This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ADS1232 Drift

Other Parts Discussed in Thread: ADS1232REF, ADS1234

Hello,

I am using the ADS1232REF together with a loadcell to try to make an accurate scale, similar to what you purchase for use in a kitchen.


I have two loadcells at my disposal, both are rated for 20kg and are 4 wire. One of the loadcells has shielding, the other does not. Using the ADS1232REF in scale mode, calibrated with 500g I get a much more stable reading on the shielded loadcell than on the non-shielded loadcell.

On both loadcells I see a constant drift downward though. I'm talking on the order of magnitude 0.1g/min.

This drift is something that is not present on a low-cost scale and I'm wondering how I can get rid of it.


Thank you for your help.

Nick

  • Hi Nick,

    Welcome to the forum! There are a couple of things that direcly effect the result on the display.  One is noise and the other is drift.  To get the lowest noise and drift you would need to make sure that the connections of the load cell create a ratiomentric measurement.  In simple terms this means that the same voltage used to create the excitation of the load cell is also the reference used for conversion.  With the ADS1232REF you should use the +5VA to excite the load cell and set as the reference at SW7.

    Another factor is the noise.  Using the slowest data rate (10sps) gives the best noise performance.  You can measure the device noise by using the shorting jumpers J7 and J8 that when connected apply a short on the inputs to mid-supply of AVDD.  When connecting a load cell make sure that these jumpers are removed (common mistake).  You also want to make sure that the highest PGA (128) is being used.  I don't know the sensitivity of your load cell, but I will guess that it is 2mV/V output at full-scale.  For a 20kg load cell this would be a maximum output of 10mV for 5V excitation.  Your calibrated weight is 1/40th of full-scale or has an output of only 250uV.  Notice on page 34 of the datasheet that best case scenario is a total of 177,385 noise free counts throughout the range of your load cell.  Depending on how the ADS1232REF is setup and the amount of external noise you may be closer to 1/2 of that, or 88,700 counts. This will give a noise free weight of about 1/4 gram.  However, calibrating to 250uV with any appreciable amount of noise will not give you much calibration margin.

    There could be a number of reasons for seeing drift.  One might be drift of the ADC (which can be calibrated out by running a SELFOCAL offset calibration.  Another source may be related to self-heating of the load cell.  A third reason can be from external noise that appears as drift.

    The best thing to do first is to determine how much noise you are seeing in your system.  Run the ADS1232REF in analysis mode to see how much noise you actually have with the load cell connected.  You may need to swap out the 0 ohm resistors (R23 and R24) with some resistance to add input filtering. If the resistor is too large you will see analog settling of the input filters when a weight is applied to the load cell.  You could try something in the range of 100 to 1k ohms.

    A couple of things to remember is you need to consider the best noise free resolution from calculation, and you need to reduce the noise as much as possible prior to calibration.

    Best regards,

    Bob B

  • Thank you Bob,


    I am using the 5V source on the board to excite the loadcell.


    When I make a noise measurement it is around 380n.

    How do I know if I should add resistance to R23/24?


    My loadcell is actually at 1mv/V. I'm looking for the notes you're talking about on page 34 but I don't see a page 34 in the datasheet for either the ADS1232REF or the ADS1234.


    Is the SELFOCAL referenced on page 20 of the ADS1234 datasheet? Is there anyway to perform the SELFOCAL in scale mode? Also I have not found any settings to adjust the PGA in scale mode. Is that possible as well?

    Thank you so much,

    Nick

  • Hi Nick,

    See my comments below in RED.

    Best regards,

    Bob B

    Nick McMahon said:

    Thank you Bob,


    I am using the 5V source on the board to excite the loadcell.


    When I make a noise measurement it is around 380n. BB> Was this RMS or P2P?  This added noise will lower the total number of noise free counts and resolution flicker.

    How do I know if I should add resistance to R23/24?  BB> This depends on the source of noise and frequency of the noise.  For example, if noise is created by small vibrations, cell phone activity, motors, fans, etc.,  you may be able to filter out the noise based on setting the cutoff frequency for an input filter stage.  Currently the resistors are 0 ohm, so the only resistance is the load cell itself.  Any noise picked up on the wiring will not be filtered.  You would set the R value for the desired filter cutoff frequency.


    My loadcell is actually at 1mv/V. I'm looking for the notes you're talking about on page 34 but I don't see a page 34 in the datasheet for either the ADS1232REF or the ADS1234. BB> Sorry my mistake (typo) it should have been page 24.  Based on the noise information I'm guessing that the noise is P2P which will lower the noise free resolution to about 17.5 bits.  By rough calculation your noise-free weight will be about 1 gram for the 20kg load cell.


    Is the SELFOCAL referenced on page 20 of the ADS1234 datasheet? Is there anyway to perform the SELFOCAL in scale mode? Also I have not found any settings to adjust the PGA in scale mode. Is that possible as well? BB> The offset cal can be issued by going to the CONFIG (pressing the two left most buttons at the same time) and selecting OCAL from the menu selections on the ADS1232REF.  The PGA is set to 128 for scale mode.

    Thank you so much,

    Nick