In designing a communication receiver for a particular range of frequencies, one must look at the natural noise power contribution anticipated at the antenna of such receiver for the range of interest. This tells the designer that the receiver should be quiet enough regarding its internal noise (noise figure) to hear this ambient noise. In the electromagnetic spectrum, noise contribution and its level varies considerably with frequency, type of source, location and season of the year.
DETERMINING NOISE FIGURE REQUIREMENTS
Fig. 1
The graph in Fig. 1 shows the overall external noise power contribution (dB above KT. )* for the frequency range of 10 kHz to 10 GHz. This information is based on measurements made with an omnidirectional vertical antenna over a perfectly conducting ground.** The sky noise data and the solar noise data were obtained with a directional antenna pointed at the sources. Maximum and minimum readings were provided in the case of atmospheric and man-made noise as shown, and Equation 13-1 was used to find the available noise power F.
F = P /KT B = T /T [Eq. 1]
Where: F. is Effective antenna noise factor
P is Noise power available from an equivalent loss-free antenna (WATTS)
K is Holtzman constant = 1.38 x 10^-23 joules per degree Kelvin T . is Ref. temperature, taken as 288° Kelvin.
B is Effective receiver noise BW (Hz)
T o is Effective antenna temperature in the presence of external noise (degrees Kelvin)
-------
•Where K = Boltzman's constant = 1.38 x 10^-23 and t o = 288° Kelvin.
**CCIR Report 322. "World Distribution and Characteristics of Atmospheric Radio Noise" 1CSh Plenary Assembly. Geneva; 1963.
---------
In HF, as seen in Fig. 1, things are quite noisy. Looking at the overall picture between galactic, atmospheric, and man-made noise, the external contribution to an HF receiver is such that the lowest noise level expected is about 18 dB. This is the determining factor in choosing the noise figure of our receiver. An 8 to 10 dB noise figure is typical in such a case, and anything quieter is not practical in this application. If a vhf/uhf receiver is contemplated, lower noise figures should be used. The limiting factors for such receiver would be the galactic and solar noise, as shown.
Designing for a given noise figure in a radio receiver involves all circuits, from mixers to amplifiers and local oscillators, with mixers being the predominant contributors. The noise figure of a receiver is expressed by the total noise added by this circuitry. An expression for defining this parameter in terms of signal-to-noise ratios is given by equation 2.
NF = 10 Log pS/pN [Eq. 2]
PS /PN o o
Where: NF is Noise figure in dB
PS. is Signal power at input
PN is Noise power at input
PS,: is signal power at output
PN. is Noise power at output
SENSITIVITY
Sensitivity is a way of expressing the ability of a radio receiver to detect a signal of a certain level. There are many ways of defining how sensitivity is calculated. Included are tangential sensitivity, signal-noise, output signal-to-noise ratio, false-alarm rate, bit-error rate, probability of detection, etc. The type of method chosen usually depends on factors, such as type and degree of modulation, i-f bandwidths employed and type of detected output as well as the receiver's noise figure. We will deal with two simple ways of expressing receiver sensitivity. Equation 3 shows one method of determining sensitivity.
S =- 174 dBm + NF + 10 log i-f-BW + K.. + K. [Eq. 3]
Where: S is Sensitivity in dBm
- 174 dBm = KTB = Thermal noise power in a one-Hz bandwidth at room temperature
NF is Noise figure in dB
i-f-BW is Pre-detection i-f bandwidth in Hz
K is Desired S+N/N in dB of the detected signal
K is Variable expressed in dB which is a function of type of modulation used
As it can be seen from this equation, sensitivity improves (becomes smaller) with decreased NF and i-f bandwidth.
Another method of measuring sensitivity of a receiver is expressed by the signal input expressed in (uV), necessary to produce an audio output 10 dB greater than the noise figure of the receiver. For example, if the specification reads: 1 AV for 10 dB S/N, it means 1 microvolt of signal at the antenna (30% amplitude modulated) will be heard at the speaker 10 dB over the internal noise of the receiver. This method is usually used in specifying amateur and commercial receivers.
In conclusion, sensitivity can take many quantitative forms. One fact that must be kept in mind is that sensitivity varies throughout the frequency coverage of a communication receiver because of gain variations in the front end. Good manufacturers will usually publish sensitivities at several points in the frequency coverage. See Figs. 13-2 and 13-3 and Tables 1 and 2.
* Chutes E. Dexter and Robert D. Glaz. Watkins-Johnson Co.. Tech-notes RF receiver design. Vol. 5 No. 2 Much/April 1978.
Fig. 2. The EK 070 communications receiver is a fully synthesized,
double conversion approach with a first i-f of 81.4 MHz and a second
i-f of 1.4 MHz. This receiver covers the frequency range of 10 kHz to
30 MHz in 10 Hz steps, and includes 30 memories for computer controlled
scanning. Another feature of this receiver Is the logarithmic gain provided
by the i-f amplifier which is a direct reading of the signal level at
the antenna when the receiver is AGC'ed. This control voltage is transformed
into a digital number which is displayed on the front panel to indicate
the signal strength of the received signal (courtesy of Rohde & Schwarz).
Table 1. Specifications of the EK 070 Vlf/Hf Communications Receiver.
PREAMPLIFIERS
Table 2. Specifications of the SR 219 A/AFC Receiving System.
Fig. 4. Communications receiver front end, has a high intercept point
and a low noise figure, due to the careful choice of devices.
Fig. 3. The SR 21 9 A/AFC receiving system can cover the frequency
range of 20 MHz to 4 GHz, with proper rf plug-in tuners. The receiver
is used for reception of AM, FM, cw and pulse signals and can be operated
as a fixed or portable unit (courtesy of Norlin Communications).
In order to keep intermodulation products at a low level, good receivers use a minimum of preamplification, usually just enough to compensate for the loss in the preselector filters and the first mixer. Fig. 4 shows a block diagram for the front end of such a receiver. This arrangement can provide a third-order intercept point of +20 dBm and a typical noise figure of 12 dB, as shown' From this example, it can be seen that in order to achieve these specifications, a very high level mixer was chosen requiring +27 dBm of local oscillator drive. The Watkins-Johnson M9E can provide a third-order intercept point of greater than +28 dBm. (The typical doubly balanced mixer LO drive requirement is +7 dBm).
Fig. 5. Good signal handling capability is maintained in this front
end arrangement which has no preamplifier.
Fig. 6
Fig. 7
A better and more economical approach is shown in Fig. 5. By eliminating the preamplifier, and by reconsidering each stage for noise figure, the signal-handling capability of the entire front end is maintained with less local oscillator drive. This last approach is used in the WJ-8718 HF communication receiver manufactured by Watkins-Johnson.
If a preamplifier must be used, typically a JFET circuit is chosen for frequencies below 300 MHz. Bipolar transistors can yield low noise figures up to 2 GHz. Above this, GaAs FETs are used to achieve low noise figure performance. Fig. 6 shows the noise figure performance for these types of devices for the frequency range of 10 MHz to 3000 MHz.
Fig. 7 shows two different approaches to high-dynamic range preamplifiers as used in a single-conversion communication receiver with a 9 MHz i-f. They provide a typical 2 dB noise figure. The CP650 is manufactured by Teledyne-Crystalonics and consists of fifteen FETs in parallel on a single chip which allows for a transconductance (gm) in the order of 150,000
µmhos (typical 100,000 ¡mhos). The input impedance of the circuit at A (1/gm) is about 20 ohms, resulting in a mismatch to the 50 ohm antenna (1 dB loss) which in turn improves the noise figure over the range by approximately 0.5 dB. The output is a combination low-pass filter, and matching autotransformer with the output set at 50 ohms for the entire frequency range.
Fig. 7 shows a push-pull arrangement which uses two uhf power transistors (2N5109) manufactured by RCA. This low-noise arrangement uses emitter feedback and provides about 12 dB of gain with a noise figure of 2 dB, and an input intercept of +22 dBms.
*Ulrich L. Rohde - Optimum design for high frequency communications receivers. HR. Oct. 1976. Ulrich L. Rohde, Eight Ways to Better Radio Receiver Design-Electronics. Feb. 1975.
Fig. 8. Bandpass preselector filters can be switched at the input
of a general coverage communications receiver.
Fig. 9. Half-octave filters designed to be used in the preselector
of modern communications receivers (courtesy of Communication Coil Co.).
Fig. 10. Block diagram for a typical preselector assembly using halt-octave
filters, showing signal flow with command line # 3 activated (4-6 MHz).
MODERN PRESELECTORS AND THEIR SWITCHING
General coverage communication receivers today depart from the traditional preselectors described earlier by using banks of switchable bandpass filters. The switching is usually accomplished with the use of low leakage diodes (IN458) or miniature relays. Signal diodes such as the IN914 or the IN4148 are usually avoided.
Fig. 11. Circuit diagram for a typical implementation of a half-octave
preselector assembly.
Fig. 12. Typical bandpass characteristics for the half-octave bank
of filters used at the input of a modern communications receiver.
Fig. 13. The Rockwell International 651S-1B is a fully-synthesized
communications receiver equipped with an automatically switched preselector
(courtesy of Rockwell International).
Because diodes are not perfect switches, and because of their non linear nature, they can contribute to the deterioration of the intercept point of the receiver. Another factor that can contribute to the degradation of the dynamic range in a communication receiver is the size of the iron-powder toroid cores usually used in the construction of the preselector filters. It has been found that larger cores are superior to small cores from an intermodulation distortion point of view.
Although general filter design theory is outside the scope of this guide, Fig. 8 shows a schematic diagram for several bandpass filters that could be used at the input of a general coverage communication receiver.
Fig. 14. Internal view of the 651S-1 communications receiver (courtesy
of Rockwell International). Preselector filters in modern receivers are
usually of the half octave*, Bessel design type, in order to keep second-or
higher-order products out of the pass-band characteristics of the receiver's
front end as well as to provide steep attenuation beyond the cut-off
points. Fig. 9 shows several half-octave preselector filters manufactured
by Communication Coil Co. for this purpose.
The block diagram in Fig.10 shows how such filters can be incorporated in the design of a front end for a communication receiver, and Fig. 11B shows a typical circuit design for this stage. This design shows how the half-octave filters can easily be switched in by digital logic levels that can be provided by the receiver's synthesizer or digital frequency display.
Fig. 12 shows the total composite frequency characteristic for the half-octave filters from 2 to 30 MHz.
These examples show that the old manual "switch and peak" method of preselecting has been surpassed by this "hands off" method. While until recently this method only applied to expensive communication receivers, new lower-cost versions of such preselectors have been introduced by receiver manufacturers especially in Japan. As technology evolves this method of pre-selection will become the rule rather than the exception in the design of communication receivers of all classes. See Figs.13 and 14.
*Musical term applied to the frequency domain describing the interval between eight notes, or a basic frequency ratio of two.
Example: one octave = 2 to 4 MHz. ti octave = 2 to 3 Mils, etc.