Handbook for Sound Engineers

Transmission Techniques: Fiber Optics 461

graded-index fibers. A large NA promotes more modal

dispersion, since more paths for the rays are provided.

Sources and detectors also have an NA. The NA of

the source defines the angles of the exiting light. The

NA of the detector defines the angles of light that will

operate the detector. Especially for sources, it is impor-

tant to match the NA of the source to the NA of the

fiber so that all the light emitted by the source is

coupled into the fiber and propagated. Mismatches in

NA are sources of loss when light is coupled from a

lower NA to a higher one.

15.4.3.3 Attenuation Measurement

In an optical fiber, attenuation measurements require

comparison of input and output power Pin and Pout,

respectively. It is measured in decibels as

(15-15)

where,

the negative sign is added to give attenuation a positive

value because the output power is always less than the

input power for passive devices,

LFOP is the level of fiber optic power expressed in dB.

Remember these are optical powers, and they are
dependent on the wavelength. Optical power digital
meters make their measurements readings in either dB
or dBm, and also display the wavelength. The optical
power level LOP is computed with the equation

(15-16)

where,

Ps is the power of the signal,

Pr is the reference power.

If the reference power is 1 mW, then the equation for

the optical power level LOP becomes

(15-17)

Notice when we know the reference power is 1 mW

the unit of level changes to dBm. When the reference

power is not specified the unit of level is in dB.

Precise fiber attenuation measurements are based on

the cut-back method test shown in Fig. 15-16. Here a

light source is used to put a signal into the optical fiber;

a mode filter is used in graded index or multimode fiber

to establish a consistent launch condition to allow

consistency of measurements. Although modal condi-

tioning (using mode filter) is beyond the scope of this

discussion, it is a very important topic for making

measurements in multimode fiber because of the effects

of modal conditioning on the values one will measure in

this test. Measure the amount of light that comes out at

the far end, then cut the fiber back (about 1 to 2 meters)

to just past the mode filter. Measure the amount of light

that comes out the new end. The difference in the light

at one end and that at the other end divided by the

length of the fiber gives you the loss per unit length, or

the attenuation of the fiber. This is the method used by

all manufacturers for testing their fiber.

Be aware, however, that this will not accurately

measure the loss that light will experience in short

multimode fibers because that loss depends on propaga-

tion of high-order modes that are eliminated from

measurements by adding a mode filter.

Similar measurements can be made on fiber optic

cables with mounted connectors by replacing the short

cut-back fiber segment with a short jumper cable

(including a mode filter if desired). That approach

simplifies measurements by avoiding the need to cut

fibers at a modest sacrifice in accuracy. One special

problem with single-mode fibers is that light can propa-

gate short distances in the cladding, throwing off

measurement results by systematically underestimating

input coupling losses. To measure true single-mode

transmission and coupling, fiber lengths should be at

least 20 to 30 m (65 to 100 ft).

Figure 15-15. Numerical aperture (NA).

Light ray outside

acceptance cone

Cladding Core

Light ray lost in

cladding by

absorption

Acceptance

cone

Half

angle

NAsin Q

Q

n 1

n 2

n 2

LFOP 10–

Pout

Pin

= log©¹§·---------- in dB

LOP 10

Ps

Pr

= log©¹§·----- in dB

Figure 15-16. The cut-back method for fiber attenuation.

LOP 10

Ps

1 mW

= log©¹§·-------------in dB

Light

source

Mode

filter

Fiber to test Far end point

Cutback point

Power meter

• 3.2
  dBm
  or
  dB/km