LESSONS LEARNED IN GTAW QUALIFICATION TESTING

To emphasize the importance of the effects of welding heat input and cooling rate on weld mechanical properties, this analysis examines how a one percent nickel electrode was successfully implemented in a welding application.

This month I would like to discuss yet another application where a one percent nickel electrode was successfully implemented in a welding application. Recently a fabricator wished to qualify a gas tungsten arc welding procedure using a 3/32 in diameter ER80S-Ni1 solid wire. For the procedure test, they assembled a test plate with precisely the same dimensions as the AWS A5.28 filler metal qualification plate. The mechanical properties of the weld would then be tested to ensure that the proposed procedures provide adequate strength and toughness.

Based on the intended service of the welds, the fabricator’s engineering staff determined that the ER80S-Ni1 weld deposits must possess a mechanical property of 70,000 psi yield strength. This level of yield strength is actually an even more stringent requirement than the 68,000 psi required by the A5.28 filler metal specification.

An additional requirement self-imposed by the fabricator was that the plate needed to be preheated to 300 deg F, and the interpass temperature was to be maintained between 300 deg F and 350 deg F. As it turned out, during the qualification test the welder kept the temperature of the plate as high as possible during the test at 350 deg F. The reason for this was quite clear: more time would be required between weld passes if the plate was allowed to cool down to 300 deg F before each pass. After all, as I’ve illustrated in previous columns, time is money.

As far as welding parameters are concerned, the constant current GTAW power source was set to an output of 225 amps and the average travel speed was about four ipm.  This four ipm travel speed is not incredibly fast, even for the GTAW process. When all was said and done, the average welding heat input was estimated to be about 42 kilojoules per inch per pass. I say ‘estimated’ because the welding output voltage is part of the heat input calculation. The welding voltage is not something that can be ‘set’ in the GTAW process and can only be controlled by the stability of the hand of the welder or the robot.

To be more specific, in GTAW a shorter arc length (the distance from the end of the tungsten to the workpiece) means that the output voltage is lower. The longer the arc length, the higher is the output voltage.

The welding qualification test assembly was welded in a total of 21 passes, with the weld bead placement shown in Figure 1. It can be seen that the first few weld layers are significantly thicker than the layers closer to the top of the plate. For the first few passes, this indicates a relatively slow travel speed (and high welding heat input) with a large amount of wire being pushed into the puddle.

In order to accommodate this large amount of wire and keep the end of the tungsten electrode from touching the weld puddle, the welder was forced to maintain a long arc length as shown in Figure 2. The problem with a longer arc is that a higher arc voltage results, further increasing the welding heat input.

This scenario is precisely what happened during the fabricator’s qualification test. To summarize, a slow travel speed, high output voltage and high interpass temperature equated to a low “quench rate”. This lower quench rate can produce lower tensile and yield strengths. It therefore came as no surprise when the all-weld-metal tension test specimen machined from the plate exhibited a relatively low yield strength of 65,600 psi. Not only was this yield strength lower than the 70,000 psi required for the intended application, but it was lower than the 68,000 psi minimum yield strength of the AWS ER80S-Ni1 filler metal classification.

Rather than blame the electrode or the electrode manufacturer, the fabricator decided to try again and weld a second qualification plate. This time, however, closer attention was paid to the variables as outlined above. The first variable, and the easiest one to control, was preventing the arc from being struck until the weld test plate cooled down to 300 deg F. Staying on the low end of the interpass temperature range helps to increase the cooling rate, and hence increase the yield strength of the weld metal.

Keeping with the theme of decreasing heat input, the travel speed was increased from four ipm to nearly six ipm. This change may not seem like much, but it had a dramatic effect on the heat input. Keeping all other factors the same (voltage and amperage), a 50 percent increase in travel speed would result in a 33 percent decrease in welding heat input. Again, this is another factor that will typically lead to an increase in weld metal yield strength.

Now that the welder has moved his torch faster he is depositing thinner weld beads, which means that the welder can keep his arc length shorter and the tungsten closer to the workpiece, as shown in Figure 3. This results in a lower arc voltage and once again a decrease in heat input.

The ultimate pass sequence settled upon by the fabricator is shown in Figure 4. The pass count was increased from 21 to 22, which resulted in one extra weld layer (eight layers rather than seven). But even more important than the pass count number itself was the fact that the beads on average were smaller and more consistent in their thickness (particularly in the middle of the weld where the tensile specimen is taken), thereby increasing the quench rate and boosting yield strength.

The results of the second tensile test were much more satisfactory. A yield strength of 72,400 psi was achieved this time around, which exceeded the fabricator’s own requirements for the ER80S-Ni1 deposits. This shows once again just how important it is to be mindful of the effects of welding heat input and cooling rate on weld mechanical properties.