Shop OBEX P1 Docs P2 Docs Learn Events
ADC0831 Resolution Question — Parallax Forums

ADC0831 Resolution Question

Chris SavageChris Savage Parallax EngineeringPosts: 14,406
edited 2010-10-01 09:53 in Accessories
This was just sent via PM, and since I like my answers to questions to benefit everyone, I am going to post it here for everyone else to benefit from as well.
What is the minimum temperature difference you can detect with the ADC0831 and the LM34 circuit? What do think would be required to get more resolution? (Hint: rhymes with "more honey" ) Explain

Hello...The ADC0831 is an 8-bit ADC, so it has 256 steps for the VRef, which if it is 5V equates to about 19.5mV per step. To get more resolution you can do one of two things...
  1. Get a higher resolution ADC, such as a 12-bit unit, which we sell. More steps means more resolution, inherently.
  2. Change the VREF to be a more closed range of voltage for the LM34. The LM34 outputs 10mV per degree F, so at its maximum range of 300 degrees F it will output 3000mV, so set your VREF to 3V, or less if you never expect the temperature to get above a certain point and this will increase the resolution as well. At 3V, VREF the resolution went from 19.5mV to 11.7mV using the ADC0831.
I hope this helps. Take care.

Comments

  • John R.John R. Posts: 1,376
    edited 2010-10-01 09:53
    First: Chris - thanks for sharing! An Excelent question from you PMer.

    Converting the mV differences Chris mentions above, with a 5V reference, and 8 bits, the 19.5mV per step is just short of 2 degrees. Changing down to a 3 volt reference, you're just over 1 degree.

    Jumping to a 12 bit ADC, you get 4096 steps, which is 1.2mV with per step (about 0.1 degree) with a 5 volt reference, or 0.7mV (a little under 0.1 degree) with a 3 volt reference.

    At the 12 bit rates, the "precision" is probably somewhat "empty", in that there are going to be enough other factors coming into play (variations in individual sensors, connections, solder joints, wire length, etc.) that you may not be as "accurate" as the numbers show.

    How useful the refined data is depends on what you're doing. If you're looking at trends or changes with a single sensor, the higher resolutions may have more value than if you're comparing the absolute value of a number of different sensors in different locations.

    John R.
Sign In or Register to comment.