Differential Amplifiers Questions and Answers

Answer: C

When the input difference is large enough, all the tail current is steered to one side, and that fixed current is all there is to charge whatever capacitance follows - so the output cannot change faster than the current divided by the capacitance. Transition frequency and load resistance govern the small-signal response instead, and common-mode range is unrelated. A large-signal limit set by a fixed current is quite separate from the small-signal bandwidth, and an amplifier can be fast by one measure and slow by the other.

Answer: D

Slew rate is the tail current divided by the compensation capacitance, so more tail current gives a proportionally faster edge - at the cost of more supply current and, unless the loads are changed, a different gain. Load resistance, current gain and common-mode range do not appear in the expression. When a limit is a ratio, either term is a legitimate target, and which to change depends on what else each one affects.

Answer: A

Dividing the available current by the capacitance:

\[ \begin{aligned} \frac{20 \times 10^{-6}}{30 \times 10^{-12}} \\ &\approx 0.67 \ \mathrm{V/\mu s} \end{aligned} \]

giving about 0.67 volts per microsecond, close to the figure a general-purpose operational amplifier quotes. That such a familiar specification falls out of two internal quantities is worth noticing, since it explains why faster devices run more tail current or use less compensation.

Answer: B

A field effect pair draws essentially no input current, so a high-impedance source is not loaded and no bias current flows through the source resistance to produce an offset. Its transconductance and its offset voltage are both worse than a bipolar pair's, and its common-mode range is a design matter. The two families trade input current against offset and gain, so the choice follows from the source impedance the amplifier has to work from. Naming the cost of a choice alongside its benefit is what turns a fact into a design decision.

Answer: C

Putting common-base devices above the pair holds their collectors nearly fixed, removing the Miller multiplication and raising the impedance seen at the output, so bandwidth and gain both rise. Offset is a matching property, the tail is a separate circuit, and common-mode range is if anything reduced by the extra headroom needed. A technique that improves two specifications for the price of headroom and two devices is why the cascode appears in almost every precision input stage.

Answer: D

The classical arrangement is two buffers feeding a differential stage, with one resistor setting the gain of the whole thing, giving very high input impedance at both inputs and excellent rejection. A bare pair has neither the impedance nor the rejection, and neither cascaded pairs nor a transformer describes it. Combining stages so that each supplies what the others lack is the pattern, and here the buffers supply impedance while the differential stage supplies rejection.

Answer: A

If the two arms of the output differential stage are not in exact proportion, a signal common to both inputs no longer cancels and appears at the output as though it were a difference. Bias current and slew rate are set elsewhere, and the differential gain shifts only slightly. Because the rejection depends on a ratio of ratios, it is far more sensitive to matching than to absolute values, which is why matched resistor networks are sold for exactly this purpose. Most errors here come from carrying a result across from a neighbouring case where it does not hold.

Answer: B

The pair is the input stage, supplying the two inputs, the high rejection and much of the gain. The output stage is a class AB follower, compensation is a capacitor around the second stage, and the bias reference is a current mirror network. Knowing which internal stage supplies which external specification is what lets a data sheet be read as a description of a circuit rather than a list of numbers.

Answer: C

A comparator asks which of two voltages is larger, which is precisely a question about their difference, so the same arrangement serves. Rejection and low offset are welcome and offset in particular sets the comparator's resolution, but the structural reason is the difference. Compensation is exactly what a comparator does not want, since it is designed to switch as fast as possible rather than to be stable in a loop.

Answer: D

A dominant pole placed to guarantee stability in a feedback loop limits how fast the output can move, and a comparator is never used in a linear loop - it is wanted to switch as quickly as possible. Offset, common-mode range and bias current are unaffected. That two devices sharing an input stage are compensated in opposite ways, according to whether they will be used linearly or as a switch, is why substituting one for the other rarely works. Framing the question as which quantity the change acts upon makes the answer follow without calculation.