Got my datalogger to record the tacho signal directly and this is the log:
The green trace is RPM reported by the ECU. The red trace is RPM calculated from the frequency of the tacho signal.
I also found that the tacho signal can't handle much of a load. An extra 5mA load of the opto isolator led I'm using is enough to drop the tacho voltage down sufficiently to stop the tacho dial working.
Anyway, it seems clear that the glitches reported by the ECU are not there on the Tacho signal!
Need to check the RPM/TDC signal too.
Question is, is the ECU reporting the glitches because its using this value of RPM in computing injector timing? I need to repeat the logging with injector timing to check...