How does image rejection work in a superheterodyne receiver and what determines the image frequency?
Understanding the Image Problem
The image frequency is a fundamental consequence of the mixing process. A mixer produces an output at the IF frequency for any input signal at fRF = fLO ± fIF. For low-side injection (fLO below fRF), the desired signal is at fLO + fIF and the image is at fLO - fIF. These two frequencies are separated by 2×fIF, and they both produce identical IF outputs. Without some mechanism to distinguish them, an interfering signal at the image frequency will appear as an in-band signal at the IF.
The most common image rejection approach is a preselector bandpass filter between the antenna and mixer. This filter passes the desired RF band while attenuating the image frequency. The required filter rejection depends on the signal environment: 60 to 80 dB is typical for high-performance receivers. The filter's effectiveness depends on the IF frequency because a higher IF places the image further away, making it easier to filter.
Image-reject mixer architectures provide additional rejection by using two mixers with quadrature LO signals and combining the outputs to cancel the image. The Hartley architecture uses a 90° phase shift network; the Weaver architecture uses a second frequency conversion. Practical image-reject mixers achieve 20 to 35 dB of image suppression, which supplements the preselector filter.
Frequently Asked Questions
How do I calculate the image frequency?
For low-side injection: f_image = fLO - fIF = fRF - 2×fIF. For high-side injection: f_image = fLO + fIF = fRF + 2×fIF. The image is always separated from the desired signal by exactly 2×fIF.
Why not just use a very high IF?
Higher IF provides better image rejection but at the cost of more expensive IF components, wider IF filters, and potentially more spurious responses. The optimal IF balances image rejection against these practical constraints.
Does direct conversion have an image problem?
Not in the traditional sense. In a zero-IF receiver, the image is the signal itself (mirrored around DC). I/Q imbalance causes incomplete cancellation of this self-image, but this is managed through digital correction rather than RF filtering.