The Visible Infrared Imaging Radiometer Suite (VIIRS) has been continuously observing the Earth with global coverage twice daily in the longwave infrared channels since January 20, 2012. These channels are primarily used for cloud detection, and for retrieving sea surface temperatures globally, as well as a number of other applications. The VIIRS sensor data records (SDR), aka level 1b data, have been shown to be accurate and stable at 0.1K level since the data reached validated maturity on March 18, 2014. However, during the scheduled quarterly warm-up/cool-down of the onboard blackbody calibration source, a calibration bias on the order of 0.1 K is introduced. The bias is further amplified by the sea surface temperature (SST) retrieval algorithm up to 0.3 K which causes an apparent spike in the SST product time series. Our previous study  reveals that this bias is likely caused by a fundamental assumption on the radiometric traceability of the VIIRS calibration equation, pertaining to the shape of the calibration curve. In this study, we further analyzed the equation and presented an improved correction algorithm known as Ltrace 2. This algorithm attempts to fundamentally reconcile the calibration curve shape assumption such that the calibration bias can be removed during the WUCD with better performance for all bands. Sample test results are presented to show the improvements using this algorithm.