• Resolved

# ADS1261: Choosing the correct Rref for ratiometric measurement

I am designing an RTD measurement device based on ADS1261. The excitation current requirement range from 100nA to 1mA, I understand that I need an external current source to support this full range.

I need to choose the Rref for ratiometric measurement to cover this range. From the datasheet and the excel sheet, Vref should be 0.9 < Vrep < 5.

That would mean that I need to have a resistor selection network with different resistor values for each range. Would a TMUX6104 work for this, I chose it due to its low leakage current ?

Does a 10Meg rref make sense @100nA range ? This will evaluate to 1V for Vref.

~~ Laziness is the mother of IoT ~~

• Hi LazyHD,

The TMUX6104 looks like it should work for your application.  Note that the operating voltage for the mux requires at least 10V unipolar or +/-5V bipolar as a minimum.

Note too that the IDAC output has a compliance voltage where the total voltage drop created from the current source cannot exceed AVDD-1.1V.  So it would not be possible to create a reference voltage of 5V using the IDAC, for example.

As to using a 10M Ohm resistor, consider the effect of noise using this large of a resistance.  Also, the reference resistor should be a low-drift precision resistor.

In case you haven't seen this material I would recommend taking a look at A Basic Guide to RTD Measurements.

Best regards,

Bob B

• In reply to Bob Benjamin:

Hi Bob,

Thank you for you detailed reply. I will choose a precision 10M Ohm resistor with low ppm. Do I have any other options for 100nA excitation current ? Is there a better approach than using large value resistor ?

One more question if you don't mind, when using AC excitation with the above circuit, I would need to add an Rbias on the other end of Rref(i.e on Ain0 side) to maintain the Vcm when the current is reversed, correct ??

I will not be using the internal IDAC, I have an external current source for that purpose.

The voltages for the TMUX are not a problem as well, but thank you for raising this point.

I checked the basic guide to RTD, and the screenshot is from it.

Thank you

~~ Laziness is the mother of IoT ~~

Hi LazyHD,

You could use the large resistance value as a ratiometric measurement, or you could choose to use a standard fixed voltage reference and lose the benefit of a ratiometric measurement.  With a fixed voltage reference you will have accuracy issues with respect to any variation due to drift of the reference and drift of the current excitation source.  With the ratiometric measurement the drift of the excitation source will cancel as both the RTD and reference are excited by the same source.  The issue with the ratiometric measurement is the drift of the resistor used for the reference which will appear as a gain error.  So these are things you need to consider in your system design.

When using AC excitation, I personally have only used this with weigh scales and voltage excitation.  In this case the total voltage drop across the bridge sensor is switched.  For the RTD circuit, the Rref resistor is used to bias the input into the correct common-mode for the RTD input.  So you are correct in that if you switch the direction of the current, you would need to bias the other side of the RTD to maintain the correct common-mode for the RTD input.

When using AC excitation you need to be aware of your conversion timing relative to the analog settling when reversing the current flow.  For the design itself it would seem like double the effort for little gain.  What I have seen is noise plays a larger factor than self-heating unless large currents are used.  Again, what you decide will greatly depend on the overall system goals and levels of precision desired.

Best regards,

Bob B

• In reply to Bob Benjamin:

Thanks a lot Bob, all is clear

~~ Laziness is the mother of IoT ~~