A logarithmic video amplifier or LVA is typically part of radar and electronic countermeasures microwave systems and sonar navigation systems, used to convert a very large dynamic range input power to an output voltage that increases logarithmically with increasing input power.
The need to process high-density pulses with narrow pulse widths and large amplitude variations necessitates the use of logarithmic amplifiers in modern receiving systems. In general, the purpose of this class of amplifier is to condense a large input dynamic range into a much smaller, manageable one through a logarithmic transfer function. As a result of this transfer function, the output voltage swing of a logarithmic amplifier is proportional to the input signal power range in dB. In most cases, logarithmic amplifiers are used as amplitude detectors. Since output voltage (in mV) is proportional to the input signal power (in dB), the amplitude information is displayed in a much more usable format than accomplished by so-called linear detectors.
Indication of the performance of a logarithmic amplifier is the measurement of the log transfer function. This is accomplished by the use of Automatic Test Equipment, in which the output of an RF or microwave signal generator is stepped over the input dynamic range of the device under test (DUT). The computer then records the log amplifier video output voltage and calculates the slope and deviation from a best-fit straight line using a least-squares method. The result is a plot consisting of the measured video output voltage and the log conformance deviation in dB.
Another indication of a logarithmic amplifier’s performance is its ability to accurately measure pulsed-modulated RF signals. Typical pulse measurements include rise time, fall time, settling time and recovery time. An important part of accurate pulse measurement is the measurement test setup. It must be able to provide an extremely high on-to-off ratio of the pulsed RF source, and be typically 10 dB greater than the dynamic range of the device under test. The setup must have adequate rise and fall times several times faster than the DUT. Carefully matched input and output impedances are essential to ensure that impedance mismatches do not contribute to distortion of the measured pulse response.
The test set is initially calibrated using a continuous wave (CW) signal at the highest input power level of the DUT. The pulse modulation is then applied to the input of the log amplifier. The video output is measured on an oscilloscope with a bandwidth at least twice that of video bandwidth of the log amplifier. The full dynamic range video pulse response is displayed and the corresponding measurements are taken. Using a step attenuator, the RF level is then lowered until the pulse plus noise is just above the output noise of the log amplifier. This level measured is known as tangential signal sensitivity.
Provide wide input dynamic range, best pulse fidelity, exceptional log conformance (commonly known as log linearity) and a limited IF output. SDLA uses multiple compressive stages of RF gain to emulate the exponential transfer function. The output of each stage is coupled into a linear detector. The typical dynamic range of each amplifier/detector stage is approximately 10 dB, therefore many are required to cover a large dynamic range.
The outputs of each detector are then summed in a single video amplifier to provide a single detected output. The main advantage of an SDLA is seen in the combination of dynamic range and rise/settling times. Because the RF gain stages are compressing and the video amplifier is operating linearly, the SDLA can achieve dynamic ranges of greater than 100 dB while retaining rise times of less than 1 ns. An additional advantage of this type of logarithmic amplifier is that it inherently provides a limited IF output from the cascaded RF gain stages. This output is typically used to drive phase detectors or frequency discriminators and as such is extremely valuable in a variety of system applications.
Detector log video amplifiers provide broad operational frequency range, excellent temperature stability and similar log characteristics. DLVA is a type of logarithmic amplifier in which the envelope of the input RF signal is detected with a standard "linear" diode detector. The output of the detector is then compressed to simulate a logarithmic input/output relationship in the following video amplifier section. In general, the DLVA offers the advantage of operating over the widest frequency range, but at the sacrifice of dynamic range. The linear/square law range of the input diode detector limits the dynamic range of a DLVA. Typical dynamic ranges for a DLVA are in the order of 40 dB. Very often, the user will parallel two detectors, one with an RF preamplifier, to extend the overall dynamic range to greater than 70 dB. A major limitation of a DLVA results from the gain-bandwidth product of a video amplifier. Because the logarithmic transfer function must be accomplished in the video section, a tremendous amount of video gain is required for low-level RF signals (near the diode sensitivity). The amount of gain required causes rise time and recovery time degradation due to the gain/bandwidth constraints in the video section. The detected video sections of a DLVA can be AC coupled, DC coupled, or pseudo DC coupled. Each has its advantages, depending upon the application (i.e., CW operation, temperature compensation, etc.)
TLA is different in that it does not provide an envelope detected output. The output signal is actually an RF signal compressed in dynamic range by a logarithmic scale. As with both the DLVA and SDLA, the output signal's voltage is proportional to the input signal power in dB. An advantage of the TLA is that the output retains both amplitude and phase information for signal processing. These types of units are typically used in applications where sound is involved (i.e., sonar, IFF and navigation systems), but they also have applications in some of the more advanced signal processing systems. They can be used prior to ultra fast analog-to-digital converter to extend the usable dynamic range of such systems.
Dynamic Range is the range of the input signals in dB over which the output linearity requirement is met.
The variation of the detector VSWR and conversion factor due to frequency will appear on the output of the DLVA as an error. The worst case error caused by the frequency variation of the detector occurs when the VSWR ripple and conversion factor variations are added to the LVA maximum deviation from a true logarithmic response. LVAs do not exhibit any errors due to Rf frequency variations; only the detector determines errors. By improving the input match of the detector, this variation can be minimized, but care must be taken not to degrade the detector sensitivity and pulse response at the same time.
DLVAs are required to operate over large temperature ranges and, as such, any errors due to temperature effects must be minimized. Temperature variations will have an effect on the detector conversion factor, noise floor, and on the gain of the video amplifiers. DC offset caused by these variations will be present in the output of a DC coupled DLVA. The temperature effects on a DC coupled DLVA are the most complicated to overcome due to the very high linear video gain associated with signals close to the start logging area in the LVA. With this high gain, any movement in the detector, or any DC offset shift in the first linear video amplifier due to temperature change, will be translated as a huge DC offset in the log output. Attention must be given in the design to achieve a large degree of temperature stability. The use of tunnel diode detectors is an easy way of achieving good detector temperature stability.
Linearity of a DLVA can be explained as the deviation of the actual transfer function of a given DLVA from theoretical value of the logarithmic ratio. A noticeable linearity problem is the point at which the detector is moving from square law operation to linear and then to compression. This effect usually starts at the level of -20 to -10 dBm. To compensate for this effect, an additional logging stage is added in parallel at this point to increase the output onto the sum line. This effectively adds more gain to the upper logging stages in order to compensate for te compression of the detector conversion factor. This type of compensation is good for about 15 to 20 dB of input level above the detector break point; anything beyond this level must bel solved in a more complicated detector design.
The input VSWR of a DLVA is determined by the match of the detector diode. The VSWR of the detector is worse at the transition point from square law to linear region. One way of improving this is by using the detector as a current source. This is done by connecting te detector into a virtual ground on the first linear video amplifier in the LVA (tunnel diode). Another way is to build a matching network at the input of the detector. This type of solution will result in loss of sensitivity and more restricted frequency bandwidth of the input RF signal.
Noise Figure of any amplifier characterizes its noise power spectral density relative to the input of the device. This is an RF parameter that cannot be measured on a logarithmic amplifier because the gain stages are nearly into saturation on their own noise. One way to characterize the noise figure of a logarithmic amplifier is to measure the noise figure using classical y-factor techniques on the input two or three stages alone. It can also be estimated from the device’s TSS. Output RMS Noise Level is the noise power as measured at the video output with a true RMS voltmeter.
Amplifier Bandwidth includes the operational bandwidth (range of input frequency over which the electrical parameters of the amplifier are met) and additional frequency range necessary to accommodate the pulse width and rise/fall time.
Video Bandwidth is usually specified to address the detected signal's rise and fall time. It cannot be measured directly via the input RF signal, but must be characterized by injecting a swept CW signal into the video section or by calculations from accurate rise time measurements.
Rise Time is defined as the difference between the 10 to 90% point on the rising leading edge of the output video pulse.
Settling Time is defined as the difference between the 10% point on the leading edge of the video pulse to the first point in time where no deviations are outside a +/- dB window of the final settled value.
Recovery Time for a logarithmic amplifier may be defined in many ways. The most common is to use multiple pulses and characterize the time between the 90% point on the trailing edge of the first pulse, to the 10% point on the leading edge of the subsequent pulse. An additional method is in defining the time from when the trailing edge of the pulse exceeds the settled value by an amount equal to the linearity specification to the time within 1 dB of specified offset. There are several other definitions; however, it is best to define this for your application based on your particular system requirements.
Fall Time is defined as the difference between the 90 and 10% point on the trailing edge of the output video pulse(typically three to four times the rise time).
Tangential Signal Sensitivity (TSS) defines the input level that results in an output signal-to-noise ratio of 8 dB. Tangential sensitivity, which is directly related to noise figure and bandwidth, aids in defining the lower limit of the input dynamic range of a logarithmic amplifier. TSS is also a convenient way of specifying a logarithmic amplifier's noise performance since noise figure is not easily measured at the detected video output.