OP-AMPS

# Op-Amps

## Overview

At it’s most basic level, an operational amplifier (op-amp) is a discrete analogue integrated circuit which acts as a voltage-amplifier with very high gain.

An op-amp’s gain can be adjusted with the appropriate external circuitry (the op-amp’s internal gain does not change, but the gain of the entire circuit does). This is almost always done for any op-amp that is used as an amplifier. External circuitry can also manpulate the op-amp to act as as a current amplifier.

## Schematic Symbol

An op-amp is commonly drawn on schematics as:

You may see this symbol with or without the voltage supply pins $$V_{S+}$$ and $$V_{S-}$$. If they are not present, it is assumed that they are connected up to a power source which should be obvious from the design intent.

## Uses

• Buffers (a.k.a. voltage followers)
• Linear amplifiers, non-linear amplifiers
• Pre-amplifiers (amplifier’s with a input voltage of 10mV or less)
• Analogue mathematical operations (differentiators, integrators, summer, subtracters, e.t.c)
• Voltage to current (transconductance) and current to voltage conversions (transimpedence)

## Op-Amp Topologies

### Voltage Follower (a.k.a. Buffer)

A voltage follower (also known as a buffer) is one of the most basic circuits you can make with an op-amp.

As shown in the above diagram, the output voltage is the same as the input voltage ($$v_{o} = v_{i}$$). Well isn’t this pointless? No, the key point to a voltage follower/buffer is that it can convert a high-impedance input into a low impedance output. Practically, this means that you can now sink/source more current from the output without the voltage changing. Buffers are great for boosting signals that travel across long distances, or for signals which get split and go to many devices (this is called fan-out, and is common with digital clock signals).

A simulation schematic for a voltage-follower op-amp is shown below: A simulation schematic for an op-amp configured as a voltage-follower (buffer).

The results of the simulation: The simulation results for an op-amp configured as a voltage-follower (buffer). Note how the output voltage mirrors the input voltage exactly.

### Non-Inverting Amplifier

A op-amp in the non-inverting amplifier configuration is shown below.

The equation for the gain of the non-inverting amplifier is:

$$v_o = (1 + \frac{R_f}{R_i} ) v_i$$

Notice the $$1$$ in the gain equation? This means that no matter what you set the resistors $$R_f$$ and $$R_i$$ to, you can never get a gain which is less than one. This is one of the disadvantages of the non-inverting amplifier (you can have a gain of less than one with an inverting amplifier).

Here is a simulation schematic (circuit) for a non-inverting op-amp amplifier running from a single-ended supply. Because R1 ($$R_f$$) and R2 ($$R_i$$) are both $$1k\Omega$$, the op-amp has a gain of:

$$G = 1 + \frac{R_f}{R_i} \\ G = 1 + \frac{1k\Omega}{1k\Omega} \\ G = 2$$

The results of the simulation are shown below. As you can see, the output voltage $$V_{out}$$ is exactly twice the input voltage $$V_{in}$$. A graph of Vout vs. Vin for a non-inverting op-amp amplifier circuit.

### Inverting Amplifier

A op-amp amplifier in the inverting configuration is shown below:

The equation for the gain of an inverting amplifier is:

$$v_o = - \frac{R_f}{R_i} v_i$$

The negative sign is to show that the output is the inverse polarity of the input. Notice that, unlike the non-inverting amplifier, an inverting amplifier lets you obtain a gain of less than 1.

Below is the schematic used for simulating the behaviour of an inverting op-amp. Note how is requires a negative voltage power supply.

And below are the simulation results for the above schematic: Vout vs. Vin for an inverting op-amp with a gain of -1.

### Differential Amplifier

A differential amplifier amplifies the difference between two electrical signals.

The output voltage is given by the equation:

$$v_o = \frac{R_4}{R_3 + R_4}(1 + \frac{R_2}{R_1})v_2 - \frac{R_2}{R_1}v_1$$

Below is a schematic for simulating the behaviour of a differential op-amp:

This schematic produces the following results: A graph Vout vs. Vin1 and Vin2 for a op-amp configured as a differential amplifier.

### Integrator

It’s output voltage is proportional to the integral of the input voltage w.r.t. time. The figure below shows an ideal op-amp based integrator.

However, this circuit is normally not practical in real world situations. Any errors such as the output offset voltage and input bias current (which all op-amps invariably have), as well as a non-perfect input signal with small amounts of DC bias, will cause the output to drift, until it reaches saturation.

A way to fix this problem is to insert a high-valued feedback resistor, $$R_f$$, to limit the DC gain, as well as a resistor, $$R_{bias}$$, on the non-inverting input terminal to compensate for the input bias current. An op-amp configured as a non-ideal (real world) integrator, with feedback resistor Rf to slowly remove DC offset.

### Transconductance Amplifier

A transconductance amplifier is an op-amp topology which is used to convert a voltage into a current. Coincidentally, it is also known as a voltage-to-current converter.

A basic transconductance amplifier can be built with an op-amp in a non-inverting configuration.

A transconductance amplifier is useful creating an industry standard 4-20mA (or 0-20mA) current-loop signal. The input voltage can come from something like a potentiometer or microcontroller (coupled with either using a VDAC peripheral or PWM/RC-filter technique to create a variable voltage).

One disadvantage with this design is that the current output is not ground referenced, that is, ground is not used as the return path for the current. This complicates the wiring.

## Important Variables

Sorted by function.

### Common-Mode Input Voltage Range

The common-mode input voltage range is the range of voltages that can appear at the input to the op-amp and it still work correctly. For standard single-supply op-amps, the typical range is approximately $$0V$$ to $$V_+ - 1.5V$$. Note how it includes the most negative rail $$V_-$$ (which is 0V for a single-supply op-amp) but only gets within 1.5V of the most positive rail, $$V_+$$

### Input Offset Voltage (Vos/Vio)

The input offset voltage $$V_{OS}$$ (or sometimes called $$V_{IO}$$) is the voltage difference required between the two input pins to force the output to 0. It is a DC measurement parameter. In an ideal op-amp, the op-amp only amplifies a difference between the inputs, and so the output is 0V when the difference is 0V. However, real-world op-amps always have some unavoidable differences in the internal components that make up the op-amps (specifically, in the input differential stage of the internal circuitry), and thus the inputs are not perfectly identical. The input offset voltage is typically in the following ranges:

• 1-5mV for good general purpose op-amps, 5-15mV for really bad ones.
• 200uV-1mV for specialized low input offset voltage op-amps
• 10uV-200uV for the best “ultra” low input offset voltage op-amps

For example, the general purpose LM324 has a typical input offset voltage of 2mV and a maximum of 3mV, at $$T_A = 25°C$$1. “Low” input offset voltage op-amps will have a $$V_{OS}$$ in the range of 50-200uV. For example, the OPAx196 family of op-amps has a max. $$V_{IO}: 100uV$$2.

A non-zero input offset voltage results in gain errors between the input and output of a op-amp.

Input offset voltages vary by op-amp transistor technology. Bipolar op-amps typically have the lowest input offset voltage, followed by CMOS and the BiFET op-amps3.

### Input Bias Current (Ib+ and Ib-)

The input bias current $$I_{B+}$$ and $$I_{B-}$$ describe the currents that flow in and out of the op-amps input pins. In an ideal op-amp, no current flows into/out of the input pins (the op-amp has infinite input impedance). In reality, always some small amount of current will flow. Typical input bias currents range from 1-10nA.

The amount and behaviour of input bias current depends on the op-amp transistor technology. A FET-based op-amp’s input bias current will double with every 10°C rise in temperature4.

Input bias currents are a problem because these currents will flow through external circuitry connected to the op-amps inputs. This current when flowing through resistors and other impedances will create unwanted voltages which will increase the systematic errors.

The input offset current $$I_{OS}$$ is the difference between the input bias current at the + pin and the - pin.

### Input Impedance

The input impedance is the internal resistance to ground from the two input pins. In an ideal op-amp, this value is infinite. For most op-amps, this value is somewhere between 1-10MΩ.

### Gain-Bandwidth (GBW) Product

The gain-bandwidth product can be initialised as GBWP, GBW, GBP or GB. It is an important parameter which basically puts a limit on the maximum gain and frequency. An op-amp’s maximum possible gain reduces as the frequency of the signal increases. The multiplication of the gain with the frequency gives the gain-bandwidth product, which is relatively constant for a particular op-amp.

Hence if the gain bandwidth of a particular op-amp is 1Mhz, and the gain is 10, the maximum frequency that the op-amp can operate linearly at (still provide a gain of 10) is at 100kHz. Or if the gain was set to 100, then the maximum frequency is 10kHz. This also means that an op-amp has a built-in low-pass filter, as the gain drops for very high frequencies.

An example of an ultra-high gain bandwidth is 1700MHz, which are present in ‘Wideband CFB” op-amps, designed for applications such as RGB line drivers (such as the OPA695). A ‘normal’ GBW can be anywhere between 100kHz and 10MHz. A low gain-bandwidth is around 1kHz (reminiscent of less advanced, older op-amps). Remember gain is unit-less (V/V), so gain bandwidth is expressed as a frequency only. Not realising this can be confusing! The GBW product is closely related to the slew rate (see below).

### High Level Output Voltage

The high level output voltage ($$V_{OH}$$) defines the highest voltage which the op-amp can drive the output to (with respect to the power supply $$V_+$$).

### Low Level Output Voltage

The low level output voltage ($$V_{OL}$$) defines the lowest voltage which the op-amp can drive the output to. The LM324 is rumoured to only be able to drive the output near ground if it is sourcing current, but only to 0.5V minimum if sinking (see this EDA Forum post, LM324 Opamp Gain Instability).

### Slew Rate

Slew rate is another op-amp parameter. It defines the maximum rate the voltage can change with respect to time. In an ideal op-amp, this would be infinite. It has the SI units V/s, and is commonly expressed in uV/s. It can be thought of as the slope of the output waveform if one of the inputs of the input was subjected to a step voltage change. This parameter usually increases as the GBW increases.

### Supply Current

The supply current is generally constant for any supply voltage. It is normally rated as the no-load current. Obviously, if there is a load on the op-amp, the current through the power pins will be the sum of the no load ‘supply current’ and the current going through the load.

Supply currents for standard op-amps are typically between 1.5-4mA. A ‘low-power’ op-amp has a typical supply current between 0.5-1.5mA (such as the LM258N). Then there are ultra-low power op-amps that only draw 5-20pA (such as the LMC6464). You normally sacrifice slew-rate and gain-bandwidth for ultra-low power.

Cascading op-amps is concept when the output of one op-amp is connected to the input of another. There can be an arbitrary number of op-amps in the cascade, but usual limits are 3-4.

For a fixed-gain, cascading op-amps can also be used to increase the bandwidth, as each individual op-amp now can operate at a lower gain and therefore has a larger bandwidth as defined by the gain-bandwidth product. Note though that each additional op-amp added to increase the bandwidth gives diminishing returns. Also important to note that op-amp bandwidth is defined as the -3dB gain points. Hence the bandwidth does not stay the same (total bandwidth gets smaller) when two identical op-amps are cascaded, as these will now the -6dB points. A practical limit for fixed-total-gain increased-bandwidth cascading is about 3-4 op-amps.

### The Gain

When cascading op-amps, the total gain is the product of all of the individual op-amps gains, i.e.:

$$A_{total} = A_0 A_1 A_2 ... A_n$$

### The Bandwidth

The bandwidth of cascaded op-amps is not as simple to calculate as the gain.

If all of the op-amps are identical, then the following equation can be used:

$$BW_{tot} = BW \times \sqrt{2^{\frac{1}{N}} - 1}$$

where:
$$BW_{tot}$$ = the total bandwidth of the cascaded op-amp system
$$BW$$ = the bandwidth of the individual op-amps (remember, they have to be identical)
$$N$$ = the number of op-amps in the cascaded system

## Feedback Resistor Values

As a rule-of-thumb, you should use the lowest acceptable resistances in op-amp feedback paths to reduce instabilities.

## Types Of Op-Amps

### General Purpose

General purpose op-amps typically have parameters in the following ranges:

• Gain Bandwidth Product: 1MHz
• Input Bias Current: 15pA
• Input Voltage Offset: 1mV
• Output Current: 20-50mA
• Icc: 1mA

### Rail-to-Rail Op-Amps

A rail-to-rail op-amp is an op-amp which supports input voltages near the power rails, and can drive the output close to the power rails. We must stress the word NEAR, as the op-amp’s output voltage will never get exactly to the rail, due to the finite voltage drop across the output-stage transistors. Rail-to-rail op-amps just support wider ranged input voltages and can drive closer to the rails than general purpose op-amps can. Look for the low level output voltage ($$V_{OL}$$) parameter in the op-amp’s datasheet. For “rail-to-rail” op-amps, this will usually be about 100-200mV about ground at normal load currents.

This also means that a rail-to-ral single-supply op-amp cannot output 0V. To achieve a true ground output, you need a negative voltage supply. There are dedicated ICs designed to provide a small negative power supply to op-amps so that they can output true ground. One such example is the Texas Instruments LM7705, a “Low Noise Negative Bias Generator”. This IC only generates -230mV, which allows the designer to use CMOS-based op-amps which usually have a maximum supply voltage of 5.5V. The typical application schematic for the Texas Instruments LM7705, a 'Low-Noise Negative Bias Generator' for the negative supply of an op-amp. This allows the op-amp to output true 0V. Image from http://www.ti.com/.

### Micropower Op-Amps

Micropower is a termed used for extremely low quiescent current op-amps that are designed for battery or energy recovery-based power supplies. The supply current of micropower op-amps is typically within the range of 50-100uA at a supply voltage of 2-10V. Because they are designed for battery-based systems, they are also commonly single-supply op-amps.

## Manufacturer Part Number Families

• LT: The prefix Linear Technology (now Analog Devices) uses for their range of op-amps.
• LT1006: Precision, single-supply op-amp.
• LT1077: Micropower, single-supply op-amp.
• OP07: Analog Devices/Texas Instruments ranges of low input offset voltage op-amps.
• OP07C: ±3-18V VCC, -40 to +85°C industrial temp. range
• OP07CP: DIP-8 package
• OP07CS SOIC-8 package
• OP07D: ±4-18V VCC
• OP07E: 0 to 70°C commercial temp. range
• OPA: Texas Instruments (previously Burr-Brown) family of op-amps.
• OPA241: Single-supply
• OPA251: Dual-supply
• OPA27: Texas Instruments (previously Burr-Brown) family of ultra-low noise, precision op-amps. Internally compensated for unity-gain stability.
• OPA37: Uncompensated version of the OPA27.
• TLE202: Texas Instruments family of “high-speed low-power” precision operational amplifiers. Belong to the Excalibur family of TI op-amps which uses “isolated vertical PNP transistors” to give unity-gain bandwidth and slew rate improvements.

## Examples

Below are some examples of op-amps that stand out from the crowd for some reason, be it popularity, years in service, or functionality wise.

Manufacturer CodeDescriptionApproximate Price (1 unit, US$) AD860xGood for high precision stuff! Awesome for photo-diode amplification (both current-to-voltage and voltage-to-voltage configurations).$3.50
LM32xA common family of op-amps that has been around for along time, they can operate of a single supply and can swing right to ground, but cannot swing to the rail voltage. The LM321 has one op-amp, the LM328 has two (dual), and the LM324 has 4 (quad).
LM833One of the cheapest 'audio' op-amps available (about US$0.20 as of 2011). Features a high GBW for it's price. OPA695This is a ultra-wideband, current-feedback op-amp. If you need an op-amp with a ridiculously high gain-bandwidth product, this is along the lines of what you want to use. It has a GBW of 1700Mhz and a maximum slew-rate of 4300V/us.$3.50
OP07A op-amp with a "ultra" low input offset voltage (resistors are trimmed at production time to achieve this), guranteed to be no more than 75uV. This op-amp also features offset nulling pins to further reduce the input offset voltage by performing trimming once the op-amp is installed in a circuit.n/a

## Isolation Amplifiers

Isolation amplifiers provide galvanic isolation between the input (sensor) and output (measurement circuitry). They are used to protect the sensor measurement and recording circuitry (e.g. a microcontroller with on-board ADC) from dangerously high voltages at the sensor, and also the opposite, to protect the sensor environment from potentially dangerous voltages on the rest of the system.

A common application would be to isolate and amplify the voltage across a current-sense resistor on a high-power motor, or to protect humans with medical sensors connected to them from the measurement system.

Basic isolation amplifiers require two power supplies (one for each side of isolation), while others incorporate built-in transformers so that you only have to provide one power source.

## Input Resistors

One of the first things you learn about an op-amp is that the input impedance on the input pins are very large (ideally infinite). So naturally you would start to question why resistors would be connected to the input pins of an op-amp like shown in the diagram below: Schematic showing a resistor on the positive input to a op-amp.

These input resistors serve to limit the input current if the voltage on the input pin goes above $$V_{CC}$$. Most op-amps have protection/clamping diodes from the input pins to $$V_{CC}$$ (typically you can determines this if in the datasheet the input pins max voltage is rated to $$V_{CC} + 0.3V$$, which is one diode voltage drop). If there was no resistor there, the built-in diode would conduct and sink a large current from the input pin to $$V_{CC}$$, possibly damaging the op-amp. The resistor limits this current to a safe value.

## Offset Nulling Circuits

Some op-amps which are designed to have very low input offset voltages also come with offset nulling pins to further trim the input offset voltage once the op-amp is installed in circuit. The OP07 is one op-amp which has these pins. Typically, a $$10-50k\Omega$$ potentiometer is connected across these pins with the wiper going to $$V_{CC}$$, as shown in the example schematic below: