ADS1216: The on-chip 2.5V refference is not stable

Hi user4904810,

Welcome to the forum! The best way to purchase TI devices is either directly from TI, or through one of the major distributors.  There is always some risk when purchasing devices from other sources.

As far as the voltage you are measuring, the datasheet specification is 2.4V minimum to 2.6V maximum as shown on page 4 of the ADS1216 datasheet in the electrical characteristics table.  Although your voltage appears to be on the low side, it is within specification.  It is not clear if you are using a calibrated voltage meter, and if there is any error associated with the meter itself.

Can you share your schematic as to how you are using the device?  Reference stability is a factor of AVDD source, proper grounding and loading of the reference output.

Best regards,

Bob B

I'm sure of that there is no problem with my voltage meter,AVDD source can't be more stable.Actually,these circuit boards have been produced a few months ago, neither of them had a stable on-chip 2.5V refference.But a few days ago,I found one of them has stabilized the on-chip refference.The voltage is 2.495V.

And I have another problem that the AD sampling results slowly drift downwards even the refference is stable.The circuit board is  used to collect the temperature value with PT100.The circuit schematic is as follows,but it's not compelet.

I used a resistor instead of PT100 for testing.All of these components are high precision - high stability ones,but the conversion result is slowly dropping down from powered on.And it seems like,the degree of declining depends on the time interval between the last time powered off and this time powered on,the longer time have been,the greater defference I have got.I have measured the voltage across PT100，it has no change during this time.

Hi user4904810,

First let me explain why I mentioned the meter.  Most handheld and even bench meters will average a number of samples before displaying the data.  If there is an AC component, such as an oscillation, you may not be able to know that it is taking place.  An oscillation can appear as a lower DC voltage.

If you have a number of boards, and some are stable while others are not, you may want to discover why that is the case.  You may want to check for board cleanliness.  Flux residue can create leakage paths.

Your diagram does not show a differential capacitor between AIN0 and AIN1.  This capacitor is very important to limit the differential voltage created by the common-mode filters.  The differential cap should be at least 10 times greater in value as compared to the common-mode caps.

Do you have the buffer turned on or off? For this type of measurement it would be best to have the buffer on to limit the affects of lower input impedance.  However you must make sure the voltage being measured is within the correct common-mode input range for the ADS1216 when using the buffer.

The ADS1216 is a rather old design.  Have you considered using a more recently developed device such as the ADS124S08?

Best regards,

Bob B

The values I had gotten are just like this

Hi user4904810,

Unfortunately you are much more familiar with your system design than I am.  For example, I do not know what the X axis represents.  I assume that it is number of samples, but I have no idea what the time frame is between samples.  It would be helpful to know the register configuration settings as well.

What resolution are you hoping to achieve?  As to initial startup, you will have some heating of the device until the device temperature stabilizes.  If you are using the IDAC output of the ADS1216, this will directly relate to the reference.

As I mentioned previously, the ADS1216 is an older device.  It is now approaching 20 years since it was designed.  I would highly suggest using one of our more recent devices that include a much improved reference.  You should take a look at the ADS124S06 or ADS124S08 devices.  You should also take a look at using a ratiometric measurement method where the reference and excitation is the same source.  This approach limits the noise and drift of the measurement.

Best regards,

Bob B