Creating a floating AC ground with DC blocking capacitors

Thread Starter

SamR

Joined Mar 19, 2019
4,117
Reading through the Texas Instruments Handbook of Op Amp Applications and came across this.

1639024762940.png
Understand about using 2 power sources in series and creating a virtual floating ground. And understand using capacitors for DC blocking. Confused where it says using single power supply by creating a floating AC ground with DC blocking capacitors. Obviously not by using a capacitor to ground for the -Vcc. Just what is it talking about here?
 

DickCappels

Joined Aug 21, 2008
8,671
Just consider that the method is ok and the writing is poor. The term "DC Blocking Capacitor" doesn't seem to mean much in this context.
1639029191796.png
This probably what the author was thinking of. When I use this circuit I put capacitors from the output of the opamp to make this kind of a "supply splitter", as shown below, I used capacitors to provide high frequency "ground" for both the +5 and -5V supplies.
1639029462587.png
The removed circuitry are over-voltage clamps, not relevant to this discussion. Be careful -some op amps become unstable with large capacitive loads.
 

Thread Starter

SamR

Joined Mar 19, 2019
4,117
Had me confused and thought I had missed something along the way. Typically I use 0.1uF snubbing caps although I have seen it mixed with a low and hi cap to attenuate more bandwidth. Had me puzzled that TI would be using confusing language as they are usually very clear to me. Thanks Dick
 

DickCappels

Joined Aug 21, 2008
8,671
That handbook is a Burr-Brown publication first published by BB in 1963, and updated by TI after they absorbed BB. The confusing text is in the original publication (I have it here), so don't blame TI for writing it, but you can blame their editor for not catching it.
 

DrBearEE

Joined Feb 3, 2020
6
In audio circuits with single supply, that "rail splitter" places the DC operating point of the op amp near 1/2 the raw supply, so it can operate on an AC signal. This half supply point is often called "virtual ground", because its the audio signal zero point, but not the power supply return.

"Blocking capacitors" would be added in series with the AC signal in and out, which couple the AC signal but isolate the driving circuit from the DC offset (such as a microphone signal). Same for the driven circuit (a speaker, perhaps) . I think this sample circuit may be acting to create this virtual split supply for other circuitry, hence the follower transisitors and large reservoir capacitors that the op amp is driving?

Agree that its very important to analyze the particular amp for capacitive loads, especially at turn-on. It may ring for a while. Or maybe oscillate forever. Part of the secret sauce in most commercial opamps is the internal frequency compensation networks, so you have to pick the amp carefully to have it work well in this job. Not sure if the followers here isolate the amp from the capacitive load.
 

Parkera

Joined May 3, 2016
96
It's a whole lot more readable than the old Walt Jung Op Amp Cookbook for me.
If you are talking about the Third Edition, I totally agree. The First Edition I have always found to be an excellent source, albeit written at a technician level as opposed to an engineering level. The op-amps used as examples in both editions are all pretty much obsolete today, but the theory still holds.
 

k1ng 1337

Joined Sep 11, 2020
413
Just consider that the method is ok and the writing is poor. The term "DC Blocking Capacitor" doesn't seem to mean much in this context.
View attachment 254613
This probably what the author was thinking of. When I use this circuit I put capacitors from the output of the opamp to make this kind of a "supply splitter", as shown below, I used capacitors to provide high frequency "ground" for both the +5 and -5V supplies.
View attachment 254615
The removed circuitry are over-voltage clamps, not relevant to this discussion. Be careful -some op amps become unstable with large capacitive loads.
So the capacitors in the bottom diagram create a floating ground that is reactive in nature as opposed to resistors resisting hence reducing real power losses by blocking DC?

I see virtual grounds having 3 cases:

1) 2 resistors in series
2) 2 resistors in series that are in parallel with 2 capacitors in series with both virtual grounds connected
3) 2 capacitors in series
 

ronsimpson

Joined Oct 7, 2019
1,996
When I use this circuit
Slight variation on your circuit. I usually put a resistor from B-E of the transistors. At very low current levels the current flows from OP-amp to load through this resistors. If the op-amp is good for 10mA then choose the resistor so that it takes 2 to 5mA to have 0.65 volts across it. Now the first 3mA is supplies by the amp and the rest is supplied by the transistors. This seems to stabilize the amplifier.
1639176939423.png
The feed back comes from the output and not the base. I would make the "47" resistor small to zero.

IMO (often not right but I try)
RonS.
 
Top