ADS1235: ADS1235 not running at desired speed.

Hi Nimit,

I still don't have all of the information regarding your setup.  For example, do you have any connection from your micro to the CLKIN, START, RESET and PWDN pins?  Or are you just using the default conditions of the EVM hardware?

The CLKIN pin defaults to low, so as long as nothing is connected to this pin then the internal oscillator is connected and the typical frequency is 7.3728MHz.  All of the timing table information is based on that frequency.

There is a calculator tool for the ADS1260-61 which uses a similar timing as the ADS1235.

https://www.ti.com/tool/ADS1261-EXCEL-CALC-TOOL

I have added a screen shot below of the timing for the settings you are using:

If you monitor DRDY with an oscilloscope of logic analyzer you can verify the timing on the EVM for the settings you are using.  The result should match what is shown above.  If the ADS1235 data rate is as shown, then there is something at issue in the reading loop of your micro code.

Bests regards,

Bob B

Hi Nimit,

100,000 counts is difficult to achieve let alone 500,000.  The ADS1235 (and any other ADC) will be limited based on the conversion noise of the ADC.  This will include the PGA and quantization noise.  To achieve the lowest noise with the least amount of drift issues will require that the measurement be made ratiometrically where the excitation of the bridge and the reference voltage to the ADS1235 are the same.  The excitation/reference noise and drift cancel using this configuration.

Looking at front page of the ADS1235 datasheet you will see a bullet item where it is stated that you can achieve 120,000 noise-free counts with a 10mV/V input and 10sps data output rate.  The application information to achieve the 120,000 noise-free count is given in the ADS1235 datasheet in section 9.2.

It is possible to achieve greater resolution based on data rate, gain and the digital filter being used.  To achieve 500,000 noise-free counts will depend on the sensitivity of the bridge.  A 1mV/V sensitivity will have fewer noise-free counts and 3mV/V will have more noise-free counts.  The reason for the difference in counts is due to the full-scale output of the bridge relative to the full-scale range of the ADC.

These calculations are of course only with respect to the ADC and ideal conditions.  The calculations do not take into account any effects of the sensor itself or noise pickup due to EMI and RFI.

Let's look at a different perspective,  If the sensitivity of a load cell is 2mV/V and the excitation voltage is 5V, the maximum voltage output of the load cell is 10mV.  If the load cell is 50kg and you want to have stable and accurate measurements of 0.1g, then you would need 500,000 noise-free counts.  What you are asking of the ADC is to resolve 10mV/500,000 counts or 20nV noise-free.  That is 0.00000002 V.  If you look at the noise table in the ADS1235 datasheet you need to find the data rate, filter and PGA setting that can resolve to this level such as 2.5sps, Sinc4 filter and Gain of 128.  There may be some other ways of improvement such as using a moving average at higher data rates that might also give you similar results, but remember this is only the best that the ADC can do and is not a reflection of the actual results you can achieve.

If your sensitivity is lower, such as 1mV/V, the voltage resolution requirement would become 10nV noise-free.  So it is not easy to give a quick answer on how to make settings on the ADC give you the result you desire.

What I have seen in most systems are issues with external noise primarily picked up in cabling.  Power line-cycle noise is the most common.  This requires that cabling from the sensor to the ADC be shielded and properly terminated.

Best regards,

Bob B