I don't know about the 109 but I do know about the 225 and perhaps the answers are similar.
On an "very old-fashioned" helicopter without a coupled autopilot, the pilot interpreted the ILS indications and was hopefully aware of such things as where the wind was, what the final approach track was, how far down the approach he was, and how the seat of his pants was doing.
On an "old-fashioned" helicopter with coupled autopilot (B412 etc), the analogue electronic autopilot interprets the ILS data and is reactive - ie the autopilot carries on until localiser starts to move out, then makes a heading change to correct that and bring it back in. Ditto for glidepath. This type is fine if its smooth air with no great wind gradient, but does not cope well if its rough especially near the bottom.
Modern helicopters such as the 225 and probably the 109 have digital (software-driven) autopilots that use the same info as the pilot - ie instant wind (by comparing gps track, heading, TAS and groundspeed), final approach track from the course setting, how far down the approach it is from the ILS DME (if there is one), and how the seats of its pants is feeling from accelerometer data.
Thus on intercepting the ILS Loc it knows that if the final approach track is x and the wind is y then the ballpark heading should be z. It doesn't have to wait for a deviation from GS or LOC to make a correction because the accelerometer data is used to predict the flight path - the autopilot then makes control inputs to maintain the flight path before a deviation occurs. That results in a comfortable flight path with ILS needles bang in the middle, even if the daft pilot decides to slow down near the bottom of the approach etc.
It also knows how narrow the ILS cone is getting by ILS DME and adjusts its control responses accordingly, though it can cope without this with slight degredation in performance.
On the 225, the FM tells us that its important to set the correct final approach track on the course bar, and if for some reason the DME is not at the threshold, to use the DME HOLD function to let the system know. I have not tried to change the course during the ILS, but my feeling is that if I did so, there would be an initial heading change because would think its "ballpark heading" was now different, but it would soon react to the LOC deviation and come back onto track, though perhaps flying with a fixed offset on the loc.
Hope that goes some way to explain.
HC
Last edited by HeliComparator; 31st May 2007 at 18:15.