Non-inverting opamp with single supply - I give up

How do I make this work in multisim?

I have this circuit:


Virtual ground of of Vcc(16V) / 2 = 8V.
There is a function generator off the top edge of the image feeding a sine wave (o.5V peak, 10kHz) into my opamp.

With both the signal generator and the 4 channel scope grounded, this is what I get. The sine wave is the input signal from the function generator. But all I get from my opamp is a flat line.



I tried grounding the the function generator to my virtual ground of 8V.
The sine wave input signal goes off the top edge of the scope unless I put in a negative offset to move it down into view.

And the output of my opamp seems to be a clipped triangle wave, oscillating about 8V, instead of an identical sine wave of large amplitude.

I don't get it. How do you implement a non-inverting amplifier with a Vcc/GND

My non-inverting opamp works terrifically if I make the supply 16V/-16V with the standard configuration straight from the opamp datasheet.

Is this just an artifact of multisim to be ignored?

The flat line shows that you didn't manage to hit the input voltage range of your amplifier. My suggestion would be to calculate first, than put the circuit together.

Secondly, LM358 isn't suited to amplify 10 kHz with a gain of 20 and 10 Vpeak output, both regarding gain-bandwitdh product and large signal bandwidth. Check this by selecting a considerably lower frequency for your simulation, e.g. 100 Hz.

FvM said:

The flat line shows that you didn't manage to hit the input voltage range of your amplifier. My suggestion would be to calculate first, than put the circuit together.

Secondly, LM358 isn't suited to amplify 10 kHz with a gain of 20 and 10 Vpeak output, both regarding gain-bandwitdh product and large signal bandwidth. Check this by selecting a considerably lower frequency for your simulation, e.g. 100 Hz.

Click to expand...

LM358

These circuits consist of two independent, high-
gain, internally frequency-compensated which
were designed specifically to operate from a
single power supply over a wide range of voltages.
The low power supply drain is independent of the
magnitude of the power supply voltage.

Click to expand...

Surely the only explanation is that the model for this component in multisim is simply wrong or there is some other bug.
From the above I should be able to hook it up as stock standard non-inverting amplifier and it should amplify the whole signal between Vcc and GND with no clipping of the negative part of the signal.

boylesg - independent on the question which opamp to be used, I like to mention that the whole amplifier configuration does not work.
A non-inverting opamp gain stage (single supply) with a gain (1+Rf/Ro) needs:

(a) a voltage divider Vcc/2 to bias the non.inv. input.
(b) unity dc feedback (to transfer the dc bias to the output); this is accomplished using a large capacitor between ground and the lower feedback resistor Ro.
(c) an input coupling capacitor at the non-inv. opamp input (because of the dc bias).

LvW said:

boylesg - independent on the question which opamp to be used, I like to mention that the whole amplifier configuration does not work.
A non-inverting opamp gain stage (single supply) with a gain (1+Rf/Ro) needs:

Click to expand...

No $hit Sherlock ;-) I am just hopelessly confused about the whole thing.

There is a myriad of consistent instructions on how to do a stock standard non-inverting amplifier.

But I couldn't seem to find the same consistent instructions on how to do one with a single power supply and virtual ground. To my untrained eyes there just seems to be so many different sorts of circuits. I tried, and failed miserably to extrapolate.

I setup a stock standard non-inverting amplifier on my breadboard using LM358AN and with a single power supply and it works just fine.

So the problem I saw in multisim is a problem with multisim and not with the opamp configuration......as in the stock standard non-inverting opamp configuration but using Vcc and GND rather than Vcc and -Vcc.

Which was what prompted me to try and make a vurtual ground in multisim. On my breadboard I simply don't need this with LM358.

Can you point me in the direction of a page that explains the bandwidth issue? Preferably not one meant for second year electronic engineers or what ever.

Multisim might not exactly reproduce LM358 parameters, but it mostly does. It's surely not primarly a Multisim problem.

I read from post #3 that you are not yet aware of OP parameters like large signal bandwidth respectively slew rate.

It's possible to evaluate OP behaviour by trial-and-error method, not bother yourself too much with theory, but then be prepared for unexpected results and ready to correct your design instead of blaming the simulation tool.

`

FvM said:

Multisim might not exactly reproduce LM358 parameters, but it mostly does. It's surely not primarly a Multisim problem.

I read from post #3 that you are not yet aware of OP parameters like large signal bandwidth respectively slew rate.

Click to expand...

Up until now I have not yet done that much with opamp so I have not bothered to get my head around all their details. But now I really need to.

With multisim......if I set up a stock standard non-inverting opamp and view it in the output in the scope compared to the input sine wave, then it truncates the negative half of the sine wave and I just get a series of amplified humps from the positive phase of the sine wave.

If you don't think the model of the LM358AN is wrong then I must be doing something wrong with the ground connections on the instruments or something. But I am buggered if I can see what it might be.

According to the datasheet and given that LM358AN is designed for a single supply, I should see an amplified sine wave oscillating between GND and Vcc shouldn't I?

The identical circuit that does not work in multisim as expected, does work on my bread board. If multisim is entirely correct about the truncataion of the negative phase of the sine wave, then I would hear massive sound distortion from my bread board circuit. There is a small amount of crackling type distortion, probably related to the bandwidth issues you raised. But other than that the sound quality was pretty good.

Basically I replaced the transistor in this circuit with a non-inverting opamp with x10 gain:http://www.circuitdiagram.org/images/555-audio-amplifier-circuit.GIF

I presume your breadboard circuit is simply a bit different from the simulation setup. LvW already mentioned the problem of biasing the non-inverting OP input, depending on what's the actual sound source is, it's probably behaving different than a ground referenced test generator.

A sound signal has only low magnitude in the 10 kHz range (although the high frequency parts are important for sound quality), so you won't get an actual slew rate problem with LM358.

FvM said:

I presume your breadboard circuit is simply a bit different from the simulation setup. LvW already mentioned the problem of biasing the non-inverting OP input, depending on what's the actual sound source is, it's probably behaving different than a ground referenced test generator.

A sound signal has only low magnitude in the 10 kHz range (although the high frequency parts are important for sound quality), so you won't get an actual slew rate problem with LM358.

Click to expand...

Oh.....may be I need to introduce a +0.25V offset to the sine wave? Because on my bread board, the audio ground and my gel cell ground are tied together

boylesg,
before speaking about "minorities" (like offset) - don`t you think the first step should be to build a circuit based on a principle that works?
The amplifier as shown in your first posting works only as a linear amplifier in case the input voltage at the pos. input rides upon a dc voltage in the range of Vcc/2.
Where does this circuit comes from?
Why don`t you follw the classical design approach for a simple non-inv. gain stage for single supply (as described in post#4) ?
And - to be independent on a possible dc bias of the test generator - use capacitor coupling.

The 'scope photo you posted shows no DC voltages. Is the output of the opamp at about 0V? Then since it has apositive-only supply, the negative parts of the input are missing from the output.
1) The input DC voltage is 0.43V lower than the voltage divider on the inverting input so the opamp amplifies the difference 21 times and forces its output to 0V.
2) There is no negative supply so with the output as low as it can go at 0V then the negative part of the expected output is missing.

The proper way to do it is like this:

Audioguru said:

The 'scope photo you posted shows no DC voltages. Is the output of the opamp at about 0V? Then since it has apositive-only supply, the negative parts of the input are missing from the output.
1) The input DC voltage is 0.43V lower than the voltage divider on the inverting input so the opamp amplifies the difference 21 times and forces its output to 0V.
2) There is no negative supply so with the output as low as it can go at 0V then the negative part of the expected output is missing.

The proper way to do it is like this:

Click to expand...

Forget that circuit. As I now realize, it was a miserable failure to try and add a virtual ground, which I have not tried before and is not needed anyway with an LM358.

I mean a stock standard non-inverting amplifier with 100k feedback resistor and 10k resistors on both - and + inputs - straight from the LM385 datasheet. Then add a 0.5V offset to my sine wave so that it oscillates between 1V and GND.

The above should match the situation on my breadboard where the the GNDs of my audio source and power source are tied together I reckon.

The situation with the GNDs is everything if my simulation is to match my breadboard and I am to get accurate info from mulisim. And I think, as detailed, I understand what I was doing incorrectly in multisim.

Many semiconductor manufacturers make the LM358. National Semi invented it but do not show a non-inverting amplifier.
I looked at a few datasheets and found this one from copy-cat ST Micro. Next time please post the schematic you are using.

It shows a DC amplifier that is not usually used for audio because it amplifies DC and if you suddenly change the gain then there will be a very loud POP.
Won't it be a nuisance to offset the audio signal?

Here is the circuit I am guessing that you used:

Audioguru said:

Many semiconductor manufacturers make the LM358. National Semi invented it but do not show a non-inverting amplifier.
I looked at a few datasheets and found this one from copy-cat ST Micro. Next time please post the schematic you are using.

It shows a DC amplifier that is not usually used for audio because it amplifies DC and if you suddenly change the gain then there will be a very loud POP.
Won't it be a nuisance to offset the audio signal?

Here is the circuit I am guessing that you used:

Click to expand...

That's it. But then I found that, in multisim, only the positive phase of the input signal was amplified. So I went off trying to figure out how to do a virtual GND without much success.
But then I re-read the datasheet and realized that you do not need a virtual GND with LM358 and that either the model for the component was faulty or I was doing something else wrong in multisim.

And I believe it is the latter. If I am tieing the GND of the signal to the GND of my bread board then, logically, the input audio signal must be oscillating entirely above GND. So on my breadboard I have a virtual GND for my audio signal of 0.25V. And I can produce in multisim by adding an 0.25V DC offset to the function generator sine wave.

Your problem is that you biased the wrong input of the opamp. My copy of your circuit explains it using Ohm's Law and simple arithmetic: