Tech Forum

Considering the installation of a 'multi-station' intercom system wired throughout your premises can be somewhat of a task and perhaps a little more expensive than necessary as compared to the variety of 'wireless' intercom and phone systems already in use.


I opted to use a Uniden DECT 6.0 three-handset cordless phone system that offers excellent long-range wireless phone and 'conferencing' ability, selective between any/all three cordless handsets anywhere around my premises.


Although there are Uniden cordless phone units available with more handsets, this system can be purchased for a very reasonable price almost anywhere electronic equipment is sold. I carry one handset with me to any remote area, such as the garage, working in the yard, etc., and when my handset starts 'chirping' I can immediately identify [who] which handset is signaling me for conversation and, even if you don't connect to the phone line through the 'base set', this system can still be used as a long-range intercom system.

The intercom system desired is very similar to the headset intercoms used to communicate with camera operators and other studio personnel in every TV station and network for many years now. Although there are variations and expansions, the basis of these systems is a two-wire party line that carries both the conversation and the power (24 to 48 volts DC) to run the various stations so that no separate power connection is needed. They are probably similar to the two-station intercom circuits you have seen and many of these can be adopted for this use with the additional components shown in Figure 1. The basic idea is that each station must separate the audio and power components that are sent over the same two wires. Also, the audio must be blocked from entering the output terminals of the DC power supply because the filter capacitors in it will have a very low AC impedance and would short the audio to ground. There were also party line systems that used a four-wire connection to keep these two components separate, but I will ignore them here because a two-wire system was requested.

One wire is a signal and power ground, while the other carries both the audio and power. The basic audio circuit is simply an amplifier connected to the two-wire party line through a simple two-pole momentary switch which allows only one of these functions to occur at a time, thus preventing any feedback. I have shown a DPDT in the schematic to allow the speaker to be used as a microphone, but a separate mic could be used with a SPDT switch.

The coil labeled L1 is necessary to present a high impedance to audio frequencies on the party line which would be shorted by the filter capacitors in the power supply. It needs to have a fairly high inductance value in order to preserve the lower audio frequencies. You need whole Henries here, not milli or micro Henries. The formula for the impedance is 2*Pi*F*L and you want an inductive impedance value in thousands of ohms at about 100 Hz or lower. At 100 Hz for a 10 Henry (H) inductor: 2*3.141*100Hz*10H = 6283 ohms. This should work. It should be rated for the combined current of all the stations. The largest inductor I could find with a quick search is a 15 H coil but it was shown as "out of stock" with a delivery time of over 100 days. There are some being offered on eBay; search for "retard coil" and you will get what you need, but they are not cheap. There are high inductance chip-style inductors but I doubt they would work because they would not pass the DC current needed to operate all the stations.

In the stations, the capacitor labeled C1 blocks the DC voltage to the audio amplifier. It must have a voltage rating that is higher than the power supply output voltage and a low capacitive reactance in relation to the amplifier's input impedance. The formula for the capacitive reactance is 1/(2*3.141*F*C). If the amplifier you use has an input impedance of 10K ohms and you have 10 intercom stations, the combined parallel impedance on the party line will be 1,000 ohms; you will want a capacitive reactance of 100 ohms or less. Again, we calculate at the lowest desired frequency or about 100 Hz. A 15 µF capacitor will give us 106 ohms of capacitive reactance at 100 Hz and should work. I would probably step up to 25 µF or even 50 µf as the added cost will be tolerable and the low frequency performance will be improved.

Any of a number of solid-state or IC audio amplifier circuits could be used. Look for a high input impedance and a low output impedance suitable for driving your speaker. The gain needed will depend on the output level of the microphone or of the speaker when used as a mic and the desired line level on the party line. Since you want to use unshielded wire, I would suggest a line level of about +10 dBm or about three volts RMS in order to keep any noise to a minimum. Assuming the microphone level is -40 dBm, you would need a +50 dBm gain in the amplifier. I would go for one that provides +60 dBm to allow the volume controls room to work. The volume controls can be simple voltage division circuits with one side of the pot connected to the input signal, the other side to ground, and the output taken from the wiper; 10K or 50K audio taper pots should work with most types of amplifiers.

The figure below shows the DC power supply to have an output voltage that is significantly higher than the regulated voltages in the stations. This is because there are no filter capacitors on the party line. So, the regulator circuits in the stations must have sufficient voltage to stay in regulation, even during the loudest audio which will both add to and subtract from the average DC voltage provided by the power supply. So, with a three volt RMS audio level, half of the P-P value of the audio will be about 4.5 volts. This — on top of a supply voltage of 25 volts DC — will give a total voltage which will swing from 20.5 to 29.5 volts. If the regulator needs a three volt headroom to operate, this is subtracted from the minimum value of 20.5 to give a maximum regulated voltage of 17.5 volts. An 18 volt regulator would fall out of regulation on loud audio and cause distortion in the amplifier. Thus, there is a need for an unregulated voltage that is about 10 volts higher than the output of the regulators chosen.

Finally, a word about the type of wiring you wish to use. Telephone wires and Cat 5 network wires are relatively small gauge, so the power current will be somewhat limited. If you have multiple stations on a single run of such wire, the current needed for a single station will be multiplied by the number of stations on the run. This may cause problems with excessive voltage drops. Also, longer runs will add to this problem due to increased resistance in the wiring. The systems in TV stations I referred to earlier commonly only have to power headsets — not speakers — so less power is needed. I would suggest heavier gauge wiring; common lamp cord or even doorbell wire may work better. If you must use telephone or Cat 5 wiring, they commonly have four conductors and you should consider doubling up for both the ground and the signal/power conductors.

Think wireless!!! Go to gadgetshack.com and look at the Westinghouse two-channel basic home intercom system. For $59 per pair of units and a reputed "unlimited number" of units that can be used with full security, this is a lot cheaper and less of a "pain" than running cable. Running cable in an existing house is very difficult and installing cable in a new house is still expensive.