ADS1262: low range, high precision RTD

Hi Maciej,

Welcome to the TI E2E Forums!

Are you going by the 3-wire RTD circuit (with a high-side reference) schematic that is in the ADS1262 datasheet? If so, I might recommend using the typical 4-write RTD circuit (with a low-side reference) shown below, as this doesn't require the bias resistor and allows for more IDAC headroom. The high-side reference configuration is only beneficial to the 3-wire RTD, where there are two IDACs.

From initial inspection of your calculations, it looks like you are utilizing the full input range of the ADS1262 (though perhaps you might make some changes if you decide to remove the Rbias resistor). Your reference voltage (i.e the maximum differential signal you can measure) is 2.35V, and your input voltage is (2.237 - 1.65) * 4 V/V = 2.348. So you're utilizing 99.9% of the ADC's input range with your current configuration.

Best regards,
Chris

Thank You Chris !

Ok, I will rearrange my circuit, thanks for this tip !

I understand that I am using 99% of ADC's input range and can measure Vrtd (0 - 0.587)V. But is there any way to measure zoomed range not from 0V but from 0.519V - 0.587V ?

My input signal will vary only in this range, so now I am using only (0.587-0.519) / 0.587 = 12% of ADC full range, because RTD won't ever drop below 1038Ohm

Hi Maciej,

I suppose it's possible to do what you're asking, but I'm not sure of the best way to go about it such that it would actually prove you with better results. Besides, the main benefit of using the 32-bit ADS1262 is that it provides a very high dynamic range for measuring small changes in input voltages for cases just like this.

The two ideas that initially come to mind for me are:

1. Subtract ~0.5V from both inputs (using a dual amplifier circuit) so that you can use a higher PGA gain on the ADS1262. The problem here is that you'd probably introduce additional errors and noise from the amplifier that would be far worse than simply using the ADS1262 as is. (i.e. There is likely to be a subtraction mismatch on each input).
   
   
2. Create two reference signals that are approximately 2.2 V and 1.6 V in amplitude. Then measure the low-side of the RTD with respect to the 1.6V (so that the differential voltage is small and the PGA gain can be increased) and then do that same for the high-side of the RTD and measure with respect to the 2.2 V reference. Again, the problem here is that amount of error in each of these reference voltages will be different, the RTD measurement will no longer be ratiometric, the reference voltages may drift independently, and by not measuring the low-side and high-side at the same time it's possible for drift errors to have a larger effect on the temperature accuracy.

Even though you're only using 12% of the ADC's dynamic range, I think it would still be my recommendation to use the 4-wire ADS1262 circuit as-is. By going to a PT1000 (over a PT100), you've already made a significant improvement to the RTD's temperature sensitively (about 3.9 Ohms/C instead of 0.38 Ohms/C for the PT100) in your temperature range of interest.

Best regards,
Chris