| ??? 08/23/10 20:19 Read: times |
#178253 - An ADC doesn't need to be scaled 1 mS/cm for each tick Responding to: ???'s previous message |
Anita K. said:
Per Westermark said:
You can cover the range 0 to 2000mS/cm with 6 bits. Or with 8 bits. Or with 16 bits.
The range doesn't matter for the number of bits. Is there any trick for that? where can i find any literature to learn it how. No trick - just a question of resolution. With 6 bits, you get a resolution of 2000/64 = 31 mS/cm/tick. With 8 bits, you get a resolution of 2000/256 = 8 mS/cm/tick. With 16 bits, you get a resolution of 2000/65536 = 0.03 mS/cm/tick. And whatever resolution you have, you still have to consider what precision you have - if you get a high-resolution value that is +/- 50 ticks in precision then the extra resolution doesn't make you so much happer unless your only problem is signal nose in which case you may be able to perform low-pass filtering. But then again - with a noisy signal, a lowpass-filtered signal can have a higher resolution than the ADC resolution as long as the noise is always enough larger than one bit of the ADC so the noise is always able to tick the ADC value up or down. |
