Virtual Ground to Op Amps - Will This Work?

[Jim Stevens]

I would appreciate comments about the linked circuit below. It is two
op amps powered on different sides of a virtual ground, with the
outputs connected across a resistive load.

http://www.4shared.com/photo/llPcodei/virtual_ground_dual_current.html

Any suggestions on which common op amps would be suitable for this
circuit?

Yes, it is a little odd. Just curious to see what happens.

Jim Stevens

 

[Jasen Betts]

I would appreciate comments about the linked circuit below. It is two
op amps powered on different sides of a virtual ground, with the
outputs connected across a resistive load.

biasing of the right two op-amps looks kind of wierd.

http://www.4shared.com/photo/llPcodei/virtual_ground_dual_current.html

Any suggestions on which common op amps would be suitable for this
circuit?

hmm, 3V supply, 12mA output probably not a cheap one, but really it
depends what the goal of the circuit is.

 

[Spehro Pefhany]

biasing of the right two op-amps looks kind of wierd.


hmm, 3V supply, 12mA output probably not a cheap one, but really it
depends what the goal of the circuit is.

If the circuit has no purpose, then may as well choose the very
cheapest one, since it will meet all the other requirements.

The circuit, as drawn, might be a problem set by an instructor to
illustrate non-ideality in op-amps.

You'll see offset, noise and non-ideal voltage swing/saturation. A few
LM358s might suffice for that kind of an exercise if you don't let the
ideal battery drop too much in voltage.


Best regards,
Spehro Pefhany

 

[P E Schoen]

"Jim Stevens" wrote in message

I would appreciate comments about the linked circuit below. It is
two op amps powered on different sides of a virtual ground, with
the outputs connected across a resistive load.

Any suggestions on which common op amps would be suitable for this
circuit?

Yes, it is a little odd. Just curious to see what happens.

The virtual ground is reasonable enough to provide a +/- supply for op-amps
from a single battery, but the ones shown are not connected that way, and
the circuits do nothing meaningful. You might want to simulate this circuit
using LTSpice (www.linear.com). It would be a good learning experience. What
are you trying to accomplish?

Paul

 

[Jim Stevens]

The second stage amp outputs will - in the best circumstances - collapse
to their negative rails. With real opamps perhaps they won't quite get
there. Yes, not very useful, and hardly worth trying to pick devices.

-Frank

And, I beleive, the degree to which they collapse depends on the value
of R1/output current.

With regard to the other comments so far, 12mA (and more) at 3V can be
done with something like a TS941. Although I was hoping to use
something on hand.

As for "uselessness", wasn't that what they first said about the
laser? It's just learning by tinkering here.

Jim Stevens

 

[Jim Stevens]

You'll see offset, noise and non-ideal voltage swing/saturation.

Spehro Pefhany

Can you please explain briefly what features of the circuit cause this
to occur?

Is the "offset" you refer to due to the output of the negative amp not
reaching ground?

Noise is not an issue here, and I realize the swing will be odd.
Hopefully odd enough to be somehow interesting.

Jim Stevens

 

[Jim Stevens]

The virtual ground is reasonable enough to provide a +/- supply for op-amps
from a single battery, but the ones shown are not connected that way, and
the circuits do nothing meaningful.

Paul

Why nothing meaningful? I didn't show the signal inputs on pin 2 of
each amp, assuming it would be implied. It is just a 2Vpp audio
frequency sinewave to both.

Hence the outputs are identical, but with a DC offset from each other
governed by the load resistance.

Several experts here have said there is a problem with the circuit,
but I would like to understand exactly why. It's an experiment, not a
hi-fi system.

Jim Stevens

 

[[email protected]]

Why nothing meaningful? I didn't show the signal inputs on pin 2 of
each amp, assuming it would be implied. It is just a 2Vpp audio
frequency sinewave to both.

Where to start... The input opamp is a fairly common configuration to
split the rails for the bias point of subsequent opamps, though often
it's set up as a low-pass filter to reduce wideband noise. R3 does
nothing good, and limits U1's output current (not a good thing).

U2 and U3 use U1's output for their positive and negative rail,
respectively. This means the output of the upper rail cannot go below
this and the lower cannot go above this. Therefore, you *must* have a
DC current in the output. This isn't generally a good thing for an
audio circuit. I guess there may be reasons for this but you haven't
said what you're trying to do so this whole discussion is pretty
silly.

The U3 is referenced to the negative rail so its output can't go below
that. IOW, if your input is a sine wave with 0V DC bias it's going to
clip. Ditto U2 and its negative rail (the output of U1). Both are
going to clip on the negative half cycle. Even if there is a DC bias
on the input signal, there is no DC bias where neither opamp is
clipping.

Hence the outputs are identical, but with a DC offset from each other
governed by the load resistance.

Nope. At least one is always driven into the rails. Even if this were
the case, why would you want such a thing? One opamp can do this,
BTW. Just sum in a DC voltage.

Several experts here have said there is a problem with the circuit,
but I would like to understand exactly why. It's an experiment, not a
hi-fi system.

If it's an experiment why ask here? Why not see for yourself that it
does nothing good? If you're trying to accomplish something specific,
give the complete picture and people will be glad to help (though it's
probably better over in .basics).

 

[whit3rd]

I would appreciate comments about the linked circuit below. It is two
op amps powered on different sides of a virtual ground,

.... which results in the outputs of each op amp being pulled
by the other outside the absolute-max limit on the output pin.

The 'output' op amps are both driven in such a way that they
can lock-up (if offset voltage is considered, the balance
equation for the output state includes TWO solutions).
Arguably, the load resistor can pull up the lower op amp to
keep it in linear operation.

I don't see any input.

The ground-establishing op amp has 110k ohms (series resistance)
on its (+) terminal; a 110k ohm series resistor to the (-)
terminal will keep the temperature dependence low (bias current
times imbalance of resistance adds to voltage offset, and bias
current is strongly temperature dependent).

 

[Jim Stevens]

On Sun, 18 Nov 2012 09:04:08 +1100, Jim Stevens

Now that I have a better idea of what to expect, I will build one up
this week and verify how it performs.

It's hard to predict an unusual circuit without alot of practical
experience ... more than I have anyway.

Thanks for the comments.

Jim Stevens

 

[[email protected]]

On Sun, 18 Nov 2012 09:04:08 +1100, Jim Stevens

Now that I have a better idea of what to expect, I will build one up
this week and verify how it performs.
*pfft*

It's hard to predict an unusual circuit without alot of practical
experience ... more than I have anyway.

If you can't predict how it's going to work, why are you wasting time
building it? You have the cart before the horse (in so many ways).

 

[Jim Stevens]

One opamp can do this, BTW. Just sum in a DC voltage.

This sounds like a good idea. Can you be a little more specific about
the method you are referring to?

As previously mentioned, this is just a learning process. There is
really no goal apart from that. I suppose everyone here is over that
stage.

Jim Stevens

 

[P E Schoen]

"Jim Stevens" wrote in message

On Sun, 18 Nov 2012 10:08:00 -0500, [email protected] wrote:

This sounds like a good idea. Can you be a little more specific
about the method you are referring to?

As previously mentioned, this is just a learning process. There is
really no goal apart from that. I suppose everyone here is over that
stage.

You should know enough about op-amps to figure out basic operation. You can
assume ideal operation and determine approximate behavior in a circuit. It
appears that you are trying to achieve a sort of bridge circuit output that
can swing between + and - rails from a virtual ground reference. But clearly
the two output op-amps can only swing between their respective rails and
virtual ground. Thus the voltage across the load will always be positive. I
doubt that is what you want. I can't see it being a learning experience to
build this circuit, except maybe using a simulator. Usually, the only reason
for building an actual prototype circuit is to observe behavior that is not
as easily predicted or modeled, such as instability and oscillation. Or
possibly behavior of op-amps under other than recommended conditions, which
can cause latch-up and other possibly destructive operation.

You might want to check out some proven op-amp circuits and learn from them:
http://www.amazon.com/Op-Amp-Cookbook-3rd-Walter-Jung/dp/0138896011
http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html
http://www.ece.ucsb.edu/Faculty/rodwell/Classes/ece2c/resources/an-31.pdf
(National)
https://instruct1.cit.cornell.edu/courses/bionb440/datasheets/SingleSupply.pdf
(TI)
http://en.wikipedia.org/wiki/Operational_amplifier_applications

and learn LTSpice:

Paul

 

[Jim Stevens]

On Sun, 18 Nov 2012 09:04:08 +1100, Jim Stevens


Please excuse the profileration of posts, and my lack of initial
clarity. Here is a revised circuit showing the signal inputs and other
changes.

http://www.4shared.com/photo/29tKWyFG/Virtual-Ground-Amps-2.html

Input signal is line audio (test tone).

Note the outputs are not ground referenced.

Jim Stevens

 

[Phil Allison]

"Jim Stevens"

Please excuse the profileration of posts, and my lack of initial
clarity. Here is a revised circuit showing the signal inputs and other
changes.

http://www.4shared.com/photo/29tKWyFG/Virtual-Ground-Amps-2.html

Input signal is line audio (test tone).

Note the outputs are not ground referenced.

** So you have a signal voltage that directly drives the virtual earth
inputs of two op-amps above and below supply rails.

Brilliant.


.... Phil

 

[[email protected]]

I would appreciate comments about the linked circuit below. It is two

op amps powered on different sides of a virtual ground, with the

outputs connected across a resistive load.



http://www.4shared.com/photo/llPcodei/virtual_ground_dual_current.html



Any suggestions on which common op amps would be suitable for this

circuit?

None whatsover.

Opamps do not like non-zero differential voltages applied to their (-)/(+) inputs, they just rail. So U2 rails to the node labeled GND and U3 rails to node labeled NEG on battery. It's the only way to drive the currents thru R4/R6 to 0.

 

[josephkk]

I would appreciate comments about the linked circuit below. It is two
op amps powered on different sides of a virtual ground, with the
outputs connected across a resistive load.

http://www.4shared.com/photo/llPcodei/virtual_ground_dual_current.html

Any suggestions on which common op amps would be suitable for this
circuit?

Yes, it is a little odd. Just curious to see what happens.

Jim Stevens

The circuit as drawn won't work worth squat.

?-)

 

[josephkk]

On Sun, 18 Nov 2012 09:04:08 +1100, Jim Stevens


Please excuse the profileration of posts, and my lack of initial
clarity. Here is a revised circuit showing the signal inputs and other
changes.

http://www.4shared.com/photo/29tKWyFG/Virtual-Ground-Amps-2.html

Input signal is line audio (test tone).

Note the outputs are not ground referenced.

Jim Stevens

Not quite as messed up as the first one, however the output is 100% common
mode with a DC bias current. None of the input signal appears on the
output in a useful way.

Try powering both output amplifiers from the battery directly and make one
inverting and the other non-inverting. You might get some useful output
that way.

?-)