Spectrum analyzers fall into two basic categories, real-time and swept. Real-time spectrum analyzer technology is derived from that of the oscilloscope. The instrument first gathers data in the time domain and then translates it into the frequency domain by means of the Fast Fourier Transform (FFT). The noise floor is lower than in an oscilloscope, but higher than in a swept spectrum analyzer. The real-time spectrum analyzer can capture transient and fast signals more quickly. The swept spectrum analyzer is less expensive.
Both types display amplitude as power on the Y-axis and frequency distribution on the X-axis. The two types differ in how they go about displaying this information and in their comparative facility in processing high-frequency signals.
The swept spectrum analyzer embodies the operating principle of the superheterodyne radio receiver. This is an old technology that gained widespread use in the 1920s, when the falling price of the vacuum tube made it feasible.
A superheterodyne radio works by mixing the incoming RF signal as received from a transmitter with a locally generated frequency. The two signals beat against one another to produce a lower, still-modulated intermediate frequency (IF) signal that is further amplified in one or more stages.
There are two advantages. The lower IF is more manageable so it can be processed more efficiently. And because there is a single frequency for all stations, the electronics are more task-specific as compared to the earlier simple regenerative receiver. The superheterodyne principle is conceptually more complex, but it is not expensive to mass produce and its advantages are compelling.
In the earliest superheterodyne versions, each station would broadcast, in addition to the primary signal, the unmodulated signal needed for mixing. Soon it was found to be more efficient for the mixing frequency to originate in each receiver. For this purpose, a two-gang variable “condenser” (really a capacitor) was deployed to tune-in the station while simultaneously creating the correct mixing signal to produce uniform IF for all signals received at the antenna. This superheterodyne principle is universally used today in virtually all radio and TV receivers.
Superheterodyne technology is also used in the swept spectrum frequency analyzer.
Another subtype of spectrum analyzer is the tuned-filter, swept-tuned frequency analyzer, which operates in a manner resembling the tuned radio frequency (TRF) radio. This tuned-filter instrument is actually a swept-tuned analyzer. It is in every sense a real-time instrument. A conventional swept-tuned frequency analyzer is not capable of evaluating all frequencies within a band simultaneously. So its use is restricted to frequency domain representations of steady-state repetitive signals. Only the real-time frequency analyzer can represent all frequencies at once. Therefore it transcends the limitations of swept-tuned instruments.
An example is the parallel-filter spectrum analyzer, combining a number of bandpass filters. The effect is to permit the instantaneous display of transient and time-variant dynamic signals. This approach, of course, entails added expense in manufacturing, so the user’s initial investment is higher.
Many elements of the modern swept spectrum analyzer closely resemble the superheterodyne radio receiver. In place of the antenna, the spectrum analyzer has an RF input where the signal under investigation feeds into an attenuator circuit. The purpose of this circuit is to reduce the amount of power and thus ensure an overload doesn’t damage the spectrum analyzer’s front end, should there be high-amplitude spikes or problematic mixer effects. Manual settings on the instrument’s front panel let the user control the degree of attenuation or, alternatively, fixed or variable attenuation may be external.
The next stage in the signal path is the pre-amplifier. It simultaneously boosts the amplitude of the signal and increases the signal-to-noise ratio. This user-controlled capability can have the effect of enhancing measurement sensitivity where the input power level is inadequate.
One problem with both oscilloscopes and spectrum analyzers is that sometimes the input signal refuses to display in a meaningful way as either time domain or frequency domain representation. There are a great many buttons and knobs, and in a modern machine, a great many menu choices. So it is unlikely the correct combination will be found by trial and error. This is where the operator needs knowledge in some detail of the instrument’s system architecture.
Following the pre-amplifier, a bandpass filter, also known as a pre-selection filter, is deployed for the purpose of blocking unwanted frequencies from downstream circuitry. This filter is not present in all instrument models, and there is considerable variation in product designs. A part of the rationale for its inclusion is to reduce false peaks that may appear in the final display.
At this point it is typical to see dc blocking, which is effective in preventing damage from overload in circuits that follow. Just as a skilled pool player plans shots with a view to limiting the opponent’s options, the electronic imagination is greatly concerned with the effect of a design element on downstream circuit elements.
The mixer is central to all superheterodyne designs, both in radios/TVs and in measurement instrumentation. The mixer, as its name implies, combines the RF from the frequency analyzer input with the tone synthesized at the local oscillator, producing a new signal of intermediate frequency. This IF resultant is comprised of the sum and difference frequencies with respect to the two inputs, as well as their harmonics.
It is common to see a number of mixing stages. They are arranged in cascade to fine-tune the degree of resolution and to establish a good frequency range. Multiple mixing stages, each coupled with well-designed filtering, have the effect of increasing sensitivity, frequency stability and selectivity at higher frequencies.
Superheterodyne methodology depends upon the presence of a local oscillator (LO). There are design variations. A common one involves the use of a voltage-controlled oscillator (VCO). Here, an increase in applied voltage results in a higher frequency at the output.
An essential part of the LO operation is its use as a dependable reference oscillator. The LO usually is based on a quartz crystal, which has the valuable property of vibrating precisely at a specific frequency. Quartz crystals were first used as very stable resonant oscillators in the 1920s. They outperformed tuned circuits that previously had been used to prevent frequency drift in amateur and commercial broadcast transmission. A properly cut quartz crystal vibrates at a stable frequency when electrical or mechanical energy is applied to it, and in this respect it resembles a conventional RLC circuit. The effect of temperature is minimal.
The output frequency of a non-crystal oscillator is adversely affected by temperature and humidity. The umbrella term for these cumulative effects is phase noise, which is harmful because it raises the noise floor and increases the difficulty of discerning small signal differences. The source of phase noise and heightened noise floor may be either the signal as applied at the instrument input, or it may be internal, within the spectrum analyzer. Phase noise may be mitigated by ensuring the integrity of the reference oscillator or by protecting the instrument from variations in the environment.
A major transition in the signal path is the analog-to-digital converter (ADC). This stage samples the analog signal and renders a digital representation that contains the information required to precisely characterize the electrical waveform of interest. To accomplish this task, the sampling rate must be at least twice the frequency of the analog signal, so as to comply with Nyquist requirements.
Following the ADC, the resolution bandwidth (RBW) filter may sit either as the final stage of the IF filter or elsewhere downstream. Like all bandpass filters, the RBW filter passes a continuous portion of the spectrum, blocking higher and lower frequencies. A graph of permitted frequencies generally conforms to a Gaussian shape.
Spectrum analyzer designers endeavor to minimize measurement error by lowering the noise floor through various means such as reducing the RBW, but they continually encounter tradeoffs. For example, narrowing the RBW boosts the time needed to scan any given frequency segment.
Real-time and swept spectrum analyzers, following the ADC, have much in common, including most prominently the flat screen technology that has, for the most part, replaced CRTs. Both instruments, together with the oscilloscope, will undoubtedly continue to undergo further development in years to come, providing insights into both quantum and relativistic realms.
Leave a Reply
You must be logged in to post a comment.