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.

signal conditioning circuit

Status
Not open for further replies.

hemnath

Advanced Member level 3
Joined
Jun 24, 2012
Messages
702
Helped
61
Reputation
120
Reaction score
57
Trophy points
1,308
Location
Chennai
Activity points
6,588
WhatsApp Image 2020-07-24 at 2.56.27 PM.jpeg

Input signal of 0.2 to 100 mV is from sensor output.

I apply a linear input signal but the output is not linear and its produces noise in the output.
 

First of all if you are doing this with discrete OpAmps and discrete R's you
have a problem getting adequate CMR.

1595587127556.png



Do you have Oscilloscope screenshots of input and output ? I assume your A/D has a
diff front end ? Does it have any recommended input constraints ? Part number of A/D
and OpAmps.

Bypassing on OpAmps a combination of ceramic disk (.1, .01, and bulk polymer tant preferred) ?

1595587380106.png


Use a laser trimmed IA from a vendor goes a long way to fixing CMR problems.



Regards, Dana.
 
As an aside if you are using a micro as well there are processors out there
with del sig A/D's in them up to 24 bits. And some with PGAs in front end and
reduced accurate Vref that can handle small signals. Also ones who can CM
outside supply rails for applications like current measurement of a load.

Cypress PSOC and TI come to mind.


Regards, Dana.
 
Sorry. I cannot change the hardware now. Could you help for the existing circuit?
 

Note the input of the DelSig has a switched C input sampler. That in
turn cause transients at the OpAmp output, you have to allow for some
settling time, or "throw away" the first sample. Settling time to 18 bits
is not necessarily small, so be aware of that.

The GBW of the 6064 is quite low, so its settling time might be suspect.
I would do some testing on that to see if its a problem. At the very low
SPS rate of the A/D may not be an issue. But check.


Regards, Dana.
--- Updated ---

Scope captures of noise ? Also spectrum of noise, eg. Spectrum Analyzer
screen capture if you have one.

Also overall schematic.


Regards, Dana.
--- Updated ---

What are your goals ? Accuracy in bits (absolute or relative), CMR, Speed, Noise, Temp range......?


Regards, Dana.
 
Last edited:

The circuit schematic is incomplete.

What kind of sensor connected? Do you understand that input common mode voltage is copied to the ADC, thus the nput must be biased appropriately.

Also no info about power supplies.

How do you determine that "the output is not linear"?

The ADC input sampler will alias wideband OP noise into the signal band, thus an RC low pass is suggested in any case, review the datasheet in this regard.
 

Hi,

Not sure if non-linear relates to LMC6064 Vos of typical 0.1 mV to maximum 1.2 mV across temperature range, if the minimum Vin is 0.2 mV.

Buffering the difference amplifier (in amp) inputs with voltage followers probably won't help, but worth suggesting.

Maybe the non-inverting inputs used to sense Vin need a series resistor of a value equal to parallel value of 110k ll 20k for input bias current cancellation, then more buffers might be convenient for input/output impedance.
--- Updated ---

Hi,

Not sure if non-linear relates to LMC6064 Vos of typical 0.1 mV to maximum 1.2 mV across temperature range, if the minimum Vin is 0.2 mV.

Buffering the difference amplifier (in amp) inputs with voltage followers probably won't help, but worth suggesting.

Maybe the non-inverting inputs used to sense Vin need a series resistor of a value equal to parallel value of 110k ll 20k for input bias current cancellation, then more buffers might be convenient for input/output impedance.
 

If the CM, its max, gained up with signal, does not exceed the ADC input CM range you
do not have to do any biasing. The ADC diff front end will handle that, rejecting the CM.

The ADC has a CM range that easily handles the offset, and the offset gained up wont
cause a problem other than it is added to input Vdiff and produces absolute error.

The 6064 has a low CM range, basically 0 to Vdd/2 for 5V supply. But you should be
OK with that depending on what your CM is.

The bias current of 10 fA obviates any need of input Rs needed in NI inputs.

You do not need any buffering of the inputs. That only adds error.

But you still have to answer " What are your goals ? Accuracy in bits (absolute or relative), CMR,
Speed, Noise, Temp range......? " as that affects dong this as discrete rather than using a laser trimmed IA.


Regards, Dana.
 

the grounding and power supply sourcing and relative levels of the whole circuit will have a large influence ...
 

With respect to anti aliasing filter -




Another consideration, since you Zin of the OpAmps quite high they are susceptible to
coupling if the sensor itself is HiZ, capacitive coupling in particular, power line coupling,
like fluorescent lights, etc.. If the sensor is remote from board then certainly shielded cables
out to sensor make sense.



Regards, Dana.
 
Last edited:

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top