Relevance of Drop-Weight Testing in the Determination of the Reference Nil-Ductility Temperature

To further illustrate this process, let’s review the methodology presented above in a real-world example.

Consider the case where the first drop-weight test coupon with a F9P4-EM2-M2-H2 submerged arc weld deposit (obtained with Lincolnweld® LA-100 wire and MIL800-H Flux after post weld heat treatment, PWHT, at 1150 deg F for 48 hours) shows a “no-break” condition at minus 58 deg F. Based on this, we now know with confidence that the RTNDT is probably lower than minus 58 deg F.

Taking this first no-break result into consideration, the next drop-weight specimen should now be conducted at a temperature where a break condition is expected to occur. The temperature at which the subsequent test should be conducted should be based heavily on previous experience (either by the nuclear fabricator/constructor or by the testing laboratory operator). In my opinion, the temperature at which the second drop-weight test should be conducted should be at least 20 deg F below the first no-break test. This, in all-likelihood, will result in a test where the drop-weight specimen is broken.

Now imagine a scenario where the testing operator intended that the second test should actually  be conducted at minus 78 deg F (a temperature 20 deg F below the first test temperature), but because of specimen handling issues the drop-weight test was actually conducted at minus 76 deg F. Under the guidelines presented in E208 this does not pose a problem, and the specimen showed a “break” result upon testing. After that, according to ASMT E208, the third drop-weight test should occur at minus 66 deg F. Despite the fact that every effort was made to hit this temperature, the transfer time between the soaking medium and the testing apparatus was considerably fast, causing the testing temperature to be one degree off at minus 67 deg F. Again this resulted in a “break” outcome.

And finally, since the third drop-weight specimen was conducted only 9 deg F lower than the very first specimen (very close to the 10 deg F difference specified by ASTM E208), the fourth and final specimen should be conducted at the original temperature of minus 58 deg F. As with the first test, a no-break result was obtained at this temperature. Therefore, based on the formalism laid out above for the determination of the nil-ductility temperature, minus 67 deg F is our TNDT.

The next step in the process of the “determination” of the RTNDT of the F9P4-EM2-M2-H2 weld deposit is the administration of a set of three CVN impact tests at a temperature of TNDT + 60 deg F, which after doing the math is minus 7 deg F. As stated above, if the average of the three CVN tests at minus 7 deg F provides a CVN impact energy greater than 50 ft-lb of force and greater than 35 mils lateral expansion, the TNDT is declared to be the RTNDT.

As is typical with a F9P4-EM2-M2-H2 deposit after PWHT, the average of the three CVN impact tests at a test temperature of minus 7 deg F was 82 ft-lb and the lateral expansion was 62 mils – far and away greater than the required “50/35”. Hence, the RTNDT for the weld metal is considered to be minus 67 deg F. (Had the results of the CVN impact test happened to be less than “50/35”, then the TNDT determined in the drop-weight test is the actual RTNDT).

Although the establishment of an RTNDT has been referred to as a determination, I have personally been better able to comprehend this procedure when likening it to a RTNDT “verification”. In this scenario, because the CVN results exceed “50/35” criterion, the RTNDT is considered to be equal to or greater than the TNDT, and the CVN or “energy data” are regarded as governing the RTNDT determination.¹ I have seen that this result is usually the case for weld metal, as embrittling elements such as copper, nickel, and phosphorus are more tightly controlled in welding consumables than they were decades ago when initial studies of NDTT were performed.²

Although it is relied upon heavily for RTNDT determination, the drop-weight test has been considered by many investigators to be arcane and somewhat unreliable. One of the reasons for this lack of confidence in the results is the performance of the test in the transition region of the ductile-to-brittle temperature curve (an example of this type of curve is shown in Figure 2). For most types of ferritic steel weld metal there is a considerable amount of variance in fracture toughness as these graphs are often fraught with scatter.¹