ADS1243 full scale range problems and software trick to double input range

fbj

Other Parts Discussed in Thread: ADS1243

Hello, doing some experiments around a suggestion I had from another member about Vrefernce issue on ADS1243 I found that the analog input range is half of what the datasheet should specify for my design. I also found a turnaround for this problem.

PROBLEM

In my application ADS1243 is powered at 4.85V and its Reference voltage is: Vref+ = 3.3V & Vref- = 0V ; Software settings: RANGE = 1 , UNIPOLAR output format

with these conditions the datasheet suggests Vref =3.3V and input full scale limits should be ADC_FSR = Vref/2 = +1.65V and -1.65V (differential inputs)

Since in my design the negative analog input is connected to a precision 1:2 divider of the full scale reference = +1.65V according to the datsheet specifications I should be able to measure a voltage applied on the active positive analog input that could span from 0V to the 3.3V full scaale refence.

In the real prototype the ADC produces output code 0 at 1.65V positive input , and code 161775351 at 3.3V (full scale) positive input, while any input voltage below 1.65V obvioulsy saturates the output code at 0.

QUESTION

The ADS1243 datasheet is maybe a little confusing about the full scale reference and usage of RANGE and UNIPOLAR/BIPOLAR software options. Would anybody be so kind to clarify the issue and telling me what is wrong in my calculations?

TURNAROUND

For the moment I am running after the problem by taking two separate measurements inverting the positive and negative analog inputs, in order to choose the right half of the full scale.

Thanks in advance!

here is the original forum discussion I am working around:

http://e2e.ti.com/forums/t/12960.aspx

over 15 years ago

0 Joseph Wu over 15 years ago

TI__Guru* 97775 points

Here's a quick explanation of the RANGE bit and the unipolar/bipolar modes of the ADS1243.

First, for the RANGE bit, this controls the method of sampling of the reference in the ADC. This has the effect of changing the apparent size of the reference and is used to change the full scale range of the ADC from +/-VREF to +/-(VREF/2).

The unipolar mode is just that, it assumes a unipolar signal (typically with the negative input tied to ground). When in unipolar mode, it only measures positive input signals and adjust the output data so that the positive full scale is FFFFFFh.

In unipolar mode, the full scale range is from 0V to the positive full scale according to the RANGE bit.

In summary:

RANGE=0, Bipolar
Full scale=+/-VREF, 800000h to 7FFFFFh in two's complement

RANGE=0, Unipolar
Full scale=0 to +VREF, 000000h to FFFFFFh

RANGE=1, Bipolar
Full scale=+/-(VREF/2), 800000h to 7FFFFFh in two's complement

RANGE=1, Unipolar
Full scale=0 to +VREF/2, 000000h to FFFFFFh

In your case, you have RANGE=1, Unipolar.

Your VREF is 3.3V meaning your Full scale is 0 to +1.65V. When the positive input is at 3.3V and the the negative input is at +1.65 (from your voltage divided input), the ADC reads near full scale (I think you added an extra digit in your decimal data). When the positive input is 1.65V, the negative input is also at 1.65V and you read 0V. When the positive input is less than 1.65V, the negative reading is clipped to 0V.

Joseph Wu

0 fbj over 15 years ago in reply to Joseph Wu

Prodigy 240 points

This is really clarifying the situation even for people like me that is not familiar with sigma-delta converters.

What was confusing on the ADS1243 datasheet is the "Full scale input range" data limits on the electrical characteristics table of page 3: in fact the datasheet says that full scale input range @ RANGE = 1 max limit is +/-Vref/(2*PGA)

What is ambiguous is the +/- that might induce to think that the full range spans from -Vref/2 to +Vref/2

TI could add a comment to remove the ambiguity in future datasheet revisions thus helping neophytes to better understand.

Thanks!

Data converters

Data converters forum

ADS1243 full scale range problems and software trick to double input range