Quote:
Originally Posted by Diver
Bye Zeity  |
Quote:
Originally Posted by NickFitz
When you say "more or less", are you allowing for minor timing variations? Sometimes cutting a little slack (aka rounding to a lower level of precision) can lead to better results, when dealing with a data source that is inherently a bit noisy.
You also need to take account of the fact that aliasing issues can arise if you are mapping a comparatively low resolution input to a high resolution output - you should determine a mapping equation that will transform the input space to the output space without introducing noise derived solely from the mapping.
 
Bye Zeity - I'm very glad to find that the background-data-gathering-exercise for your latest Cunning Plan is working, even though it'll probably lead me to a life of servitude in the Sugar Mines |
The slant is caused by the difference in clock frequency at their end compared with the crystal clock at my end.
Since the variable doing the counting is an integer, and currently set to 799 (good) as against 800 (slanty), I'm a bit stuffed really...
The sampling thingie at home has 4 bits of resolution, which gives a reasonable response with (Sample and 31) * 8. I'm sure that some further frigging about will improve it a bit.
The PIC16F88 version has 8 bit resolution, but goes from about 0 (roughly 0.3V) to 0xc0 (about 3.7V) because of limitations with the PLL.
I may have to change the program in the Pic a little too.
Linear Mode
