0
\$\begingroup\$

I'm using AD7190 with a 4-wire load cell to build a pressure acquisition board. From ADI reference design(CN0102), I understand that connect sense+- lines of a 6-wire load cell to the ref+- pin of AD7190 create a ratiometric configuration which eliminates the need for a precision reference. However, I only have 4-wire load cell. The CN0102 points out that due to lack of the sense line, 4-wire system is not completely ratiometric.

I want to achieve relative high precision and >100Hz date rate, so I config the AD7190 to 150Hz output data rate.

Whether a precision reference is needed for my 4-wire load cell application? If I use a precision reference, is my performance giong to be better or worse?

\$\endgroup\$
1
  • \$\begingroup\$ I would expect other error sources in the load cell could be just as big or bigger. Also if you want meaningful data at 150Hz, you should sample faster (3kHz at least IMO), and attend to aliasing issues. My feeling is you can benefit from a precision reference even in a ratiometric measurement, as it can help achieve good stability in excitation V or I (in combination with usual filtering). \$\endgroup\$
    – Pete W
    Feb 3, 2021 at 15:29

1 Answer 1

2
\$\begingroup\$

Just make the cables from your 4-wire load cell as short as possible and rewire from that point using 6-wire cabling to the AD7190: -

enter image description here

All that this adds to the error budget is the variable volt drop along the blue wires shown in my picture above. The shorter the blue wires are, the better the accuracy over temperature and the load cell maximum range.

Whether a precision reference is needed for my 4-wire load cell application? If I use a precision reference, is my performance giong to be better or worse?

It certainly won't hurt to use a precision reference but, a better method is to use current excitation.

You use a current source that produces the same approximate level of voltage excitation. This then keeps current flow through the blue wires constant so it's one less variable error to worry about AND, if your load-cell is not a full active bridge you will certainly get improved linearity when using a current source. See this table for the relative formulas: -

enter image description here

Table taken from this document. See also this document that makes linearity comparisons when exciting with constant voltage and constant current: -

enter image description here

\$\endgroup\$
2
  • \$\begingroup\$ Thanks. My load cell is full-bridge, is that means voltage and current excitment practically the same for me? Another is that my load cell has 350Ohm of output impedence, did I need to enable buffer inside the adc? \$\endgroup\$
    – livecf
    Feb 4, 2021 at 10:42
  • \$\begingroup\$ No, current excitation is better - read the first table and my words before it. I don't understand your 2nd question i.e. what buffer are you talking about and has this question got anything to do with the original question? \$\endgroup\$
    – Andy aka
    Feb 4, 2021 at 10:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.