Even quite expensive cameras still only have a 3.5mm stereo
socket to allow an external microphone to be connected.
This has two main limitations:
1. The 3.5mm socket and plug are not
very robust and are easily damaged. Unfortunately, damage
is at least as likely to be done to the socket as the plug,
so it pays to take extreme care to avoid the cable being
pulled, perhaps by tying it to part of the camera. By contrast,
the industry standard XLR connector is very solid and although
like the 3.5mm jack it still only has 3 connections they
are wired in a different way.
2. The input is Unbalanced. This makes
it prone to mains hum pickup.

The XLR and 3.5mm Jack plugs. Each
pin on the XLR is 2.3mm diameter and solid.
The outer body is metal so also screens the connections,
whereas a typical Jack is plastic bodied so exposes short
lengths of the cores to the open air (plastic doesn't
stop electro magnetic radiation)
In order to understand why a Balanced
cable (line) is superior at rejecting unwanted interference
it is necessary to have an appreciation of what is at each
end of the line.

In this Diagram (Fig 1) the Microphone
is represented as a coil, and wih no other connections made
to it each terminal will produce a signal which is antiphase
to the other end (Antiphase - a mirror reflection). These
signals vary with the sound that the microphone is exposed
to and they are represented above as a waveform which indicates
the voltage appearing at the terminal with respect to time.
However, to make use of this signal in
an unbalanced line we need to first establish a reference
point, normally called the Earth (or ground, Ov). When we
connect our reference point to one end of the coil this
has the effect of fixing that end to ground, so the level
(or Amplitude) at the other terminal is forced to double.
Figure 2 shows this.

The ground and the signal in Figure 2
represent the two wires that are in a typical piece of single
core screened cable. Any interfering signal such as mains
hum will affect both wires (the ground and the signal wire)
equally, but of course the ground is fixed to, err, ground,
so it can't carry a signal. So the interference appears
only on the signal wire. If the signal wire is carrying
a high voltage (say 10 volts) and the hum is only 1mV (1
millivolt, 1/1000 volt) then the ratio between them is 10000:1.
Ratiometric measurments in electronics are made with the
Bel, and usually quoted in decibels (1/10 Bel), 10000:1
is 40dB. So the hum of 1mV would be 40dB down on the signal
of 10 volts. Loudspeaker cables will pick up hum (as will
any cable) but the voltages driving the speaker are far
higher than the hum so it will not heard.
Unfortunately, microphones usually only
produce a few tens of millivolts, so 1mV of hum would become
a lot more audible...

This is a piece of high quality
Neoprene covered LOC (Low oxygen content) twin screened
audio cable. The screen is braided to give a more even coverage
(some cables have twin screens), the red and white cables
are stranded (more flexible) and it also has a core of string.
This gives the cable a lot of strength and also makes it
more flexible.
Now on to the right way to do it - the
Balanced transmission line shown in Figure 3. Here the two
signal wires are connected to each end of the transducer,
and an equal and opposite signal (antiphase) appears on
each core.

The Black wire in the diagram is connected
to a clean (I.e. not noisy) earth, usually at the other
end. It would usually also be connected to the microphone
body if it is metal - but it carries no signal, it just
acts as a screen. This can also have the effect of earthing
anyone holding the microphone, as a human body can also
reradiate hum. Touch a microphone input wire (This one,
Simon...) and you'll hear the hum.
We haven't got to the other end yet,
so let's look at that now:

Here the input is unbalanced, so
there is little choice but to connect the white wire of
an XLR wired cable (shown in Green) to ground leaving just
the red wire to carry the signal. This effectively makes
the balanced cable unbalanced. A single core screened cable
would be adequate here, the screen also becoming the Earth.
The
input of a theoretical Differential Amplifier (Figure
5) has a Positive and a Negative input. If a positive going
signal is applied to the positive input, the output goes
positive. If a negative going signal is applied to the negative
input the output also goes positive.
So if a positive signal is applied to
both inputs they will cancel each other out and the change
at the output terminal will be zero. We're nearly there.

Now (Figure 6) we have a balanced cable
connected to a balanced input. Because the signals on the
two wires are in opposite phase (see Figure 1) they both
contribute to the output.
Of course the two signal cores will still
pick up mains hum and other interference, but because the
hum will appear as two signals in phase, (I.e., they are
both going positive at the same time - Hey Presto, they
cancel each other out in the differential amplifier.
So you get the signal and reject the
interference. There are always non-linearities in electronics,
so perfect rejection of interference is not possible. The
ability to reject an interfering signal which is common
to both cores is known as the Common Mode Rejection and
is quoted in dBs.
If you only have a choice of an unbalanced
input, try to get a microphone giving a higher than average
output level, this will increase the signal-to-noise ratio.
Let me know if you find this useful and
interesting.
Copyright / Tony Wright / 2005-2010 |