Well, the manual/help files are quite explicit - re-scaling loses the data. I resorted to simply replotting all the points from a 'history' array I stored as the strip developed. I could successfully re-plot several thousand pts with no serious delay.
The fun comes when the array gets full. You can keep a counter, check for overflow and start again at the beginning, but this makes plotting a waveform that is split up within the array tricky. I made my history file twice the size I needed (which was N pts), and saved pts into it at (say) posns a[x] and a[x+N]. When x reaches N, simply reset it to 1. Then the whole history (up to N pts) can be re-plotted (when the scale changes) in one go, starting anywhere within a[1-N] for up to N pts.
Worked for me (twice so
far). HTH, Martin
