Why is it often the case that when transferring magentization from 1H to 15N, the 1/4J time is often a little less than expected - for example, it is often set to 2.25 ms. On the other hand, when transferring back form 15N to 1H for detection, 1/4J is typically set accurately to 2.7 ms?
I'm assuming the answer is something to do with relaxation of 1H during INEPT transfer. How necessary is this? Is there a way to optimize the values based on t2 time rather than sit at the magnet and manually optimize them?