ADC accuracy

Hi Guys, i am using STM8S003F3 controller and i want to know that how can i achive 0.1v accuracy or at least 1v accuracy.if my ADC resolution is 1023 and my total count is 2000 and Vref is 5v.
Is it possible to achieve it without changing in hardware ???????

i want to know that how can i achive 0.1v accuracy or at least 1v accuracy.

Click to expand...

Not sure what this means. Do you want a range of 0 - 1V with absolute accuracy of .1V ?


Your ADC absolute accuracy controlled by Vdd accuracy and of course the ADC errors.

10 bits, 5V Vref implies ~ 5 mV / LSB.

So its good to .1V, and its range 0 - 5 Vdd.

But you have to do an error budget analysis to see what total absolute error
is.

If you use a "normal" 5% regulator for the Vdd supply chip then your absolute accuracy
laregly dominated by this.

Regards, Dana.

Hi,

The ADC accuracy is specified in the datasheet.

And for the rest I recommend to read through your own thread.
It won't help to start several threads asking for assistance but ignoring the answers.

Klaus

danadakk said:

Not sure what this means. Do you want a range of 0 - 1V with absolute accuracy of .1V ?


Your ADC absolute accuracy controlled by Vdd accuracy and of course the ADC errors.

10 bits, 5V Vref implies ~ 5 mV / LSB.

So its good to .1V, and its range 0 - 5 Vdd.

But you have to do an error budget analysis to see what total absolute error
is.

If you use a "normal" 5% regulator for the Vdd supply chip then your absolute accuracy
laregly dominated by this.

Regards, Dana.

Click to expand...

what is error budget??

When you have a signal path it is subjected to errors like :

1) PSRR, effects of power supply noise on amplifiers in signal path
2) Offset errors, voltage and current, of amplifiers in signal path
3) Gain and non linearity errors of amplifiers/attenuators in signal path
4) Noise sources, 1/f, popcorn, surface state noise, added into the signal


5) T and V effects on signal path components
6) A/D component errors, its Vref, its INL and DNL.....

So you take these errors and by superposition add them all up (noise
added RMS) and that is your total error. Note when you calc the errors
you normalize them all, say to ppm or LSBs, so you can add them up.


Attached an ap note of basics (more on web).


Regards, Dana.