OPAMP questions

[Kevin Weddle]

You can see from the evidence, a lot of opamp configurations, that the designers are willing to put anything around an opamp. Every single one of those configurations acts differently on the bias. Does anybody else think that the opamp is being too lenient when it comes to external circuitry.

[MP]

I have never looked at this as a leniency. I have always looked at it as versatility. That is what makes the op-amp so popular. But there are rules. There are lots of publications regarding the use of op-amps and complete libraries of circuits built with them.

MP

[Kevin Weddle]

What is the accuracy difference between using the opamp differentialy as opposed to not. It is my understanding that the inputs require changes in current to function best. The example opamp shows no signal applied to the input. I think the changes in current are too small, although the signal appears credible. Is this taking advantage of the opamp? Does the signal need to change the input current more?

[arock1957]

I have a question that is baffling me! The word "operational amplifier" originated from mathematical operations that could be performed with it. Assume you wanted to produce a circuit for which the output voltage was given by the expression V(out)=-3A-2B (A and B are variable input voltages). Show how this operation could be accomplished with a summing amplifier by drawing the circuit. Show values for resistors. Any ideas?

[Kevin Weddle]

I think you are talking about the amplification and the voltge dividing. Zero current into the opamp and that sort of stuff. I think the gain in a nondifferential situation is high but is limited by the zero current factor. The gain would have to be high because of the small change in base current producing the output voltage.

[Kevin Weddle]

How does one go about making sense of an opamp configuration? Start with a zero bias? Would you replace the capacitors with resistors? Do you use the maximum supply voltage? What are some primary concerns when designing an opamp?

[MP]

You might want to start here:
http://pasture.ecn.purdue.edu/~abe591s/Opamp.pdf

MP

[Kevin Weddle]

I have noticed that the BW is very narrow on some opamps. Why is it so narrow? Do you think it is the transistors they use? Maybe the sheer number of transistors and their associated capacitive properties contributes to the overall reduction. So I think this is the frequency at which the capacitance becomes the overall problem. What is the net result of negative feedback in a transistor anyway?

[audioguru]

Most general-purpose or audio voltage-feedback opamps have a low frequency gain of 100,000 or more and then start a 20 dB/decade roll-off at 10 Hz to 100 Hz. They have the built-in roll-off so that at a high frequency such as 1 MHz or more, where the op-amp circuit will have close to 180 degees of phase-shift, then the roll-off causes the gain to be less than one.
When you add negative feedback then the op-amp's phase-shift will cause that feedback to become positive at a high frequency but since the roll-off reduces the gain to less than one then the circuit will not oscillate.
Negative feedback is used:
1) To reduce the gain to a useful number and to make that gain very accurate.
2) To extend the bandwidth depending on gain (higher bandwidth with less gain).
3) To reduce distortion depending on gain (less distortion with less gain).

[Kevin Weddle]

Did you know that positive feedback is different than phase shift. The phase shift margin needs to be less than 360 degrees, but positive feedback can occur at any time.

[Kevin Weddle]

I don't think opamps would be very popular if the phase shift were a problem. The problem is that the opamp could become unstable.What does this mean? The input signal can not drive the output to it's voltage level. Am I right? This would be because of the negative feedback of the individual transistors.

[ante]

I must say that op-amps have been around for many years and I have accepted them as they are because they are very useful!

Ante ::)

[audioguru]

Kevin,
A compensated (with built-in roll-off) opamp that is using negative-feedback IS stable because the roll-off makes certain that at high frequencies, the gain is reduced so that the circuit will not ring nor oscillate. The compensation ensures a good phase-margin.
Stability of an opamp is referring to its lack of ringing or oscillation at a high frequency. The only times that a compensated opamp with negative feedback cannot drive its output to a certain voltage level is:
1) When the load is drawing more current than the opamp's rating then the opamp will current-limit the output swing.
2) When the frequency is high enough (an old 741 at 8kHz or a more modern TL071 at 100kHz) so that the opamp is slew-rate-limited where its internal circuitry cannot charge and discharge its compensation capacitor quickly enough. When the opamp is slew-rate-limited then a sine-wave input results in a reduced-level triangle-wave output. Spec sheets show the maximum frequency where slew-rate does not limit the output swing in a graph called POWER BANDWIDTH.

[Kevin Weddle]

Opamp compensation serves another purpose. The input capacitor, often seen with compensation, changes the critical frequency. If you notice the bode plot graph, you will see that the critical frequency can be anywhere. When you add a capacitor, the critical frequency moves from the 40db per decade back to the 20db per decade. This is difficult to see because it shows that the bandwidth is extended. Which is true, when you change the critical frequency. So what I am saying is that the lag through the opamp at high frequencies is due to the reverse collector base of the transistors.

[audioguru]

Kevin,
A compensated opamp works VERY well when used within its simple rules and specs. Don't try to build your own opamp from scrach, just use the ones that are so popular.
The only problem that I have ever had is that some opamps oscillate at a very high frequency when driving a shielded audio cable and/or the shielded cable to my 'scope (capacitance). But that problem is simply solved by adding a 100 ohm resistor in series with the opamp's output.

[Kevin Weddle]

Remember that an oscillation has to do with a change in polarity across the input. This is what causes the voltage to go in either direction. An oscillation is simply anything around an opamp circuit that would cause the input to change polarity. This could include the feedback. Any good scientist will tell you that a configurable device such as an opamp is just as good as the product. But if your talking DC, then there is no additional phase lag and the opamp works fine.

[Kevin Weddle]

I must elaborate on an earlier statement. The input capacitor, associated with phase shift, does not inherently change the characteristic of the opamp. Instead, it changes the input signal such that it will not produce 40db roll-off. This makes the graph misleading because it shows the 20db being extended. If you accept
that the capacitor is separate, then you will be left with the graph. The graph shows a change in critical frequency thus extending the 20db.

[audioguru]

Kevin,
The input capacitor, if used, to an opamp circuit is external to the feedback and therefore has no effect on the high frequency roll-off.
The input capacitor, souce impedance and opamp's bias resistor determine the low frequency phase-shift and roll-off.
Where did you get 40dB from? Most internally-compensated opamps use a single capacitor for their 20dB/octave high frequency roll-off.
The flat frequency-response part of the graph gets extended due to your gain-determining negative feedback. This feedback exchanges the enormous open-loop gain, for bandwidth.

[audioguru]

OOOOps
Kevin,
The opamp roll-off is 6dB per octave (each doubling of frequency) or 20dB per decade (ten times the frequency).

[Kevin Weddle]

I have it. The input capacitor causes the gain to increase by 20db per decade. This is because the capacitor produces a 20db per decade roll-off and it is placed as an emitter element. As the frequency is increased, the gain is higher because it is an emitter element. A higher gain means it is able to keep the 20 db.

[audioguru]

Kevin,
The input capacitor does NOT cause the gain to increase. It simply rolls-off the low frequencies, below its "corner" frequency at 20dB/decade. The corner frequency is determined by the capacitive reactance and the source and bias resistors.
Below the corner frequency, the capacitor passes less signal because its reactance is high (making an attenuator).
Above the corner frequency, the capacitor passes the full signal because its reactance is low (like a piece of wire).

[Kevin Weddle]

The input capacitor is to ground. It is not in series like a differentiator. Do you understand the concept of the emitter element? This is sometimes the resistor R1. R2 is the feedback resistor. The fact that it is an emitter element is why it sets the gain.

[audioguru]

Kevin,
What opamp circuit has its input capacitor connected to ground?
Input capacitors usually couple in series an AC signal source to the opamp input and bias resistor for noninverting, or opamp input resistor for inverting circuits. The low-frequency rolloff of an input coupling capacitor is like a differentiator.
What is an emitter element?
If you are talking about an inverting opamp circuit, then its gain is simply R2 (feedback) divided by R1 (input) plus the source impedance.

[Kevin Weddle]

I have seen many configurations of the opamp and am still unconvinced of a few things. I believe it should be operated with differential inputs. I believe it should have the correct output bias current. Anything else is cheating the operation. It should also have the correct input bias current which is easy to calculate.