High Pass speaker filter for CTCSS suppression

Overview

This article describes a high pass audio filter for an external speaker on an FM transceiver, particularly for suppression of CTCSS tones.

Use of CTCSS has gained popularity as a means of providing some level of protection against intermodulation products with the progressive increase in the noise floor, and our tendency to weight intermodulation performance lightly when selecting receiver technologies.

Older transceivers even many newer transceivers that directly support CTCSS have quite good audio performance down below 100Hz, which results in annoying feed through to the speaker of the CTCSS tones, especially where higher frequencies are used. Low end performance is often suppressed by the choice of small internal speakers, but use with an external speaker with moderate low end performance will result in annoying audibility of the CTCSS tone on signals that use the same.

CTCSS tones range from 65Hz to 255Hz. Common practice in these parts is to use 123Hz, which I would guess is a compromise between audibility of the tone and audio intermodulation problems caused when certain types of modulators are driven to very high modulation indexes. (66Hz CTCSS at 500Hz peak deviation has a modulation index of 7.5, which is higher than the 2 to 3 caused by the speech component.).

Design

The filter is a T-section high pass filter with a cut off frequency of 300 Hz and nominal 18dB per octave slope for an 8 ohm load, using off the shelf components intended for hi-fi speaker cross over networks. The components used are an air cored 2.7mH inductor and two 100uF non polarised capacitors. The whole network is assembled in a die cast aluminium case with a bypass switch and spring loaded speaker terminals, although if space permitted, it could be built inside a speaker enclosure. The bypass switch allows the filter to be switched out for best fidelity on local signals without CTCSS.

Figure 1 shows the circuit as modelled. The 1.3 ohm resistor is not a discrete resistor, but to model the loss in the inductor (do not build it into a circuit). The circuit is just the two capacitors and choke.

Figure 2 shows the model amplitude vs frequency plot. The model indicates response at 123Hz is 19dB down on 1Khz, and 26dB down at 91.5Hz. Now the model is based on an ideal load of 8 ohms resistance, and practical speakers depart from that somewhat. Nevertheless, acoustic measurements gave similar results.

Figure 2: Modelled amplitude vs frequency plot

Implementation

Figure 3 shows an implementation as a stand alone for insertion between the transceiver and the external speaker. The components including a bypass switch are assembled in a small diecast aluminium box. The inductor is clamped between the box and lid with a 1mm thick rubber packing. The inductor has been attached to the box with silicone adhesive to prevent it moving around. The lid of the box forms the base, and has four self adhesive rubber feet affixed.

Figure 3: Prototype construction

Figure 4: Font panel with monitor jacks added.

Figure 4 shows the front panel with a pair of banana jacks added for monitoring receiver audio output.

Sourcing parts

The parts were obtained from Jaycar. Table 1 lists the parts and Jaycar catalogue numbers. Any equivalent parts may be used, but be careful to avoid lossy inductors.

Links / References

• LTSPICE model

Changes