My set up is a Pico ADC 200/100 and a PP/USB converter using Pico Log V5.20.3 software. We have just realised that the step interval between ordered data points does not equate to that predicted. For example, using a +/-500mV range with 8 bits, I expect to see a 1000/256 = 3.9mV step when the data set is sorted into increasing voltage. But the data shows a curious pattern in that for 9 intervals the step is 4.15mV and the 10th is 3.9mV (ignoring variation in the 4 place), see attached file. Although my attached data does not show it, we have seen this 9 * 4.15mV & 1 * 3.9mV pattern reversed. Any ideas of what is happening and is there a work around?
ziko wrote:Hmmm, the number of bits is not always 256 this is because in each of the ranges there is calibration and offset information.
How curious, though I am unclear as to how this explains the 4.15mV / 3.9mV increment pattern.
Or are these values indicative of the inherent error span in the accuracy of a single point measurement? Presumably if we went for extracting raw data using the get_values routine, then we would have to calibrate the output because of this intruding data?
We have been considering upgrading to a 12 bit / 4096 step logger (the 4224). Does this reduction in bits available also occur with the newer devices and software? If so, do you know at what level of accuracy of the mV increment does it impact (for say the +/-500mV range)? Also if we did purchase a 12 bit logger, does the .PLW file have the same internal format structure for a 12 bit as for a notional 8 bit logger?