Converting trajectory frames of different MD codes to each other is not recommended since you face with many forcefield and topology errors as using them. Why do you want to convert it to NAMD trj.?
I have to perform covariance and network analysis on my proteins. If I'm not wrong, R accepts only .dcd files as input and not .crd. That's the whole point of converting the file formats.