Best Way to Remove Offset from Low-Frequency Signal

Thread Starter

SSPhysicsMeasurments

Joined Feb 18, 2019
4
Hi! This is my first post, so I'll try to keep this brief as possible.

I have a small (~uV or lower) square wave signal (f~500Hz) which is on top of a much larger DC voltage (~50mV). I need to determine the amplitude of the square wave signal to high accuracy as part of my experiment.

Currently I am using a lock-in amplifier, which works well for strong signals (~uV). However the signal becomes undetectable below around 100nV. I have a DC amplifier (adjustable 100x to 100,000x gain) to amplify to the signal before I put it into the lock-in, however because of the large DC offset (~50mV) it overloads immediately.

So the question is: What is the best way to remove the background DC voltage without affecting the amplitude of the 500Hz signal? I have a very low-ripple computer controllable voltage source available. So if I measure the offset to be 50mV I can for example set it to output -50mV. I just do not know the correct way to apply the offset to the signal

Because of the low frequency of the signal, I guess a low-pass filter is not really an option here?

Thanks in advance!
 

ericgibbs

Joined Jan 29, 2010
11,869
hi SS,
Welcome to AAC.
One method that is used with higher level signals is to use a OPA to apply a cancelling DC Offset voltage signal to the source signal.
Which type of amplifier do you use on the nV signals.?
Do you have a diagram that you could post.?
E
 

Thread Starter

SSPhysicsMeasurments

Joined Feb 18, 2019
4
Thanks Doros. Got it. I am sure I could buy a differential amplifier of some sort which could do it if necessary. I just thought there may be some easy way to apply this offset to the signal.

Thanks OBW0549 as well. The AC coupling has a cut-off frequency of either 1kHz or 100kHz so is not going to work with my low-frequency signal unfortunately.

Basically I got the wrong amplifier for this application.

I realize in my original post I said "low-pass" filter when I meant "high-pass". I also now realize I may be able to it with a really low cut-off frequency high pass filter (like 0.1Hz). As long as I leave my experiment running for a minute or so before I measure I think this may work...
 

crutschow

Joined Mar 14, 2008
26,324
Why do you think you need such a low frequency cutoff?
For example, a simple 8Hz HP RC filter will be down ≈1mdB @ 500Hz (see below)
It has a settling time-constant of 20ms at start-up.

upload_2019-2-18_8-57-28.png
 

OBW0549

Joined Mar 2, 2015
3,566
Thanks OBW0549 as well. The AC coupling has a cut-off frequency of either 1kHz or 100kHz so is not going to work with my low-frequency signal unfortunately.
Ummm...no. You are misinterpreting the specs for this amplifier: it has a high-frequency cutoff of either 1 kHz or 100 kHz; the lower end of its frequency response is 1.5 Hz. See the specs for "Lower Cut-Off Frequency":

amplifier.png
 
Top