A 20 Input Audio Mixer for Multiple Receiver Setups

Many of us have more than one receiver, especially those of us who are afflicted with the Boatanchor bug.  In fact, some of us might reasonably be accused of being obsessed.  Of course, I’m sure no one would ever use that term about me and my 27 receivers.  Still, I do like all my receivers and I do like to use some of them simultaneously.  Sometimes I listen to different things at the same time and other times I like to compare the reception of the same signal on various different receivers – classic A/B testing.

One of the more common configurations around here is for all of the following to be happening at the same time:

Another time might find me comparing various receivers:  say a 75S-3B, then a 75A-4, then the SX-115, then an HQ-170, and then the Flex 5000.  Because of this audio mixer, I can compare the various receivers just by twiddling the gain controls.  No switching required.  Makes A/B comparisons a snap. (They are all driven by the same antenna).

All of this used to be easy – I just used a different speaker on each radio.  However, once the receiver count soared up through 6 or 7, this solution started to lose its luster.  Not only the cost, but also the space, consumed by all these speakers seemed to be getting out of whack.  So another solution began to seem attractive.  I put up with various frustrating switching arrangements for years before I got around to doing anything about it.

This desire to listen to multiple rigs at once led me to develop a 20 input audio mixer.  I run the audio from each receiver to the mixer, which then drives one speaker – typically a Collins 312B-3.  This allows me to listen to whatever radios I want without worrying about any switching.  I still have several speakers to choose from, but whichever one I use is driven by the mixer – and, therefore, all the receivers.

Some of the mixer’s inputs are 8 ohms and directly accept amplified outputs that would normally be connected to speakers.  Other inputs are 600 ohms and accept line level signals.  Some receivers, like the R390A, have only a 600 ohm line level output.  Others – such as the Collins S line - have both types of outputs.  I use whichever outputs exhibit less distortion.

The design has been very successful.  The mixer provides adequate output power (up to a couple of watts) with very low noise, and is nearly bullet proof with respect to RFI.  The noise delivered to the speaker when all the radios are connected but turned off is only about 6 microwatts, which is just barely audible when I put my ear directly on the speaker grill and the room is quiet.  I tried to keep leads very short in the vicinity of the circuit's most vulnerable areas - the inputs at each amp and the power connections to the amps.  It seems to have quite well.  However, I still find myself thinking about looking around for even lower noise amps.  But I do realize that noise reduction then might be becoming an obsession.  Oh, well.

The schematic of the mixer is shown in Figure 1.

Figure 1

Each input and output is protected from RFI with a ferrite bead and a high quality bypass capacitor.  The speaker level input is terminated with an 8 ohm 5 watt resistor.  Then each signal is connected to the summing junction through a resistor.  The value of this resistor sets the gain for that input, so the value of the input resistors are different for each of the two types of inputs – reflecting the different voltage levels presented to those two types of inputs at any given reference level of sound pressure.

The Summing amp stage:  These input resistors are all connected to the inverting input of an NE-5534 op amp (U1).  The NE-5534 datasheet is available here.  I chose this op amp because of its reputation for low noise in audio applications.  One of the characteristics of an op amp (assuming it is “happy”) is that it functions as a “nullator”.  It forces the voltage across its inputs to be "nulled out" - to be essentially zero.  This interesting characteristic simplifies circuit analysis.  Since the non-inverting input is held at half the power supply voltage by R7 and R17, this means that the inverting input of the op amp must also always be at half the supply voltage.  This means the current through each of the input resistors is the same (assuming the same input signal level).  Therefore, we have a perfect summing junction and the gain (at least for the first stage of amplification) for each input is easily calculated – it is simply the ratio of the feedback resistor R8 to the input resistor.  I set the gain to 1 for the 8 ohm inputs and about three for the line level inputs.  These gains produce about the same sound pressure levels.

C16 rolls off the response at about 16khz.  The output of U1 swings around a DC level equal to half the supply voltage.  As a result, the next stage must be ac coupled, which is what C12 does.

The Power amp stage:  I chose an LM380 for the power amp.  Admittedly, I didn't have a lot of good reasons for picking this chip other than the fact that I'm familiar with it (National Semi is my alma mater)  There are better ones around now.  This amp has a fixed voltage gain of 50 and can supply up to 2.5 watts of audio power.  I wanted the nominal gain of the whole system to be one, so it was necessary to offset the gain of the LM380.  This is done by the resistive divider formed by R1 and R6, which reduce the signal from the first stage by a factor of 50 before it is applied to the LM380 input.  I chose very low values for these two resistors to keep the impedance at the input of the power amp as low as practical in order to reduce noise and RFI susceptibility.  This preserves the overall gain of the system such that the mixer will produce the same sound pressure level from the speaker that the input signal would have produced if connected directly to the speaker.  C10 provides output AC coupling to the speaker.  C11, C42, and L2 are for RFI protection.  I used metal film resistors throughout the entire mixer because they generate less noise than most other types, especially carbon resistors.

It’s basically a simple circuit repeated 20 times to get 20 inputs.  I probably should have made it with 24 inputs while I was at it.  I may go back and add more inputs, but it will be tougher now to drill holes in the enclosure without messing up my nice paint job.

A photo of the completed unit is shown below.  Since these were taken, I have added ferrite beads to every signal coming into or leaving the box for more RFI protection.  You can see that I had one of these ferrite beads on one input when this picture was taken (near the top center).  The wire makes 3 loops through the ferrite bead, which is from Palomar Engineers, model FB63-77.  The beads, which form RF chokes, made a huge difference.  It turns out that certain of my receivers pick up lots of RFI (even when they are turned off).  So if they are plugged into the mixer I was getting RFI into the mixer via the input from the offending receiver.  The ferrite beads and a .1 uf cap to ground just after the 8 ohm load resistors really killed it.  The enclosure is a Hammond 1590WFFBK, available from Mouser.  I like it because it is diecast and has mounting flanges.

Also since this photo was taken, I have wired each ground lug on each connector to ground.  I didn't have any trouble with these, they were all grounded well to the case simply by being screwed down against the inside of the case, which is unpainted.  But I did have some issues of that sort with another project.  I also realize that the surface of the aluminum case will oxidize over time and this may compromise those mechanical-only ground connections.  So I overcame my innate laziness and impatience and went back and grounded each connector's ground lug directly.  Now the aluminum can oxidize all it wants but it won't have any effect on operation.

I used the "ugly" method of construction.  It was a lot faster than laying out and etching a PCB.  It works very well, partly because the whole unetched PC board, which serves as a substrate for this, is a ground plane.  The unetched PC board material was cut to size and then "tinned" with a product called Liquid Tin.  This is just like tinning the board with solder.  It keeps the copper from corroding, makes soldering to it easy, and preserves the ability to solder to it in the future (whereas copper will oxidize and will then not take solder).  This style of construction makes it very easy to change things.