The thing, SNS3, is that flying an NDB procedure with an
RMI in the pilot's primary field of view, like you have in your 747, is a piece of cake, compared to flying one with the usual ADF kit in a GA spamcan which has the ADF screwed into the panel somewhere 4 feet to the right just above the instructor's right knee - OK if the instructor is an female with good legs and a short skirt
The RMI tells you the bearing to the navaid
instantly, and you can immediately see which way the bearing is drifting so you can fix up the heading.
The traditional-GA fixed-card ADF, OTOH, has to be continuously compared with the DI, and since the card will either be fixed with 360 at the top, or (if rotating) will often be set by the instructor to have 360 at the top, you have to do some constant mental arithmetic. I did all this crap with a very traditional VERY BRITISH CPL/IR (who could naturally reproduce from memory the whole matrix of all the different map projections and the properties of each one) and he just loved the mental arithmetic bit; adding and subtracting 5 or 10 or 20 degrees or whatever the whole time. While the DI itself is a piece of absolute crap drifting off by a few degrees every few minutes and having to be constantly adjusted from the liquid compass (impossible in any turbulence) while the DI in your 747 is slaved to a fluxgate mag anyway....
That's why I paid a nice 4-digit sum to have a KI-229 RMI (ADF & VOR)fitted into my plane, which also has a slaved compass system, right from day 1. It removes a whole layer of messiness from NDB approaches.
What it doesn't do is fix up NDB radiation field distortions due to terrain, which make the accuracy so loved by all these ancient instructors largely illusory........
So, I fly NDB approaches with the GPS, on autopilot, while checking the ADF needle on the RMI, and noting with amusement as the difference between the two creeps up and down...... but don't tell anybody about this; we can't go undermining the old instructors can we?
