This article will address a historical perspective on a common type of negative feedback amplifier topology used for amplifying standard definition video signals (i.e., composite video).
With the advent of high-quality, broadcast-quality video tape recorders such as the ones used in the 1960s to 1980s and their associated TV test signal generators, the “final” video output amplifier fed the transmission links or TV monitors for evaluation of the video signal.
Some of these video output amplifiers were feedback amplifiers that included a differential pair input stage similar to op amps, but certain specifications that were important for instrumentation and general operation amplifiers were not necessary.
For example, a typical op-amp's performance is rated by some of the following characteristics:
- Input bias current
- Input offset current
- Input offset voltage
- Common mode rejection ratio
- Power supply rejection ratio
- Frequency response to -3dB amplitude
- Gain bandwidth product
- Slew rate or power bandwidth
- Harmonic distortion
- Frequency response and distortion, tested with sine waves at any number of amplitude levels up to the clipping point of the op-amp
An op-amp is often tested with a symmetrical AC signal such as a sine wave. However, a typical video op-amp for standard-definition video is rated for the following:
- Frequency response to 4.2 MHz to 6 MHz within 0.2 dB or better
- Power bandwidth to 4.2 MHz to 6 MHz within 0.2 dB or better
- Harmonic distortion is not measured for video amplifier performance; instead differential gain distortion or phase variation with a DC offset for a composite color TV signal is utilized.
- Video signal is tested for 2V p-p into 150Ω. Or put in another way, the amplifier usually includes a 75Ω output resistor, which then is connected to a 75Ω load or terminating resistor for a 1V p-p signal across the load.
A video amplifier is tested with an asymmetrical waveform signal instead. See Figure 1 that shows a synchronizing signal, a front and back porch signal, and a pixel area signal. Note that Figure 1 does not include signals in the vertical blanking interval that contain vertical synchronizing signals.
To measure frequency response, sine wave signals of different frequencies are added to a pedestal signal (DC offset signal). See Figure 1.
To measure the differential gain of a TV signal, the test signal includes a low-frequency ramp signal and a high-frequency sine wave signal. See Figures 2 and 3. The ramp signal is then high pass filtered out, and the sine wave is measured for amplitude variation across one TV line.
Figure 3 shows an enlarged view including the sync tip level, which is typically -0.286V for a 525 TV line system and -0.300V for a 625 TV line system. The backporch level, which is the blanking level of the video signal, is a reference voltage of 0V for both 525 and 625 TV line systems.
Not shown in Figure 3 is the peak white level, which is the maximum luminance level of + 714mV from blanking level for a 525 TV line system and slightly lower to +700mV for a 625 TV line system
Now let's take a look at typical op-amp topology in Figure 4.
Here we see a differential input stage voltage gain stage with Q1 and Q2, and its emitter current source I Bias 1. A second voltage gain stage consists of Q5 and Q6, with a Miller compensating capacitor CC . The second stage includes an active load current source I Bias 2. The output stage is a typical Class AB complimentary push pull amplifier with Q8 and Q9. With two voltage gain amplifying stages, a pole splitting technique such as using a Miller compensation capacitor is required for stability and a well behaved frequency response with negative feedback. Including two voltage gain stages ensures high DC open loop gain (e.g., 100,000V/V), which translates to high power supply rejection ratio when negative feedback is applied (e.g., a voltage follower circuit with 100% feedback). Also, note the current source I Bias 1 is used for improved common mode rejection ratio, typically 90dB at low frequencies.
Also note that the typical op-amp circuit in Figure 4 uses an active load for the differential input stage via a current mirror circuit via Q3 and Q4. In contrast, see Figure 5, a typical three-transistor video op-amp, which does not use active loads.
Looking at Figure 5, we see the following differences from Figure 4:
- Only one voltage gain stage, so the open loop DC gain is much lower than a typical op-amp
- Dominant pole set by CC on the collector of the differential amplifier
- Passive load resistors are used instead of a current mirror and current source.
- Common mode rejection is “inferior” because of using REE instead of a current source for Q1 and Q2 emitters.
- Single ended class A output stage is used instead of a push pull circuit.
For this example, we have RL = REE = RE3 = 1000Ω, with +VCC and -VEE as well regulated ±12V supplies.
The open-loop gain is low, roughly = RL /(2/gm,Q1Q2 ) = 1000/10 = 100V/V or 40dB, where gm,Q1Q2 ≈ 5.5ma/26mV.
However, the open-loop pole is very high at more than 16MHz because the dominant pole formed by RL = 1000Ω and CC with the parasitic capacitances of Q2 and Q3 have a total of less than 10pF.
Because the video op-amp is configured for typically a gain of 2 by having RF = R1 = 1000Ω, there is almost no need for any compensation capacitor CC , other than the capacitances internal to the transistors. For example, with Q1, Q2, and Q3 transistors as MPSH10 and CC = 0pF, this amplifier had no problem having a flat frequency response better than 0.1 dB up to 4.2 MHz, and the differential gain distortion and differential phase were at less than 0.1% and 0.1 degree.
For 1.00V p-p into 75Ω at Vout , the emitter of Q3 must provide 2.00V p-p. Measured maximum output voltage at the emitter of Q3 is + 6.70V and -1.5V when Vout is loaded with a 75Ω resistor. However, for 2.00V p-p, the sync pulse only needs to go to -0.6V p-p, so the maximum negative voltage swing of -1.5V has more than enough headroom.
Also note that the input transistors are not necessarily matched, and thus the input offset voltage and input offset current are “poor” by op-amp standards. However, in the world of video, the output at the emitter of Q3 is coupled to a DC servo circuit that samples the back porch region of the video signal and sends an error voltage back to the (-) input, via a resistor to the base of Q2 to ensure 0V at blanking level at the emitter of Q3.
The common mode rejection ratio is terrible by op-amp standards because of using a resistor, REE , instead of a current source for the emitter tail current. The differential mode gain is about 100, and the common mode gain is about RL /2REE = 1000/2000 = 0.5V/V. The common mode rejection ratio is thus about 200 or 46 dB. However, this common mode rejection ratio is “flat” out to the MHz region, which is different from most op-amps that show their common mode rejection ratios falling at frequencies beyond the dominant pole frequency, which is typically <1000 Hz.
In part 2 we shall examine other types of video op-amps and a DC servo circuit known as a back porch feedback clamp circuit.