Do you mean 20Joules @ -40°C?

I do not have the ASTM spec for 588 so I cannot comment on the base materials having to meet the Charpy requirements. If memory serves its a weathering HSLA - however, I'm not 100% sure about that and if there are any requirements by the ASTM spec.

From the welding point of view you must watch your interpass temperature, only use stringer beads and make sure you minimize your heat input into the joint. Also make sure you choose a consumable with good impact properties. Here in Atlantic Canada, where many of the offshore structures must meet 47J @ -40°C the FCAW consumable of choice is Air Liquide's T12 wire or Nittetsu's SF-3A.

Perhaps you could provide a bit more detail... joint type, welding process, position, etc.

Regards,
Sean

Were these CVN samples taken from the HAZ (base metal) or from the weld metal?

I will not comment from the welding side, but will direct my response toward the testing side. It has been many years since I performed Charpy tests, but the first thing I did when performing the test was to inspect the Charpy specimen itself for dimensional conformance to specs. If we did not perform (or contract for) the machining, quite regularly I would find variations outside the dimensional tolerances for the specimen. The test is extremely dependent upon consistently machined specimens, as well as well maintained testing machine. The overwhelming majority of the rejected specimens was due to not correctly radiusing the notch. Too small a radius, or no radius, would cause the specimen to break at a lower level than it should under correct conditions. When the client was contracting the machining, I would always be very careful to get with the client to let him know exactly what was allowed so they would not waste their time and materials to weld a test coupon and then not get the required specimens out of it. This may not be your problem in this case, but all who need Charpy tests done should keep this potential problem in mind.

We have quite a few CVN tests done for materials we use. So far, it has been rare to see a failure. Having said that, our testing has been for bridge work and most tests were done at +40 deg F (+4 deg C) and not as minus 40 deg C.
A few tests were inadvertently done at minus 40 deg F (-40 deg C).
Those still met the minimum absorbed energy values of 15 ft/lbs (20J), although the test values were on the low side.

I know there are situations where the colder test tempertures are required, but not knowing your application, I am wondering if your tests are supposed to be at minus 40 deg F (-40 deg C), or should they be at + 40 deg F? I ask only because I have seen that mistake made many times.

Second set of questions: Is the testing required to be longitudinal or transverse to the rolling direction?
If longitudinal is required, is it possible that samples are being cut in the transverse direction? Maybe rolling direction is mis-marked?

Lastly, are you sure the lab technicians know the proper methods for performing CVN testing? I don't want to assault anyone's technical abilities but I have run into that situation before, and feel it is worth checking that out too.

Chet

Does the original product specification require CVN to your test temperatures? If not, there is no reason to expect the material will comply with your criteria. Your acceptance criteria, your requirments, must precede the purchase order for materials. All the details of sampling and specimen preparation must be known/planned in advance of testing and verified during the course of testing.

In all my years of impact testing on welds, I've failed with too much heat input, non-compatible filler material, or not reading my radiographs correctly, taking the specimen from an area with even a tiny defect.

A 588 does not requires any charpy testing however it refers to ASTM A 6. ASTM A 6 calls for chapy testing with absorbed energy of 27 J at 21 deg c unless specified. If you are looking for -40 deg c then it may have to be specified as special requirement while purchasing