X-12-ARIMA support for large data set (> 600)?

I recognize that US Census X12A implementation has a hard program limit of 600 for input data set, but would it possible for NumXL to dynamically pick the last 600 data points by dropping earliest observation and adjusting the start date?

This need to be dynamic so, when we add a new data point, it auto-adjust without having user modifying the input cells range?