded the inner bottom sections of the tankers on welding tables spread throughout the yard. They were turned upside down for better support. The design provided by the Maritime Commission (Figure 2) called for 3/8-in. fillet welds, 3-in. long on 12-in. centers. We didn’t want to reposition the work once it was laid out on the tables, so we had to do half of the welding overhead. The small space between the ribs, called the intercostal, made for cramped working conditions. It was immediately clear that, because of the size of the weld and the overhead position, it would take three passes to put in the specified 3/8-in. fillet.

We made several inner bottom sections as specified and found that making a multi-pass overhead fillet in this way often resulted in undersized welds. The inspector had to crawl through the “lightening holes” in the ribs to gain access into the intercostal. It was dark, so each weld was inspected with the aid of a flashlight. When rework was required, it was similarly cumbersome and time-consuming.

In contrast, it was easy to make a 5/16-in., single pass fillet weld in the overhead position (Figure 3). With some quick computations, I realized that a 5/16-in fillet weld, 3-in. long on 10-in. centers, had the same capacity as the three-pass intermittent weld specified by the Maritime Commission. While the weld was the same length, the closer spacing yielded a total increase in weld length when we considered the full length of the joint.

My next job was to convince the sometimes intractable Maritime Commission that smaller, more closely spaced welds were the solution. Fortunately, over the years we had made many suggestions to the group and we had earned their trust. They approved the alternate weld detail. The results?

  • A 16.7 percent savings in weld metal.
  • A 16.7 percent savings in welding time.
  • The job was accomplished in one pass, requiring only one inspection, rather than in three passes, which would have entailed three inspections and probably some rework.

More than 60 years later, the same principles apply.