There has been some confusion surrounding the exact implementation of the Doppler effect on the QSI Upgrade Chips.

How not to do it!

First let's address the two common ways that users think they are implementing the Doppler effect.

First is to use function 6, theoretically this is correct, but the way you use it will greatly affect the outcome. Some people think that you blow and hold the whistle, engage function 6, and then release the two at exactly the same moment. This will give you a slight shift in tone but not really the full sweep of sound you would expect.

Second, I'd try to blow the whistle for more than a second, quickly release and give a second short tap on it. With most of the steam locos this will give a neat sort of tailing effect on the whistle (on the diesels it doesn't do anything). This tailing effect is what a lot of people were confusing with the Doppler effect, this is not the actual Doppler though the correct method to implement it is very similar.

How to do it!

To experience the full sweep you'd expect from the Doppler effect:

1: Hold the whistle/horn button for more than a second.
2: Very briefly release the button and re-engage it, but unlike the tailing effect this time hold the button for roughly another second. Upon the second touch of the button you should hear a dramatic Doppler shift.
3: Using Function 6. To implement the Doppler push function 6 while at speed and without the whistle/horn engaged and then blow the whistle. The difference between this and the double touch method is that the sound sweep here is much more subtle, the whistle drops from high pitch to low pitch very quickly, almost imperceptibly so.

The important thing to remember about the Doppler effect in these locomotives is that it is a speed responsive effect. The faster you are going the more pronounced the effect.  If you want the effect to be more pronounced at lower speeds I recommend that you adjust CV 5 (V-High or Max Speed) and turn the maximum speed of the locomotive down so that the Doppler responds more dramatically at scale speed. I set the value on my tester to 200 with good results.