Low pass filter

You need to be more descriptive of the power levels and frequencies involved. If you're talking about line level audio, then download TI's FilterPro software. It does differential input filters too.

I assume that you must not have access to the internals of this amp or you need to remove noise from the other end of the leads? And, if you don't have access to the input to where this is going? then you must make an inline filter. you didn't specify what spectrum you're trying to filter out? if this is spike noise ? a common mode choke will work, you simply pass both outputs through it..

No, you need a balanced filter. That assumes you are going to feed a balanced load. However, common mode problems might make the two filter approach better.

Tam

Thanks for the response

I am trying to design a chopper stabilized amplifier and the last stage is a LPF. The stage before LPF is a modulator and its output is a modulated signal. So I try to use a LPF to keep the first harmonic and filter the other harmonics of the signal coming out of the modulator.

My input is a 1khz, 5V DC signal, so I think I should have a LPF with cutoff frequency at 1kHz.

Can you example what do you mean by pass both outputs through it.

capacitor. The -ve pin of the opamp connects to one output of my modulator and the +ve pin of the opamp connects to ground.

So, do you suggest that I should connect the other output from modulator to the +ve pin of the opamp instead of the ground?

Thanks

Hi Tam Can you example how I can implement a two filter approach?

Do I need to make additional connection beside connecting one LPF to each modulator output?

Thanks

I have always used a balanced filter, but if your amplifier can drive the signals with respect to ground you might be able to use two filters . Can't think of any other connections you would have to make. Unless the two filters are absolutely identical, the phase shift and amplitude will be slightly different for the + and - signals, and things that should cancel out won't.

Can't you use another op amp, and convert to a single ended signal? This would take care of the common mode, and give you a clean signal with respect to ground. From your original posting I couldn't tell if this was at microwave , or what.

You really should simulate this in SWCad, or the like. For two filters, see what happens if the components are slightly different.

Tam

You can simply take two equal resistors from the outputs into a common capacitor with half the computed (f=RC*2pi) value. An Instrumentation Amp would then give you the single ended output. IMHO if you do the switching right, you actually don't need a filter.

Hmm, i just replaced a Chopper Op-amp on a voltage sensing board at work, it just used a RC network to smooth off the edges. In any case, i would try a simple common mode filter and check it on the scope .. I'm assuming that you must be using a very square wave signal that you're trying to remove the harmonics that are getting generated from it ? Most likely, any filter you use is going to curve the left and right shades.. if that is fine with you, i think maybe just having a non inductive type resistor on each output would fix your problem. a little by pass capacitance should be applied on the output side of the R's I'm thinking 100 Ohm R should do it with like a .001 uf by pass, but that is just off the top of my head.

filters like Active types may tend to take your nice square wave and truncate it. That has been my experience on that subject, it only makes sense that it would do that. But then again, the idea i gave you above may also give you some deforms thus giving you what i call a hook tooth wave, in which case an inductor added to the scene can handle that. I think you need to experiment a little.