small current amplifiers

[Kevin Weddle]

I have this idea about the operation of opamps. We all know the input transistors are low current. But how accurate are they. We all want to believe that they are perfect. But I think the  manufacturing process for these devices may not be as good as the higher current ones. Specifically their high beta creates a problem. By sheer means of the number. You take a beta of 500 and the base current is one then the number of the collector current is 500. You take a base current of 2 and the number of the collector current goes to 1000. It is sort of like the high gain principle in opamps. When you gain the signal at a higher number, you might get a gain of 20k or you might get a gain of 25k. If the gain is low, then you might get a gain of 20 or you might get a gain of 25. Just look at the difference in the numbers. So what I suggest is that by the sheer value of the numbers, you can expect a certain amount of error.

 

[audioguru2]

Kevin,
The input transistors in opamps use high beta for an extremely low bias current and an extremely high impedance, not for voltage gain. The input transistors use emitter-degeneration resistors to straighten their linearity which reduces their voltage gain a lot. Opamps use very high impedance current-source circuits instead of collector resistors, and multiple gain stages for very high voltage gain.

The voltage gain of an opamp is typically about 200,000. If you ended up with one at its limit of only 25,000 then when you add negative feedback to them for a closed-loop gain of 10 to 100, the error and difference between them is almost nothing. The tolerance of your feedback resistors causes far more error. You are not just attenuating the output to get the gain you want, you are adding negative feedback for its error and linearity improvements.

Look at what negative feedback does for the harmonic distortion of an opamp. Without negative feedback, the harmonic distortion might be as high as 1%. With negative feedback improving its linearity the distortion is 0.0005%!

I can't think of any application for an opamp where its gain is maximum, or "open-loop" without any negative feedback. Electrocardiogram circuits use a special "instrumentation" opamp with its extremely high gain and balance determined by very high-accuracy built-in resistors.

   

 

[MP1]

I can't think of any application for an opamp where its gain is maximum, or "open-loop" without any negative feedback.

 

[Kevin Weddle]

And what about the numbers. Don't you agree that there is more error in higher numbers. Odds are if your gain is 100,000 then +-1% error is a 2000 difference. But your gain is only 10 then +- 1% is a difference of 2. I know that theoretically you cannot go by percentage, but it can be used as a basis of comparison.

 

[audioguru2]

Kevin,
You missed the decimal point. 1% of 10 isn't a spread of 2, it's a spread of 0.2.

Let's compare the gain accuracy of an opamp circuit that has a gain of 10 with negative feedback:
1) The opamp has its minimum gain of only 25,000. According to the attached tutorial, the circuit's gain will be 9.995602.
2) The opamp has its typical gain of 200,000. The circuit's gain will be 9.999451.
3) the difference in gain between the circuits is only 0.0003849, pretty darn small.  ;D

View attachment 36775