I think that the real complaint is the reading is not a steady RPM to=20 the full resolution of the display. Could be not enough consecutive=20 readings are used to calculate an average reading, but then, the more=20 readings, the poorer the shown response to actual speed changes. It=20 wouldn't surprise me that the single units digit of the display, or even=20 10's was driven as a "0" permanently) I still question whether the RPM=20 isn't a full digital circuit. And then is there noise in the circuit. :)=20 Perhaps better (and not trying to stop this thread here) forum might be=20 in the homebuilt aircraft area. On 1/18/2012 11:44 AM, John Coppens wrote: > On Wed, 18 Jan 2012 20:06:47 +1300 > IVP wrote: > > =20 >> There's a minor problem with the tachometer that I'd be interested >> to know if others have >> =20 > Note that, to measure RPM with a resolution of 20, you have to measure > at least 3 seconds, if the speed is actually taken from the axle. > > If exactly 3 seconds are measured, the result will always be +/- 1, as > the number of transitions will randomly coincide with the 3 second gate. > > To improve precision, you can either: > > - enlarge the gate time, then divide. But then the response time will > be unreasonable - the time from changeing throttle to readout. > > - Put in a processor which filters, but not simply filter the RPM, for > the same reason as previous. You have to measure the period instead. > > - Put in a PLL (CD4046) which can filter in the feedback loop. > > > Note: I somewhat doubt that the RPM will pass through the ADC, unless > the RPM sensor is linear/analog. > > Though I like precision too, 2500 +/- 20 looks quite precise to me. > Jumping LCD/LED displays are the curse of modern technology - nobody > cared when the old mechanical RPM needle was nearly 100 RPM wide... > > John > =20 --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .