Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

current excitation for sensor

Status
Not open for further replies.
Accuracy: designing for accuracy better than 0.025%, I don't mind about the linearity now because that can be solved using interpolation.

concentrating on the stability of the circuit. That's why posted the sensor excitation circuit and ADC reference circuit,
 

Hi,

Maybe you mix accuracy with precision?

You can't get 0.025% accuracy with all what you wrote before.
Even 0.025% precision is hardly possible.

And can you please explain how you want to improve linearity with interpolation.

Klaus
 

Can we use a single voltage reference for both current excitation and reference voltage for adc as shown in the fig below.

Kindly share your comments.
 

Attachments

  • Schematic Prints1.pdf
    9.6 KB · Views: 48

Hi,

I don't know what you want to achieve with your circuit.
Please tell us your ideas.

My view:
You overcomplicate things ... and sadly this way you reduce performance.
* Using the same reference for both sensor and ADC has some benefits:
- initial accuracy as well as drifts of the reference is cancelled out, thus you may use a really cheap reference.
- you just have to focus not to introduce much noise and that both sensor and ADC really see the same signal.

Thus the use of "U$A" does nothing but introduce errors. It's offset drift and noise is sent to the ADC only and thus can't be cancelled out in ratiometric style.
The same applies to "U$B"(because it's errors ar only sent to the sensor).
The use of R2, R3, R4 do nothing useful but introducing initial errors, drift and noise to the ADC only and thus can't be cancelled out.
And it makes the signal high impedance = weak ... and thus prone for external influence
It reduces VRef to the ADC, which makes no sense, since the noise of the ADC is higher than it's resolution.

U$B makes the sensor to be supplied with "constant current" (and not constant voltage).
Is this really a benefit?
Here you need to consult the datasheet if there is some benefit ... and whether this benefit is not cancelled out by the additional errors like noise and drift which is introduced by the circuit.

******
My recommendation: keep it simple and use the benefit of a ratiometric system:
The simplest way with very good performance is to use a well low pass filtered signal derived from VCC
A simple zener could give additional benefit to decouple from VCC
--> R, zener, C
Use the capacitor voltage as sensor supply.
The use 4 identical R C low pass filters to feed the signals to the ADC.
* capacitor_GND to ADC_Ref-
* capacitor_+ to ADC_Ref+
* sensor_out- to ADC_in-
* sensor_out+ to ADC_in+
--> 4 x R, 4 x C
(If the wiring to the sensor is lengthy, then consider to use kelvin wiring: feed back the sensor supply to the ADC with extra wires. No additional electronic components needed.)

That's all. And I'm sure it gives more accurate and more precise ADC results than your expensive and complicated circuit.

Klaus
 
I have drawn some configuration. If this is also wrong, can you hand sketch and show me.

Where can i connect REF+ and REF-?

Sorry for it.
 

Attachments

  • IMG_20190924_155546.jpg
    IMG_20190924_155546.jpg
    500.4 KB · Views: 63
  • IMG_20190924_155557.jpg
    IMG_20190924_155557.jpg
    500.8 KB · Views: 61
  • Schematic Prints2.pdf
    8.4 KB · Views: 48

Hi,

First you should clarify if the sensor needs constant current excitation.

If constant voltage excitation is OK, then try Fig1.
But as already written..I recommend to add 4 RC filters in all signals to ADC.

I generally use appropriate filters with analog signals. Especially with ADCs, because if there are alias frequencies you can't safely filter them out on digital (software) side.

Cost and effort is low .... and they introduce about no error.

Klaus
 

sensor which I have is constant current excitation type.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top