Hello all!

I have beeen working with high-precision ADCs for some time now, and one thing that I notice in almost all high precision ADcs frrom prominent manufacturers is that most of them have one or more integrated IDACs within the same single package. I do understand that they are to facilitate ratiometric measurement of physical quantities, and that ratiometric measurement has a lot of advantages. What I am not able to understand is why current sources play a monopoly in ratiometric measurement?

Aren't voltage sources as good for excitation sources in ratiometric measurement, and that the high precision ADCs may offer simple DACs to act as the excitation sources?


With best regards,
Abbas.

 

What is your specification of "high-precision ADC"?

 

Can you supply some example part numbers -- or better yet data sheets -- of the ADC's you are talking about?

 

For example, the latest one I am working with is the ADS1120.

 

The IDAC is used with a RTD (low resistance) or resistance changing element. With the current(s) set the voltage across the RTD in series with this current is a direct factor of temperature and a precision reference resistor (for the ADC reference voltage), not the current source. The ADC measures this RTD voltage and the controller scales this voltage to a temperature. Using voltage sources here would make the absolute accuracy and temperature drift of the excitation voltage a factor.

From the ADS1120 doc:
For three-wire RTD applications, the matched current sources can be used to cancel errors caused by sensor lead resistance (see the 3-Wire RTD Measurement section for more details).

 

Thanks for your response, nsaspook.

With the current(s) set the voltage across the RTD in series with this current is a direct factor of temperature and a precision reference resistor (for the ADC reference voltage), not the current source.

That's the same with a voltage source as well!

Using voltage sources here would make the absolute accuracy and temperature drift of the excitation voltage a factor.

Nope, not possible! We are talking about ratiometric measurement, here. source disturbances are eliminated in a ratiometric measurement, irrespective of the nature of the source.

 

Thanks for your response, nsaspook.
Nope, not possible! We are talking about ratiometric measurement, here. source disturbances are eliminated in a ratiometric measurement, irrespective of the nature of the source.

Look at the three-wire RTD circuit again. It's ratiometric with 'matched' sources. On-chip matched current sources are much easier to fabricate and it's usually better in these types of applications to generate a voltage reference from a current source.

Current sources are a versatile means of forming voltage references. Why not just use a voltage reference? With a current source, a single resistor provides a programmable voltage source of any value. Low voltage references are often needed, and with this approach, it’s as easy to get a 1mV reference as it is to get a 10V reference. Also, the voltage can be referenced anywhere—to the positive rail, the negative rail, or floating anywhere in between.

http://www.ti.com/lit/an/sboa046/sboa046.pdf

 

Look at the three-wire RTD circuit again. It's ratiometric with 'matched' sources. On-chip matched current sources are much easier to fabricate and it's usually better in these types of applications to generate a voltage reference from a current source.


http://www.ti.com/lit/an/sboa046/sboa046.pdf

Thanks for your response, again, nsaspook.

But those matched sources help in eliminating the errors due to RTD leads.

Also, how can current sources be easy to fabricate, when a current source basically uses a curent stabilizing circuitry in conjunction with a voltage source?

 

Thanks for your response, again, nsaspook.

But those matched sources help in eliminating the errors due to RTD leads.

Also, how can current sources be easy to fabricate, when a current source basically uses a curent stabilizing circuitry in conjunction with a voltage source?

http://web.iku.edu.tr/courses/ee/ee525/PDF/EE525_Exp1_Current_Mirror.pdf