Block a DC Component : A Simple Capacitor and a DC Servo Circuit?

I have always used a capacitor in series with a resister to ground to filter high frequency AC noise, probably not what you want here. I use a capacitor between stages, that is in series between an output and the next stage input, to block DC, specifically for charge pumps. Audio amplifiers commonly have AC coupling capacitors between stages to eliminate any DC bias.

Thanks, Stephen.

I found the information of DC servo from google search. I do not have experiences on those things.

Audio amplifiers do commonly have AC coupling capacitors to block the DC bias for protecting speakers. It seems some of the high-end audio amplifiers adapt DC servo, like the method 2 in my original post.

So I wonder if a pair of capacitor and resister has the same cut-off frequency what's the difference between the method 1, passive capacitor and method 2, an active feedback with a low pass function?

Does the DC servo one, with proper design, provide more "accurate" or "stable" DC elimination?

A resistor and capacitor in series to ground will not affect DC. There will be no electron flow in or out of the capacitor with DC so it might as well not be there. The resistor and capacitor are used for a different reason. A resistor in series with a capacitor slows the charge rate of the capacitor so it responds to a wider frequency of AC. It would be used to reduce the AC noise produced by motor. The resistor and capacitor in series are called a bouchcell or zobel network. You will find zobel networks across speakers to reduce the microphone effect of a moving speaker - either just varying impedance or actually feeding voltage back to the amplifier.

The op-amp circuit removes AC so maybe both circuits are useful - one to remove DC and the other to remove AC noise.

Hi!
I have to manipulate an AC signal comes with a unwanted DC component in two cases:
A. wanted AC signal ~ 800 ~ 1.2K Hz, Vpp ~ 300 mV, like a deformed sine wave; the un-wanted DC component : ~ 400 mV.
B. wanted AC signal ~ 15~ 60 Hz, Vpp ~ 50 mV, like a deformed sine wave; the un-wanted DC component : ~ 4 V.

It seems there are two ways to remove the DC components:
1. block the DC by applying the signal to a capacitor + resistor to GND, then I can get the AC signal only at the point between the capacitor and resistor., see the attached image.

2. remove the DC by a DC servo circuit like the one on page 21 of the data sheet of AD8221,see the attached image.

I can not tell what's the difference between the method 1 and 2. For example for one capacitor and one resistor (means one pole solution ?) case, the frequency response of the "filter" is the same for both the method 1 and method 2. Isn't it?

Which one method do you suggest for the Case A and Case B? (Especially in the points of stability, robustness, response time. Thanks!

Maybe my question is a kind of a question to passive filter and active filter.

Thank you for your time!

Hi

Please see the new attached image for the passive method with one capacitor and one resistor. One resistor and one capacitor can be used to eliminate the DC in this way.

I don't see an attachment ...

Sorry, the added film was added in my original post. Please refer in the very first post of this question. Thanks!

that looks like an ac coupling cap with a load to ground. Your input circuit may already have a path to ground so the resistor is an unnecessary load.

Hm... yes. By this capacitor/resistor arrangement, there should no(or very small) DC offset cross the R as well as the next stage. This is actually very similar to some of the audio power amplifier for protecting speaker.

Although the electronic signal I have to manipulate is not an audio signal. I found, by articles reading, not real experience, most of the audio amplifiers use a passive capacitor to block DC offset for protecting speakers, while some of the high-end audio amplifier use DC servo for the same purpose.

I wonder is it better to go DC servo for my case...?

It depends on the application - if you want very low frequency signals (i.e. those that are time varying but almost close to DC) you might need active DC nulling...however this too will need a large time constant so that it does not block the wanted "almost close to DC" signals. The disadvantage here is that it adds complexity to the circuit and subject to the same rules of failure as an active circuit.

A simple RC filter would be a lot more reliable (at least if the resistor and capacitor are not stressed) with the only downside being the limit it imposes upon the lowest usable frequency that you can get (the largest available capacitor value and in some measure the resistor value will impose this limit). Where possible this should be the choice. Now this is only as far as DC blocking is concerned...if you want a filter function then an active filter would be needed especially for 2nd order or higher slopes. That said even the passive filter would probably need an active buffer stage before and after it to minimise loading effects.

Thank you for your hints.

I will have a chance to do my first try next week. Here are my guesses for both Case A and B in my first try:

For the Case B, signal @ about 15 ~ 60Hz, I guess a high order high pass filter may needed for DC
blocking and minimizing freq. filtering to the targeting signal.

For the Case A, signal @ 1K Hz, I will adapt a RC filter with buffer stages.

For the purpose of DC blocking, a RC filter can be used for both cases. Simply use the lowest frequency of interest as the reference which in this case is 15 Hz. So a RC filter with cut-off of 7 Hz or less will block DC but at the same time not have much attenuation at 15 Hz. Just add a buffer stage after the RC filter to minimise loading effects.

Let's say R = 100k, then for a cut-off of abut 7 Hz, C turns out to be 0.22uF (220nF) - since 220nF is the bare minimum, you can use a slightly higher value such as 270nF or 330nF...upto about 1uF to block DC.

Thanks! Your reply makes me more clear for the RC way I have to do. The example also help me a lot.

I will also find a chance to build a DC servo and try to experience the differences between RC filter and a DC servo... Hope I can tell the differences in my work soon. Do you have any comments about DC servo?

For one, DC servo will be a lot more complex but achieve more or less what the passive RC scheme would. DC servos are useful where we absolutely need direct coupling, or cases where outputs of two or more similar circuits are connected in parallel into a load. In this case a DC blocking capacitor might not be optimal - an example is 2 or more power op-amps paralleled into a speaker.

Thank you for your input.

I can not image "output of different circuits are connect to the same load in parallel". But I got several key words in your reply then found a very useful reference( for me, it tells me many things...):

http://www.ti.com/lit/an/snaa021b/snaa021b.pdf

Yes, the TI article is about BPAs (bridge-parallel amplifiers) and a very useful reference.