There are actually FOUR different sources / types of altitude displayed, transmitted or visible in the respecitive cockpits plus the FAA radar decoding of same. Now add that the Kollsman window altimeter settings could be set differently, too … adding some correlation / actual altitude discrepancies.
First off, altitude ON the airplane is generated either by a “blind” encoder (separate from the altimeter) or by a combination altimeter / encoder. MOST GA aircraft use a separate unit. So there could be a correlation issue between what the blind encoder is transmitting (or being data recorded) and the actual altimeter reading the pilot sees. FAA Regs require that correlation to be within 125’. In view of what happened and the requirements for holding a 200’ altitude, this large correlation error number seems ridiculous; It would be possibe for the helicopter pilot to SEE 200’ yet be at 325’. Same thing for the CRJ in reverse.
Further, there are two types of blind encoders; I have no knowledge of the systems in either airplane. Older discrete parallel wire encoders were only accurate at 100’ increments. Newer serial altitude encoders are accurate to either 25’ or 10’ … I can’t remember. Maybe Larry Anglisano can shed light on this. My point it, the encoders themselves may not have been reporting that accurately to FAA radars.
I do not know what type of altitude encoding equipment was aboard the helicopter but – think of it – the helicopter pilot may have been seeing 200’ when the helicopter may have actually been at 325’. Assuming the CRJ was at the correct RNAV approach altitude and on the 3 deg. VASI path (SO far, that seems to be the case), then this small allowable altitude error could be where the problem was.
Allowing ANY helicopter ops on that approved ground path at night when runway 33 is in use at DCA was the beginning of the accident chain IMHO.