Click here to go to our main page on microwave receivers
Click here to go to our main page on mixers
Click here to go to our page on noise figure
Click here to go to our page on low noise amplifiers
A superheterodyne receiver contains a combination of amplification with frequency mixing, and is by far the most popular architecture for a microwave receiver.
To heterodyne means to mix two signals of different frequencies together, resulting in a "beat" frequency. Actually, two signals are always created, the sum frequency and the difference frequency. These are referred to as the two sidebands. The sum frequency is the upper sideband, and the difference frequency is the difference sideband. In most microwave receivers, the upper sideband is ignored.
The word "superheterodyne" is often hyphenated to "super-heterodyne". A common contraction of the word is simply "superhet".
Here's a clickable index to our super material on superhet!
The superheterodyne receiver is still the most popular microwave receiver, and it was invented during and directly after the Great War and patented in 1918. Edwin Armstrong (see him in our Hall of Fame) was truly one of the great minds of the twentieth century. Unfortunately, he had a bad habit of inventing stuff independently of the corporate anti-Christ of the time, which was RCA (Radio Corporation of America); RCA and others pretty much ruined Armstrong's life. From this perspective, nothing's changed, big companies ruin many engineers' lives today, especially in arguments over inventions. As an engineer you'd best remember your place at the bottom of the corporate food-chain or you'll be living in a cardboard house behind the Circle K pretty soon. Don't ever think your services are as valuable as the work of more deserving corporate lawyers! But we digress....
The word "heterodyne" was coined by Reginald Aubrey Fessenden, also in our Hall of Fame. Heterodyning is the same as mixing, we offer a little history on the topic here. Although Fessenden was able to receive signals by mixing them, there is no evidence that he used a detector to sort out the baseband; therefore Fessenden only managed to put together just one of the five pieces of Armstrong's superhet receiver (RF amp, mixer, local oscillator, detector and audio amp). Order "Man of Fidelity: Edwin Howard Armstrong" by Lawrence Lessing on our book page to corroborate that fact.
Prior to the superhet receiver, radio listeners had to constantly play with a set of knobs on a radio to keep locked on to radio programs of the day. Prior attempts at amplifying radio signals included the regenerative and super-regenerative receivers (both invented by Armstrong, but stolen by greedy RCA), which at least offered some gain to boost far-away signals, but were prone to drift. The superhet solved this problem for all time by using frequency conversion that was controlled by a local oscillator. By tuning the LO frequency, the RF frequency that you received was thereby tuned. It turns out it is much easier to control the LO frequency and filter out unwanted channels at IF, rather than try to filter out unwanted channels at the RF input. This characteristic is known as receiver selectivity.
To learn more about the history of radio, go to our book page and click on the link to Empire of the Air.
The block diagram below shows typical components of a superhet microwave receiver. The pre-LNA components all have a direct affect on noise figure, so low-loss is a key characteristic of these parts, because loss adds directly to noise figure. Often, waveguide components are used in front of the LNA because it offers the lowest loss available. Stuff after the LNA doesn't have such a drastic affect, but everything must be taken into account in maximizing dynamic range.
This component restricts the frequency band that is permitted to enter the receiver.
Limiter (receiver protector)
This component is what protects the LNA if a stray high-power signal makes its way into the RF input. This function can be accomplished by a passive or an active limiter. An active limiter, also known as a blanking switch, must be commanded somehow to protect the LNA. This is often the case of a co-located transmit/receive system such as a radar. A passive limiter doesn't need to be commanded, it automatically reduces the signal whenever it is hit with enough power.
The power that leaks out of the receiver protection circuit must be specified and measured under both pulsed conditions (spike leakage) and CW condition (flat leakage).
A pre-LNA switchable attenuator can be used to increase the dynamic range. When signals are strong enough to saturate the LNA, the attenuator can be switched in to reduce the signal strength to the LNA. The noise figure increases by the attenuation value, unfortunately.
Low noise amplifier
Nothing is more critical to receiver performance than the LNA. This is the active component that increases the power of the received signal (it produces gain). The power handling of the LNA means how much power can it be hit with before it is permanently damaged. Typical MMIC LNAs can handle from 10 milliwatts to half a watt. The damage to an LNA is not necessarily catastrophic, only a small change in gain can result from overd-riving it. The noise figure usually does not degrade by much unless the gain of the device is reduced by several dB. Designing an LNA for high-power use is called "hardening".
Image rejection filter
Used to reduce image noise foldover.
The device that converts the incoming RF frequency to intermediate frequency is called a mixer.
The cleanup IF filter's purpose is to remove unwanted high frequency signals such as the RF and LO, which could cause additional distortion products when they pass through the IF amplifier. It is usually a low-pass filter (LPF), and could use low-cost lumped components (inductors and capacitors).
You should always design a receiver with extra gain, then you can used attenuators here and there to "pad" out poor VSWRs. Places that are notorious for VSWR problems is the mixer/filter interfaces.
This is where gain can be added to the receiver without high cost. At one GHz, gain can be purchased for less than 10 cents per decibel!
The local oscillator signal must be stable (time-invariant) and maintained at a fixed or minimum power for the mixer to work properly. Often an amplifier stage is put in the LO path, and kept in compression to reduce variation in LO power coming from the exciter.
Advantages of superheterodyne receivers
The advantages of superheterodyne receiver are many. An obvious advantage is that by reducing to lower frequency, lower frequency components can be used, and in general, cost is proportional to frequency. RF gain at 40 GHz is expensive, IF gain at 1 GHz is cheap as dirt.
The second advantage is in the superior sensitivity that we almost take for granted. Filtering out unwanted signals at IF is a much easier job than filtering them out at RF, because the desired bandwidth is much higher after the signal is mixed down.
Further advantage in that many components can be designed for a fixed frequency (and even shared between different receiver designs), which is easier and cheaper than designing wideband components.
Single down-conversion versus double downconversion
A receiver that has only one frequency conversion device (mixer) is said to be of single-downconversion variety. Quite often a high performance receiver has two down-conversion steps. In this case we would label the frequencies RF, LO1, LO2, IF1, and IF2. There are advantages to the second downconversion approach. Having a higher IF1 frequency splits the RF and image frequencies far apart, which makes an easier task of preventing image noise foldover. For example, if your RF frequency was 10 GHz, and your IF was 3 MHz, in order to filter out the image, you'd need a- filter that passes 10 GHz yet rejects 10.003 GHz. This is for all intents impossible. But if your IF frequency was 3 GHz, it is quite easy to filter out incoming noise at 13 GHz.
There is a solution to image noise no matter how close the image is to the RF, that is an image rejection mixer.
A mixer is used in a receiver to create the IF frequency. Because a mixer creates both sum and difference frequencies,
The sidebands of a mixed signal occur at the sum and difference of the RF and LO signals. Thus a 12 GHz RF signal mixed with a 10 GHz LO will produce a lower sideband of 2 GHz and an upper sideband of 22 GHz.
Mixer spurs (MxN products)
This topic is coming soon!
A mixer is used in a superhet receiver to create the IF frequency. Because a mixer creates both sum and difference frequencies, by corollary, there are two RF frequencies that will produce the exact same IF frequency. The unwanted signal frequency is called the image frequency. An example:
RF is 12 GHz
LO is 10 GHz
Upper beat frequency is 22 GHz
Lower beat frequency is 2 GHz
Image frequency is 8 GHz
The RF passes through the mixer, and two possible outputs are (sum and difference) are 22 GHz and 2 GHz. We don't have to tell you that the IF is 2 GHz, do we? The 22 GHz is the upper side-band, and is discarded by filtering the IF (usually a low pass filter is all that is required). The image frequency that would produce a 2 GHz IF is 8 GHz. What's up with this? This is because an 8 GHz RF signal will mix to 18 GHz, and minus 2 GHz. Guess what? the minus sign is unimportant, there is no such thing as a negative frequency, is there....?
Negative frequencies do exist, it's just that we are all used to thinking of the magnitude of a frequency as would be displayed on a spectrum analyzer. This clarification came from Tim in May 2016, thanks!
Whilst it is probably true that frequency polarity is unimportant for the purposes of the example analysis, negative frequency does exist. It certainly has a mathematical definition and is of particular importance in relation to digital phase modulation.
The mathematical definition of negative frequency is a phasor with a negative complex exponent, that is to say, if a positive frequency phasor rotates anti-clockwise, a negative frequency phasor will rotate clockwise.
The significance from the point of view of phase modulation is that in a negative frequency image the quadrature component will lead the in-phase component, vice-versa for a positive frequency signal.
Twenty dB of image rejection is about all you need before you can neglect image noise. Worst case, image noise foldover can degrade receiver noise figure by 3 dB.
Image noise foldover can be prevented in two ways: use an image filter, or use an image-rejection mixer. The image rejection filter is located between the LNA and the mixer. If you place it in front of the LNA, it doesn't do the job.
1 1 1 1 1 1 1 1 1 1 Rating 4.06 (8 Votes)