opamp won't output zero

[Kevin Weddle]

I did this experiment with an opamp and found that the output had quite a bit of DC offset. It looked as though the difference between the inputs was being gained too much. And yet I was able to get a reasonable sized amplification. So if it's gain, which I think it isn't, why is it still high? And why does my small input offset cause the output to rail. It doesn't do that in the recommended configuration.

The answer is that it really never gained that high. A gain of even 10 can cause your output to rail. Maybe my gain was as high as 100, but you can never tell because I didn't use the standard gain setting resistors.

 

[audioguru2]

Why not post your circuit so we can see what went wrong?
Then we can recommend a solution for you. ;D

 

[gsmaster1]

There is an error in the inputs. It is caller Vio (Input offset) usually it is some

 

[audioguru2]

Hi Gsmaster,
Most opamps have a max input offset voltage of only 10mV. With Kevin's opamp configured for a DC gain of only 100, its output certainly won't hit the rail.

Kevin says that he didn't use the standard gain setting resistors so I am guessing that his circuit has an offset current problem.
Maybe Kevin's circuit has a megohm input or feedback resistor and therefore has excessive input offset current. Half a uA into a 1M resistor will cause the opamp's output to hit the rail when the gain is 100.
Maybe Kevin's opamp is operating beyond its input common-mode voltage range.
Maybe, maybe and maybe!  ??? ??? ??? We won't know until Kevin posts his circuit.

BTW, The middle pin on an offset adjustment pot usually goes to the opamp's negative supply, not to ground. I know, you were thinking of a single supply. Some FET-input opamps have the offset pot connected to the positive supply. ;D

 

[Kevin Weddle]

Audioguru, I like to think in terms of collector and emitter resistors. A base resistor is like an emitter resistor and an output resistor is like a collector resistor. This is how I lower gain. I make sure the output resistor is at least 10K and that I don't overload my generator. It's not perfect, but the gain is low enough. Negative feedback looks good because of the current through the external resistors. Positive feedback looks good too, but it's harder to keep the input near zero because there is no inversion and thus the input signal is large. Simple voltage dividing. The only thing remaining is why I can't just use an output resistor to ground and lower the gain if it's only a collector resistor. Well a collector resistor lowers the gain considerably, and if you feed it back even more so because of the external current. The current around the opamp works with positive feedback too, but it's plagued because it cannot stay near zero.

 

[audioguru2]

Hi Kevin,
Where's your opamp circuit? Post it so we can see if you are overloading its output emitter-followers with your output resistor. Your output resistor certainly isn't a collector resistor.

Transistor circuits and opamp circuits operate completely differently.
The collector of an NPN transistor is a very high impedance current sink.
The output of an opamp is a very low impedance voltage source but has a limited amount of current.
The input of a transistor is a fairly low impedance.
The input of an opamp is an extremely high impedance.
A single transistor has a max voltage gain of about 100.
Many opamps have a voltage gain exceeding 1 million.

A non-inverting opamp circuit doesn't have positive feedback. The output is fed back to its inverting input and it becomes a follower. If this negative feedback is done with a voltage divider then it has some gain.

Since an opamp has a gain of about 200,000 then of course its input voltages are nearly the same. The input voltages are near zero only if it is an inverting circuit.

 

[Kevin Weddle]

Audioguru, I don't want to post my circuit when it's easy to explain. All I am trying to do is experiment. Searching for a broader meaning. I know you don't think the only way to lower gain is with negative feedback. I mean, if I wanted, could I not design an amplifier with whatever gain I wanted and not use negative feedback. As I have said, it is my belief that the opamp is just a simple amplifier that is widely used because you can select the gain you want. If I want a gain of 100, or 1000, it can be done with this chip. Very popular, but just use one and be done.

 

[audioguru2]

if I wanted, could I not design an amplifier with whatever gain I wanted and not use negative feedback?

It won't be repeatable because each transistor would have a different Re and therefore would have a different gain. You would need to select the transistor from a bucket full of them, or tailor the collector resistor to match each transistor. Also, the distortion would be excessive without negative feedback.

The gain of an opamp or pile of discrete transistors in an amplifier with excessive gain is determined by two negative feedback resistors, and most of the distortion is cancelled. ;D