high accuracy gain stage (op-amp)

[danial_1860]

hi guys, is there any body who can help me designing a very accurate gain satge, which gain should be 8±0.001, it means error value must be less than 0.001. input voltage is 125 mv and therefore te output must be 1 volt. please help me, it's an emergency case!!

 

[scanman]

Hi danial_1860,

First you will need to select a high quality OP amp with good linearity and thermal stability. Good thermal specs on the resistors as well. You could use a simple inverting design with a variable feedback resistor to calibrate the exact gain.
Example using first circuit in the op amp doc attached:
R1 = 10K
R2 = 75K in series with 10K multiturn potentiometer.

Cheers,
Scanman

 

[keith1200rs]

You will need an opamp with very high open loop gain and autozeroing. Depending on you temperature range you need to choose you resistors very carefully. Any trimmer needs to be tiny relative to the fixed resistors to ensure its temperature coefficient doesn't affect the results. You are likely to need to use 0.1% resistors to get a decent temperature coefficient.

Keith

 

[danial_1860]

First you will need to select a high quality OP amp with good linearity and thermal stability. Good thermal specs on the resistors as well. You could use a simple inverting design with a variable feedback resistor to calibrate the exact gain.
Example using first circuit in the op amp doc attached:
R1 = 10K
R2 = 75K in series with 10K multiturn potentiometer.

@ scanman: thank u. the problem is the desired op amp u are talking about, I need to find that of course. it must be CMOS based and its f(3dB) is 100 MHz.can u please give me an example of such op amp.

Added after 2 minutes:

@ Keith: thank u. the problem is the desired op amp u are talking about, I need to find that of course. it must be CMOS based and its f(3dB) is 100 MHz.can u please give me an example of such op amp.

 

[FvM]

it must be CMOS based and its f(3dB) is 100 MHz

What comes next? E.g. on-chip resistors?
The CMOS OP offset voltage will already dwart your accuracy requirements and the noise involved with 100 MHz bandwidth
will exceed the permissible voltage error. I wonder if you're aware of analog electronic basics?

To determine if the design is feasible, you should give a full specification.