We are Using ADS1232 ADC for our application. We are facing the problem that, *During the power on condition, the ADC value is zero for few seconds.*After that the ADC value which we are getting from ADS1232 is not the actual value. It takes sometime to settle down. After that it is showing the actual count.
Welcome to the forum! Can you send me your schematic? If you have a large cap across the inputs it may take a long time to settle. What type of sensor do you have at the input? Is your reference stable at power-up? How are you determining when to read the data? Are you polling DRDY, or is your system interrupt driven? Do you do an offset calibration at power up?
please check the following excel document to check the ADC count variation.
This appears to be a cap charge/discharge. It could be on your reference, or your inputs. Please send me your schematic. Layout could also be a factor. Do you have a ground plane or just ground traces? Are you using any inductors in your supply or ground connections?
We are using the ADS 1232 for our weighing applications. This is giving the stable ADC counts for 10 SPS but for 80 SPS ADC counts are still oscillating. We are also implemented the Averageing of ADC counts but still it was oscillating. Please give me the suggestion for this.
Inherent to the ADS1232, you will lose about 1.5 bits due to noise moving from 10sps to 80sps. So it is quite possible that you can achieve stable readings at 10sps and less stable readings at 80sps. Bottom line is you will not see the same level of performance at 80sps that you did at 10sps.
So the question to you is, what level of performance are you expecting to see? Is this a reasonable expectation? And what level of instability are you currently getting?
As far as reducing noise, all of the things I posted before should be considered.
Thanks for your valuable reply.
1. You told that 1.5 bits would lost due
to noise while we were moving from 10sps to 80sps. But we are using only 17-
bits for our application from the 24-bit ADS 1232. Remaining 7-bit (LSB) we are
2. Now by using these 17-bits we
get ADC counts, these ADC counts will use for weight calculation.
3. If, we configure the ADC with 10sps
then weight variations are happening even if I place small weight (<10 grm)
on load cell . But the same case is not happening with 80sps. If small weight
placed on the load cell (ADC count variation and Weight variations are not
there). If large weight (>0.5 kg) placed then only ADC and Weight variations
occured but it took some time to stabilized.
we are expecting the
weight variation for (<10 grm) with 80 sps.
I will comment each item as you have numbered them in your last post.
1) The concept of noise can be confusing. If the PGA is set to 128 and using 5V for both supply and reference, the absolute best the converter can do noise wise is 18.4 bits (Table 1 of datasheet) at 10sps. That is just the converter itself with shorted inputs and no external connections. It is also assuming a very noise free supply and reference. The absolute best the converter can do at 80sps is 17.2 bits (Table 2 of the datasheet.) If you are on the edge of stability with your system at 10sps, then it will only get worse at 80sps.
3) You are now confusing me. What you are stating here is not what you said in an earilier post. In the earlier post you said you were seeing more fluctuation at 80sps than at 10sps. That is what I would expect. Now in this last post you are saying the opposite. First of all, the small weight will be much more difficult to measure than a large weight because the weight itself is closer to the noise. The bottom line is I need precise information.
I asked for your schematic early in this posting. If you have long settling times, it sounds like you have heavy filtering on the input. Also, for me to better understand what you are trying to accomplish I need to know the maximum weight (grams/kg) you can apply to the load cell, the mV/V output of the load cell, your voltages you are using ( supplies, reference and excitation), whether or not you are using the excitation as the reference supply, and the gram resolution you want to achieve.
Sorry, I think You couldn't understand my previous post.
Now the problem is,
1. With 10 sps, even If I place a small weight on the loadcell it is getting the corresponding weight with out variations. If I place small weight on the loadcell with 80 sps it is not getting that corresponding weight, but if I place large weight it is getting the weight reading correctly but it takes some time to settle.
2. And also with 80 sps, if I press the load cell and again released it but it takes sometime to come back. For example, Initial weight is 1.000 kg, Now I press the loadcell the weight goes to 2.000 or2.500 kg (for example). Now I release the load cell, normal case it will comes to 1.000kg but here 1.000 to 1.300 kg (nearly) varying and after sometime it comes to 1.000 kg. But the same problem is not happening with 10 sps.
Please give me your mail ID then I will send the schematic. My mail ID is vijay 'at' lcsembedded.com.
From your valuable replies we solved the stabilization problem of ADC counts.
But, after a long time we identified one more problem in our ADS 1232 board.
Our Schematic description:
1. We are taking the load cell output as an analog input to the ADC. Load cell gives differential output. This output was in millivolt range.
2. This differential output will pass through the common mode filter and LPF for eliminating the noise in load cell output.
3. The LPF output will be given to the Instrumentation amplifier AD8553 (RGA and RGB) with gain of 50.
4. The output of instrumentation amplifier is given to the input of ADS1232 (AINP1). VREF = 2.5V(constant) is given to AINN.
5. AINN2 and AINP2 are not used for our application. Those are simply grounded.
6. Speed is 10sps. TEMP pin is always grounded.
7. We are using the common ground for both load cell and ADS1232 ground.
After power on the indicator, The ADC value will take some time to settle.
While testing our indicator, At initial time the ADC value is 20080 and this value is slowly increase for some time. After reaching the 20148 value the ADC was appear to be settled.
After settling the ADC whatever the calculations we made was correct. In this case the settling time for ADC was around 10 to 15 min (we observed).
If the indicator is continuously on then we identified the +/- 10 counts variations we observed from the settled ADC value 20148. Please check the attached file.
Why it was happening? whether it may due to temperature variation or any environmental noise?
What is the need of TEMP pin ADS1232? previously, we simply connect to ground. If we connect this to 5V what will happen? whether this pin can able to eliminate temperature drift? please give us the solution for this.
Thanks and Regards,
Let's start with the TEMP pin. This pin will force the device to read the internal temperature diode instead of the ADS1232 inputs. I was unable to get to the attached file, so I can't analyze your data. As far as your drift issues, you will have temperature drift of every part in your system. The ADS1232 will drift and you can evaluate the amount of drift from the datasheet. Along with the ADS1232 drift, you will have drift of your excitation voltage, reference voltage and drift of your amplifier. Some of the ADC drift can be compensated by running periodic offset calibrations. Some of the noise and drift can be canceled by using a ratiometric measurement where the excitation voltage is used as the reference voltage. This is a bit complicated by the fact that you have an external amplifier between the load cell and the ADS1232 which will add some noise and drift as well. I would start with the periodic offset calibration to see if that helps you.
The initial startup settling will be to some degree the settling of the reference and supplies. Once again, a ratiometric measurement will help with the settling of the reference.
Please check the above file which will gives the clear idea about the count drift.
Some of the ADC drift can be compensated by running periodic offset calibrations.
How we can run periodic calibration? whether by using software or hardware changes?
The offset calibration can be initiated via software by providing at least two additional SCLK pulses. See the section on Offset Calibration on page 20 of the datasheet.
All content and materials on this site are provided "as is". TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with regard to these materials, including but not limited to all implied warranties and conditions of merchantability, fitness for a particular purpose, title and non-infringement of any third party intellectual property right. TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with respect to these materials. No license, either express or implied, by estoppel or otherwise, is granted by TI. Use of the information on this site may require a license from a third party, or a license from TI.
TI is a global semiconductor design and manufacturing company. Innovate with 100,000+ analog ICs andembedded processors, along with software, tools and the industry’s largest sales/support staff.