We have a batch (92AE7) of OPA604AU devices that fail to meet their specifications in terms of slew rate and settling time when used as a voltage follower on +/-15V rails. The parts were sourced from a franchised dealer.
The example below shows the response to a +/-6V square wave input to both the suspect batch (92AE7, red trace) and a sample of another batch (68CKK, green trace). The vertical scale is 2V/div and the timebase 1us/div. The effective slew rate of the two parts is 13V/us for the suspect batch and 22V/us for the other batch. The datasheet value is 20V/us.
Is this a known problem with this batch, and, if so, how did it come about and how did it get through production batch testing?