DC Offset problem - Instrumental Amplifier

Thread Starter

rodgers3332

Joined Mar 25, 2022
4
I am currently building a circuit that takes in a ~10uV signal ranging from 1-20Hz. I am using an INA126P for the initial stage of amplification, and amplify the rest through a few stages of non-inverting op-amps. However, when I build the circuit I am getting a negative dc offset out of the instrumental amplifier of around -300mV. The gain for the INA126P is set using a resistor and is currently set to only 100v/v. An attached photo of the ltSpice model can be seen below. Am I making a mistake or is it possibly a faulty amplifier? I was also looking into solutions to fix this: could I use an AC coupling capacitor and is it possible to not effect my 1-20Hz signal input range? Any help is appreciated!
 

Attachments

Thread Starter

rodgers3332

Joined Mar 25, 2022
4
Does the input have any DC offset?

You can use AC coupling as long as it's low frequency rolloff is below 1Hz.
The input is supposed to be from an EEG headset measuring brain waves, however, for testing I was generating a 1vpp signal generator and using a voltage divider with a 1M and 10 ohm resistor to output around a 10uVpp signal.
 

ericgibbs

Joined Jan 29, 2010
18,734
hi,
If you are using LTspice, why don't you download some EEG wave files for testing.?

Is the output from the EED device differential or single ended.?
E
 

Thread Starter

rodgers3332

Joined Mar 25, 2022
4
hi,
If you are using LTspice, why don't you download some EEG wave files for testing.?

Is the output from the EED device differential or single ended.?
E
That is a good idea I will try that, thank you! The device measures the voltage from two different sides of the head at the difference is amplified.
 

MrChips

Joined Oct 2, 2009
30,618
With such low frequency signal it would be more effective to use a DC offset adjustment on the input of the opamp.
Here is an example:

1648224261625.png

btw, you don't need three stages of amplification, one stage will do.
 

ericgibbs

Joined Jan 29, 2010
18,734
That is a good idea I will try that, thank you! The device measures the voltage from two different sides of the head at the difference is amplified.
Hi,
In that case, create a simple differential signal source in LTS, drive both inputs of the INA126, and check for the original Voffset.

BTW: you can use free Audacity to create your own 'EEG' files for testing, save as wav format.
E
 

Ian0

Joined Aug 7, 2020
9,618
The usual way to reduce the DC offset is to put a capacitor in series with Rg. 470uF would give you a roll-off at below 1Hz.
Try to get a bipolar electrolytic. They are superior to the normal sort in that application.

What is the output impedance of your sensor? The INA126 at 35nV/√Hz certainly is not a quiet part. 1nV//√Hz can easily be achieved, it also quite a high 1/f noise corner for current noise, which would be a problem for high impedance sensors.
To minimise the output noise, put as much gain as you possibly can in the first stage, that avoids re-amplifying the noise of earlier amplifiers in the chain, and watch the 1/f noise spec!
 
Top