Resolution

GHaines · Post by **GHaines** » Mon Jul 31, 2006 8:26 am

Hello, could somebody explain what is meant by, for example, 20 bit resolution. I imagine that a higher bit resolution is a better resolution but I have no idea how to quantify the concept of bits in resolution terms. If you could point me in the direction of a good website which explains the idea then that would be great. Thanks

Michael · Post by **Michael** » Mon Jul 31, 2006 10:14 am

Hello,

Thank you for your post.

Right, the resolution we are referring to is the vertical or voltage resolution.

This is calculated in the following manner:
If the voltage range = Â±2500mV, this gives you an effective range of 5000mV bottom to top. This is then divided by the digital resolution, this digitisation will convert an analogue voltage number to a binary number that can be handled by a PC. The resolution is calculated by dividing the effective voltage range by the quantization levels.
Like this:
Volt steps = Voltage range/2^bit resolution (quantization level).
Our example as above:
4.768uV = 5000mV / 2^20 (1048576)

Higher bit resolution means a much smaller fraction of measurement graduations. But this also takes more time to resolve. This high resolution is intended for very small signal changes moving slowly.

Try this Wikipedia link:http://en.wikipedia.org/wiki/Analog-to- ... _converter

I hope this helps.

Best regards,
Michael

David · Post by **David** » Mon Jul 02, 2007 2:39 am

I read that, using the +/-39mV range and the longest conversion time, the ADC-24 datalogger has a 17bit resolution. Does that mean that it can measure signals with a 0.6E-6 V precision?
If so, what precautions should be taken during the measurements?
(envisaged application: measurement of Hall generator signal, typical signal amplitude <100E-6 V)

I also notice that, up to 313mV range, the number of precision bits increase by 1 when the range gets twice bigger. What is then the merit of using the smallest range?

ziko · Post by **ziko** » Mon Jul 02, 2007 4:59 pm

Hi and thank you for your post.

Yes your calculation is correct this is the resolution you will be getting. In terms of precautions, you can use standard noise reduction methods like having a screened twisted pair cable.

I do not fully understand your last query, a smaller voltage range will give you a greater resolution.

Kind regards

David · Post by **David** » Tue Jul 03, 2007 12:07 am

In the manual of the ADC-24, there is the following table, for 660ms conversion time:
voltage range(mV) | noise free resolution(bits)
+/-39 | 17
+/-78 | 18
+/-156 | 19
+/-313 | 20
+/-625 | 20

For ranges below 313mV, it looks like the absolute resolution (in mV) does not increase, e.g. 2*313 / 2^20 = 2*156 / 2^19 = ...
So what in which regard are the lower ranges better?

ziko · Post by **ziko** » Tue Jul 03, 2007 7:41 am

True, but if you choose +/-156mV you can then have a conversion time of 340ms.