OK, I have problems with your instrumentation and testing methods.
First off, the intake manifold pressure is highly dynamic, especially the closer you get to the ports. A pressure switch can read peak wave pressures which will give false "high" signals. A better choice of pressure instrumentation for the manifold would be an absolute pressure transducer (or a vacuum gauge). This will tend to average out the peaks and valleys.
Secondly, intake manifold pressure is controlled by throttle position, not engine speed. It is proportional to engine load, so gearing obviously has an affect too. So when you were tipping in and tipping out and noticing that the pressure switch was changing, well, you were measuring your own foot.
Also, I wonder what the limit of your switch is. It sounds like it's about .45 to .5 bar absolute (about 6.6 to 7.4 psiA). At light load (small pedal angle) that sounds about right for 3100 rpm. That would also explain why your light went on when you tip-in quickly as there is a quick change in load.
So, pressure wouldnot be the best way to determine when the secondaries open. I would suggest reading the signal from the ECU to the IMRC controller. I know your'e not reading actual valve position, but you'll at least know when the computer wants the valve to open. It's fully opened / fully closed anyway. Or if you want to measure actual secondary valve movement, try rigging a rotary or linear potentiometer to the assembly. Or have DemonSVT use his scan tool.
The priciple that the secondaries operate under is fundamentally engine speed based. There can really be only one good switching point between these two runner lengths, and that indeed IS the lag inthe power curve (trust me, if there was another reason, the calibration would never have been released).
If you'd like a more detailed explanation of how this works, I can surely provide one here. But I'm already too wordy to begin with!
So, I would say that your intentions are good, but you need to change how you're getting your data. Then I would suggest testing at constant vehicle speed points (like every 10 mph) in 3rd, 4th, and 5th. And even 2nd if you want to.
Good luck!
