• Background

The earliest transmitters, Spark Gap transmitters (or carrierless transmitters) propagate an extremely wideband signal composed of many frequencies.  A coil of wire, a diode and an earpiece can be used as a rudimentary receiver for this type of transmission.  This is fine for sending low data rate information like Morse code.

Eventually, people developed carrier modulated transmission systems.  The benefits being the ability to transmit information at higher data rates, the ability to transmit complex data (like voice), physically realizable antennas and the ability to modulate many baseband data streams onto one RF carrier.  This ushered in the era of AM (Amplitude Modulated) and FM (Frequency Modulated) receivers.  Next, various forms of Digital Modulation were (and are being) developed to support higher system capacities, improved quality via digital coding techniques, etc.

• Receiver Evolution

In order to support Analog and Digital modulated transmission systems, some sort of processing is required to extract the desired information from the carrier signal.

• Tuned Radio Frequency (TRF) Receivers

This class of receiver is actually not conversion based. TRF receivers are based on a series of tuned amplifiers back to back leading to an audio detector. The amplifiers successively filter and amplify the desired signal until it reaches an acceptable level for detection. The problem with this design is that it is difficult to achieve high gain at the relatively high frequencies of the carrier without oscillation. This necessitated the use of many lower gain successive stages to produce enough ultimate gain. It
was still difficult to preclude oscillation if there were many amplifiers on at the same time.

• Amplified Sequenced Hybrid (ASH) Receiver

This class of receiver is a variation of the TRF receiver. In the ASH receiver, rather than cascade the amplifiers in series, they are switched in and out of the circuit to preclude the oscillation that would occur if the signal passed through all the amps together. As the signal passes through each amplifier under control of the switch, the signal that had passed through the previous stages must be stored in some way (a delay line). In this fashion, there is never more than one amplifier on at a time.

An advantage of the TRF and ASH receiver architectures is that as they have no local oscillators or radiating sources, they emit no RF interference, and hence do not require FCC approval.

• Super-Regenerative Receivers

This class of receiver is actually not conversion based either. They function on the buildup/decay of an oscillator. Positive feedback was used to take an amplifier to the brink of oscillation. The amplifier oscillation was allowed to decay and then the process was started again. This created, in essence, controlled oscillation which was used to tune the receiver. Developed by Armstrong in 1912.

While this class of receiver provided a large gain (as compared to TRF receivers) and obviated the need for many amplifier stages, it was difficult to tune and very difficult to control.

• Homodyne Receivers (Direct-Conversion Receivers or Zero-IF Receivers)

This class of receiver converts the RF directly to baseband for processing. This type of receiver will be covered in detail in a future app note.

• Low IF Receivers (LIF Receiver or Near Zero-IF Receiver)

This class of receiver is similar to the Zero-IF design except that the RF is converted not directly to baseband, but to some low (relative to RF) frequency. There are some advantages to this over the Zero-IF approach, like mitigating the problem of DC offset to some degree.

• Correlation Receivers

In a classical receiver, the detector forms an estimate of the transmitted symbol by making a decision (based upon various decision rules) on an observation vector.  If the detector makes it's decision of the transmitted symbol estimate based upon more than one observation vector, then the structure is called a "correlation receiver".   Detection and Estimation theory will be covered in more detail in a separate app note.

• Super-Heterodyne Receivers

Developed by Armstrong in 1918, these are single or dual stage conversion based receivers, where the RF is converted to IF in one or more stages.  This method provides better sensitivity and improved selectivity over Super-Regenerative Receivers and Tuned Filter Receivers.

The sensitivity of a receiver can be approximately expressed as (see Modern Communication Circuits, by Smith, pp.82):

S_i = F * K * T * B * (S/N)_o in the linear domain

where:

S_i = Sensitivity

F = Noise Factor (linear)

K = Boltzmann's constant

T = temperature (in degrees Kelvin)

B = Bandwidth

Alternatively:

S_i (dB) = NF -174 + 10 Log (B) + 10 Log [(S/N)_o] in the logarithmic domain (at room temp.)

where:

NF = Noise Figure (in dB)

Two things should be kept in mind in connection with the formula for Rx Sensitivity.   Firstly, the signal must be strong enough to manifest at the detector (even in the absence of channel noise) above a required threshold, to overcome the Noise Figure (which is the noise induced by the radio system).

Secondly, the S/N ratio must be high enough (for the given modulation scheme) to decode acceptable BER.  In other words, the signal must be greater than the noise contribution.  This means that even if the signal power is high, it must be high in comparison to the noise power.

To a large extent, the communication link performance is gated by (for a given bandwidth) the Rx Sensitivity and the Receivers ability to handle interference (selectivity).

Receive Sensitivity is in turn predominantly gated by NF and Rx front end gain (LNA).

• Benefits of Superheterodyning

A local oscillator signal (usually at a slightly higher frequency than the desired tuned frequency) is mixed with the desired signal to produce sum and difference frequencies.  While the sum frequency signal can be filtered out and discarded, the difference frequency (Intermediate Frequency) is used for further processing.

When the carrier RF signal is downconverted to IF, the modulation information it carries is also downconverted to IF.

Since the LO signal changes in accordance with how the receiver is tuned to the desired frequency, the frequency spacing between the two is always the same and therefore the IF frequency is kept constant.

There are several advantages to the design since the IF frequency is kept constant.

1. Can optimize the amplifier for higher gain without incurring oscillation because operating at a lower frequency (IF).
2. Can optimize for improved selectivity (interference rejection).   For a given percent bandwidth, the range of frequencies passed at resonance is smaller for a lower center frequency.  For example, given a certain % bandwidth filter specification, close in interferers will be passed through at the RF carrier frequency, while those same interferers, after downconversion, will fall outside of the IF % bandwidth filter passband.  Remember, the LO is not retuning to mix with the interfering frequency, so the interferer will mix down to a higher value at IF than the desired frequency.
3. Since the IF is at a fixed frequency, can standardize on components,  realize economies of scale and savings on manufacturing costs and realize less variability over time and temperature.

• Tradeoffs

The choice of the frequency chosen for the final IF stage as well as the bandwidth chosen for the filters used in the IF stages represent engineering tradeoffs.

In connection with the center frequency chosen for the If stage;

Two common frequencies to employ in commercial designs (with readily available off the shelf components) are 10.7 MHz and 455 KHz.  On the one hand it is preferable to choose the lower frequency of 455 KHz as the IF frequency because it is easier to achieve higher gain at the lower frequency without incurring instability (as compared to 10.7 MHz.).  Use of  10.7 MHz  for the IF frequency would mean lower gain and this would necessitate the use of higher gain at the RF stages.  This would require RF mixers with increased dynamic range.  This would mean higher cost.  Secondly, at lower IF frequencies, crystal filers can be used to provide selectivity which again translate to lower cost.  Additionally, it is also easier to sample at lower frequencies.

On the other hand, it is preferable to choose an IF frequency of 10.7 MHz (as compared to 455 KHz.) as the Image frequency falls further away from the center frequency (more rejection) at the higher IF.

In connection with the bandwidth chosen for the If filter;

A smaller IF bandwidth improves both the sensitivity as well as the selectivity of the receiver.  However, a smaller bandwidth has associated with it a longer settling time (more ringing in the time domain) and thus implies a longer switching time when tuning to a new frequency.

On the other hand, a larger IF bandwidth has associated with it  the ability to handle higher data rates.

There are many other tradeoffs associated with the choice of Receiver architecture.   There are considerations like design for size, cost, power consumption, operating channel, frequency plan, interference environment, etc...

There are also lower level design considerations like whether to design the impedance for optimal Noise Figure (improve Rx Sensitivity) or to design the impedance for optimal power transfer (from the antenna).  The two design targets are not the same.  As in all other aspects of engineering, there is a trade-off.  Typically, one must use a Balun.  This does not even address the issues of designing the Noise Figure for optimal gain, optimal BW or optimal stability.

The next app note in this series will focus on Direct-Conversion Receivers