HP-PN8504-1_Measurements of Lightwave Component Reflections

8/14/2019 HP-PN8504-1_Measurements of Lightwave Component Reflections

http://slidepdf.com/reader/full/hp-pn8504-1measurements-of-lightwave-component-reflections 1/20

Measurements of lightwavecomponent re flec tions wi th theHP 8504B precision refle ctometer

Pr oduct Note 8504-1



2

The precis ionref lectometer

A new d evelopment in optical

reflectom etry, th e HP 8504B pre-

cision reflectometer, significan tly

extend s m easurement capabilities.The tw o-event resolution is better

than 25 m icrom eters, while the

dyn am ic range exceeds 80 dB.

A m easurement tool now exists

specifically for the designer and

m anu facturer of precision light -

wave components and connectors.

The sensitivity exceeds th at of the

best power m eter solutions, wh ile

the resolution is sufficient to iso-

late each ind ividu al reflection

within a small, complex optical

assembly.

Table of Conten ts

HP 8504B operation su mmary 3

Return loss concepts and measurements 4

Basic concepts of reflect ion 4

Problems that resu lt from reflected ligh t 4

S urvey of return loss measurem ent m ethods: 5

Power meters 5

Opt ica l t ime-domain reflectometers 6

Opt ica l frequency-domain reflectometers 6

The precision reflectometer technique 6

The general measurement process 8

Inst rument warm-up 8

Reference extension cable select ion 8

Cleaning connectors 8

Select opera t ing wavelength 8Measurement calibrat ion: 9

Balance receiver 9

Magnitude ca libra t ion 9

Measure the test device 10

Opt imize the instrument setup 10

Measurement example:Connector pair 10

Measurement procedure 10

Increasing the measurement ra te 11

Measurement example:Char acterizing a photodiode assem bly 11

Measurement procedure 11

Opt imizing the inst rument setup 11

Applications 13

Measuring the reflect ions from an opt ical isola tor 13Character izing reflect ions within a laser assembly 13

Char acterizing devices pigtailed with mult imode fiber 14

Pr ecision measu rem ents of differential length. A 1XN coupler 15

Char acterizing high retu rn loss term inat ions: Index ma tching gel 16

Common quest ions and answers 17

Understanding measurement accuracy 18

Bibliography 19



3

HP 8504Boperat ion summary

The following one page operating

summa ry is intended as a brief ref-

erence. Detailed inform at ion is dis-

cussed within the pr oduct note.1. Warm-up the instrum ent: For

maximum dynamic range and mea-

sur ement s tability, allow the inst ru -

ment to warm up for two hours.

2. Clean al l conn ection s: Clean

all fiber connectors a nd instr um ent

test port s used in both t he calibra-

tion a nd m easurement process.

At th is point you can follow th e

operating procedure below, or use

th e instr umen t’s “Guided Setup”

featu re. Press SYSTEM, [Guided

Setup].

3. Se lect wave length: Press

PRESET, MENU , and select th e

appropriate wavelength (under

[SOURCE MENU]).

4. Determine reference exten-

s ion length: Measure the length

(L2) of fiber cable leading to th e

device under test (DUT). Attachextension cable L1 between t he r ef-

erence extension ports with length

equal to or slight ly less th an L2. If

the extension cable length L1 is

greater tha n L2, the device may

not be seen in the instr umen t’s

measurement range.

5. Pe rform a cal ibration:

This r emoves DC offsets an d polar-

ization s ensitivity, an d sets a cali-

brat ed r eference level. Select CAL,

[Guided Cal]: Termina te the instr u-

ment test port with the high return

loss load (>40 dB) supplied with th einstrum ent. Adjust the polarization

balance as directed.

Note: Once the polarization calibra-

tion has been performed, the reference

extension cable L1 and the polariza-

tion adjustment knobs must remain

stationary. Attach a fiber cable of

length L2 or slightly longer, with

known return loss (typically a Fresnel

reflection) to the instrum ent t est port.Measure the sta ndard a s directed.

If the system is operating correctly,

ther e should be a single response seen,

similar to th e display sh own below.

6. Connect the test de vice:

Attach th e device under test (DUT)

to the instr ument t est port. Let the

instr umen t complete a full sweep.

Locate the reflections of interest

and reduce the measurement spa n

as m uch as possible using the MKR

FCTN and span keys.

7. Increase sensitivity: Increased

sensitivity can be achieved thr ough

averaging. Pr ess AVG, [AVERAGIN G

ON]. The nu mber of averages is set

using t he [Averaging Factor] function.

The num ber of avera ged traces is

displayed at th e left border of th e

display.

L1 L2

DUT

27 cm (in fiber)measurement

range

0<L2–L1<25 cm

HP 8504B Measurement Setup

Typical Display

0 dB return loss

Reflectiondisplayed indB return loss

50 dBreferencelevel

Instrumentnoise floor



4

Return loss conceptsand meas urementtechniques

Th e reflection coefficient ‘ρ’ of this

interface is th e ra tio of the r eflected

electric field to the incident electric

field and is given by th e following:

ρ=n

1–n

2

n1+n

2

where n1

is the index of refraction

for th e mat erial the light is propa-

gating from and n 2 is th e index of refraction of the mat erial th e light

is tr aveling to.

The reflectance ‘R’ is a s imilar ter m

and is defined as the r at io of the

reflected beam inten sity to the inci-

dent beam int ensity and is given

by:

R= ρ2

= (n1–n

2)2

n1+n

2

(In the a bove cases, it is ass umed

that the light is tr aveling norma l

to the inter face plane.)

An exam ple of th is phenomenon is

an a ir-to-glass int erface. Air h as a n

index of refra ction of 1, while glass

has an index of approxima tely 1.5.

The r eflecta nce of this int erface is

then 0.04 or 4% while the reflection

coefficient is –0.2. A n egat ive re flec-

tion coefficient is an indication that

the reflected field has experienced a

180-degree phas e shift relat ive to

the incident field.

The reflective properties at an

interface can also be described loga-

rith mically in decibels. This par ame-

ter is called retu rn loss (RL) an d isgiven by:

RL= –10 log10 (power reflected)power incident

–10 log10

R

The air-glass int erface would ha ve

a r etur n loss of 14 dB.

Notice tha t th e smaller th e reflec-

tion, the lar ger the ret ur n loss. This

shows the ut ility of using ret ur n loss

to describe very small reflections.

Note: In measu ring reflections,

intu itively you would expect low

reflections to be displayed lower t han

high reflections. Consequently, the

HP 8504B uses a different conven-

tion in displaying retur n loss. Retur n

loss is computed a s:

10 log (power reflected)power incident

as opposed to:

–10 log (power incident )power reflected

Thus a retu rn loss of 20 dB is dis-

played as –20 dB and a retur n loss

of 50 dB is displa yed as –50 dB.

The m ajority of this documen t dea ls

with components and devices used

in systems tha t u se optical fiber.

Another phenomenon th at m ust be

consider ed is th e coupling of the

reflected light back into the fiber. In

general, th e light from a r eflective

inter face is not collima ted a nd only

a portion will ret ur n back thr ough

th e fiber. Thus, th e definition of

return loss must be modified again.

Return loss is then defined in terms

of what actua lly propagates ba ck thr ough the fiber rat her tha n what

would be predicted solely from the

differences in t he refractive indices

at a bounda ry.

Problems that resultfrom re f lecte d l ight

The m ost obvious pr oblem th at

occurs when light is reflected is t ha t

the t ran smitted signal is reduced.

In t he pr evious glass-to-air inter face

example, since 4% of the light was

reflected, only 96% of the or iginal

signal was transmitted. This corre-

sponds to a (–10 log10

(0.96)) 0.2 dB

loss in power.

Sometimes more important

th an power loss is the effect t ha t

reflected light ha s on the per for-

ma nce of lightwave component s.

Today’s h igh-speed lightwa ve com-

mu nication system s typically usenarrow linewidth lasers. The rela-

tive intensit y noise (RIN) and m od-

ulat ion char acteristics of such laser s

can be significantly degraded by

very sma ll amount s of backreflected

light. Reflected light can also cause

“bit errors” in digital communication

systems and distortion in analog

comm unication systems.

As light reflects off one int erfa ce,

the reverse traveling waveform may

be re-reflected off another inter face

(closer to the source). This results

in two forwar d tr aveling waves. Thema gnitude of the composite forwar d

tr aveling signa l is the vector sum of

these two waveform s. Depending on

the ph ase r elationship of the t wo

signals, which is in tur n dependent

on wavelength and t he path length

tha t th e reflected wave tra veled,

th e two waves may add eith er con-

str uctively or dest ru ctively. Su btle

changes in source wavelength a nd

environmenta l chan ges (such as

temperat ure) can cause this phase

relat ionship to var y significantly.

Thus, the total power seen at t he

detector will vary with time.

For these above mentioned reasons,

the components used in lightwa ve

systems such as isolators, connectors,

an d ph otodiodes t ypically have very

high return loss (low reflections)

an d as few reflections a s possible.

Bas ic conceptsof reflectio n

When light t ra vels across the

boundary between materials with

differen t in dices of refra ction or

densities, some portion of the light

will be reflected. The figure below

shows th e simplest form of this con-

cept, a plan e wave tra veling per-

pendicular to the boundar y between

the two materials.

Boundary

N1 N2

Incident

Reflected

Transmitted



5

Survey o f returnloss measureme nttechniques

Contemporary methods

Virtually all return loss measure-

ment techniques employ some t ype

of optical r eflectometr y. This consist s

of illumina ting th e device un der t est

and measur ing the light th at is

reflected back. Optical reflectometers

include power m eter/coupler based

systems (sometim es called optical

contin uous wa ve reflectometers or

OCWR), optical tim e-doma in r eflec-

tometers (OTDR), optical frequency-

domain r eflectometer s (OFDR), an d

th e HP 8504B precision reflectome-

ter. Each of the a bove ment ioned

measu remen t techniques offer

unique advanta ges and benefits.

Pow er me ter/couplerbased measurements

Return loss measurements can be

made using a power m eter such as

th e HP 8153A light wave multime-

ter a nd th e HP 81534A return loss

module. The ret urn loss module

conta ins a sen sitive detector a nd

a dir ectiona l coupler. The sour cemodule of the HP 8153A illumina tes

the test device, while the dir ectional

coupler a nd detector sense only the

power th at is tr aveling in the reverse

direction.

This system is best suited for accu-

rat e measu rement of devices with a

single reflection su ch as conn ectors,

splices and att enua tors. It is also

used to measur e the aggregate or

“tota l” ret ur n loss of a device with

mu ltiple r eflections.

Return loss when thereare multiple reflections

While the power meter system is

both economical and easy to use, itdoes not pr ovide any spat ial inform a-

tion for resolving mu ltiple reflections.

There ar e two important implica-

tions t o consider. Fir st, ident ifying

and quan tifying each r eflection is

importa nt in the design and ma nu-

factu ring of high ret ur n loss compo-

nent s. Second, th e total retur n loss

of a component can vary significan tly

when t wo reflections creat e a

Fabry-Perot resonator.

When m ultiple reflections exist, the

spectral char acteristics of th e light

source must be considered. If th e

light s ource has a very nar row line-

width, it will then ha ve a very long

coherence length. In brief, this

implies that the phase chara cteris-

tics of the light a re very st able.

When light reflects off two discrete

inter faces th ere will be two reverse

tr aveling waves. The total reverse

tr aveling power will be relat ed to the

vector su m of th e two waves. The

phas e relat ionsh ip of these t wo sig-

nals is dependent u pon the source

wavelength an d path length betweenthe two reflecting interfaces. The

relative phase between the two

waves will then det ermin e whether

th ey will add const ru ctively or

destru ctively. This is r elated t o the

Fabry-Perot effect. An example of

th e phenomenon can be demon-

str at ed with a simple connector

pair with an air gap of 1 mm.

Forwar d-traveling light will first

encounter the glass-air interface

at th e end of the first connector.

Approximately 3.5% of the light

will be reflected. Th e ma jorit y of

the light will continu e to th e air-

glass in ter face of the second connec-

tor. Again, t her e will be a 3.5%

reflection.

At several discrete wa velength s

(such as 1300 nm), the air “cavity”

length is such th at t he two reflected

waves will be precisely in pha se, andthe reflected power will be at a maxi-

mum . However, at other wavelengths

(such as 1300.4 nm), the t wo wave-

form s will be out of phas e, and t he

two signals will add destructively.

The r eflected power will be at a

minimum 1.

Return loss for two equal reflections1 mm air gap

R e t u r n L o s s ( d B )

Distance (mm)

1 mm

Return loss for two equal reflections1 mm air gap

N1N2

N1L

Er2Er1

ErTotalEr1

Er2 Er2

θ=?Er1

1 In t heory, if th ere were n o coupling losses, the

stimulus wa s monochromatic, and a ll of the

signals r e-reflected in t he cavity were consid-

ered in addition to th e primar y reflections, the

ret urn loss can go to infinity, implying a reso-

nant condition.



6

The HP 8504Bprecis ion ref lectometer

A new development in optical

reflectometry, the HP 8504B pre-

cision reflectometer, significan tly

extends measurement capabilities.

The t wo-event resolution is better

tha n 25 micrometers, while the

dyna mic ra nge exceeds 80 dB. A

measu remen t t ool now exists spe-

cifically for the designer and man-

ufacturer of precision lightwave

components and connectors. The

sensitivity exceeds th at of the best

power m eter solutions, while the

res olut ion is sufficient to isolat e

each individual reflection within a

sm all, complex optical as sem bly.

The HP 8504B precision reflectome-

ter is bas ed on th e Michelson inter-

ferometer an d uses t he techniques

of “white light” inter ferometr y. A

1300 nm or 1550 nm low-coherence

light source is sent t o a power split-

ter. One pat h leads t o the DUT, while

the other pa th leads to a reference

mirror. Light reflected off both the

reference mirror and th e DUT is

recombined an d detected. If the pa th

length from t he source to th e reflec-

tion in t he DUT is the sam e as fromth e source to the mirr or, a coheren t

interference signal appears a t t he

dete ctor. By moving th e position of

the reference mirror, the instrument

can t hen “scan” the t est d evice for

reflections over a 400 mm r ange

(equivalent a ir dista nce). (The

It is worth ment ioning again that

th is effect will only occur when a

highly coherent source is used. This

has some important implications.

The ret ur n loss of a component can

vary significan tly depending upon

the stimulus. Return loss testing of

a component m ay not accurat ely pre-

dict its behavior in a working system

if the source char acteristics ar e not

similar in both situa tions. In addi-

tion, when component s ar e used in

a system with a h igh coherence laser

(such a s a DFB or N d.YAG), the com-

ponent retu rn loss char acteristics

ma y chan ge drama tically as wave-

length an d temperatur e are even

slight ly varied.

To more completely under sta nd

the reflection chara cteristics of

components a nd su bassemblies,

the individual r eflections mus t be

isolated an d char acterized. Othermeasur ement techniques are used

to both locat e an d qua nt ify individ-

ua l r eflections.

The OTDR

Optical time-doma in r eflectometr y

is the most familiar a nd comm only

used reflectometry measurementfor the installation and maintenan ce

of both long- an d short -hau l fiber

links. OTDRs locate faults by prob-

ing a fiber with a n optical pulse tr ain

and measuring the reflected and

backscat ter ed light . OTDR’s are typi-

cally not used for component -level

measurem ents due t o limitations

in r esolving sm all or closely spaced

reflections, unless very short

impulses an d very high-speed,

sensitive detectors ar e employed.

The OFDR

The two-event resolution and dea d-

zone problems inher ent with OTDR’s

can be impr oved by using a s wept

modulated lightwave instead of

pulses of light . The amplitu de and

envelope pha se r esponse is r ecorded

over a wide frequency span . An

inverse Four ier tra nsform is per-

form ed on t his dat a t o yield th e time-

domain r esponse. Depending upon

the m odulation bandwidth of th e

instr umen t, the t wo-event r esolution

can be better t han 10 mm (20 GHzbandwidth). The r eceiver is syn-

chronously tun ed to th e source,

thu s decreasing the susceptibility to

noise and increas ing the dynamic

ra nge to levels near 40 dB. (See

product specific litera tur e for t he

HP 8702 and H P 8703).

Precision lightwave reflectometerblock diagram

Detector Display

DUT

1300

1550

WDM

Coupler

Reference Mirror

ReferenceExtension

Return loss, two equal reflections vs. Wavelength

R e t u r n L o s s ( d B )

Wavelength

∆

Lambda0.5 mm

1300

A plot of retur n loss versus wa ve-

length sh ows th is result graph ically.

This effect is repetitive as a function

of wavelength so tha t m aximumsand m inimum s will occur a t several

wavelengths.



7

400 mm m easurement window can

be offset by simply adding length to

the reference path at the instrum ent

front pan el).Note that in this measurement

scheme th ere ar e no pulses of light

tra nsmitted as there are for the

OTDR, an d th e light is n ot modu-

lated as it is for th e OFDR. It is the

short coheren ce length of th e LED

source in the precision reflectome-

ter which leads to very high r esolu-

tion measur ements.

To demonst ra te t he u tility of such

an inst ru ment , consider a high-speed

photodiode.

The ph ysical dimen sions of

th e device are mu ch less tha n a

centimeter, yet there are several

inter faces within th e device, each

potentially generat ing a reflection

an d contribut ing to the overall

ret ur n loss of th e device. Very h igh

resolution is requir ed to locate,

identify and qu ant ify each r eflec-

tion. The precision reflectometer

easily performs th is task, as seen

in the following measu rem ent.

Lens

Diode chip

Heat sink

In this measur ement, return loss is

displayed versu s th e one-way path

distance. The measurement span

is 6 mm so the horizontal scale is

0.6 mm per division. Retur n loss is

displayed in decibels. The top of th e

display is 0 dB retur n loss, the cen-

ter of th e display is –50 dB. Higher

retu rn loss levels are displayed as

a lower value on th e display since

they correspond t o lower va lues of

reflection. (Intuitively, you would

expect low reflections t o be displayed

at the bottom of the screen). Thus a

51.9 dB ret ur n loss value is dis-

played as –51.9 dB as noted by the

measurem ent m arker, which corre-

sponds to the r eflection a t t he front

face of th e diode chip.

The first r eflection is t he end of

the fiber connector. It appr oaches

a F resn el reflection since the fiber

does not physically contact the

device. The r eflections off the front

and ba ck of the lens, and th e front

and back of the diode chip are clear ly

seen. From this m easurement, you

see tha t th e retu rn loss of this device

is dominated by t he front face of the

diode chip. However, a power met er

measurem ent would be dominated

by the r eflection off th e end of the

fiber, making it difficult to extractany inform at ion about t he photodi-

ode itself.

It is also import ant to note the high

dynamic ran ge that t he interferom-

eter technique offers, several orders

of magnitude beyond th e capabilitiesof tr aditional meth ods.

The measurement ran ge of the

precision reflectometer is determined

by the length that the r eference mir-

ror can t ra vel and is 400 mm (equiv-

alent a ir distance). If measurement s

beyond t he 400 mm window are

required, the window is simply off-

set with t he appropriate reference

extension pat ch cord.

Knowing the ma gnitude and spa cing

of th e reflections yields informat ion

th at is useful in determining com-

ponent performan ce in systems with

narrow linewidth lasers. Specifically,

how much th e retur n loss may vary

for a given chan ge in wavelength.

In sum mar y, there ar e several tech-

niques for measuring r eturn loss.

The easiest m ethod is to use a power

meter. When reflections need t o be

spat ially resolved, the OTDR is used

for coar se mea sur ements of long

lines. The OFDR can be configured

for long-line m easur ement s or close-

in measur ements, depending on t he

modulation bandwidth u sed. Forthe highest resolution and dynamic

ra nge, the H P 8504B precision

reflectometer technique is optimum.



Generalmeasureme nt processw ith the HP 8504B

8

measur ements. This will ensure

tha t var ious DC offsets within t he

instrum ent ha ve stabilized and can

be effectively removed from subs e-quent measurements.

Once the ins t ru ment h as been

warm ed up, the “Guided Setup”

featur e can be used.

Clean f iber ends a ndinstrumen t test ports

It is a good meas ur ement pr actice

to clean fiber int erfaces before per-

forming calibra tions and m aking

measu remen ts. Clean fiber ends

and connector ports a re essential

for good mea sur ement s. All thr eeinstrum ent ports and a ny fiber ends

(including the reference extension)

should be kept clean. Dirt y connec-

tors can result in spur ious responses

and reduced dynam ic range. To clean

the instrument ports, the connector

adapters are removed from the

instrument front panel, exposing

the cable ferrule.

Caution: Ex treme care shou ld

be taken to avoid damaging the

instrument con nector ferrules .

Dama ged connectors r educe mea-surement integrity and ar e not user

serviceable.

Please r efer t o the connector care

docum ent in the m anu al for connec-

tor cleaning an d care.

Select operat ingwavelength

The sta ndar d configurat ion for th e

HP 8504B includes both 1300 nm an d

1550 nm measurement capability.

When not using Guided Setup, to

select between 1300 or 1550 nm oper-ation, press MENU , [Sour ce Menu],

and either [1300] or [1550].

Select referenceextens ion

The first step in making a m easure-

ment is to select t he appr opriate

length of reference extens ion cable.

In gener al, the r eference extension

cable length s hould be equal to

th e “pigtail” or path length t o the

device to be tested. The length of

the r eference extension m ay be

slightly shorter than the pigtail,

but should not be longer. If the ref-

erence extension is slightly longer

th an the DUT cable, some of the

events to be exam ined may not be

in the 400 mm (air) measurement

span of the instrum ent.

The HP 8504B is designed so

tha t if the r eference extension

length is identical to that of the

DUT fiber, the first response will

appear appr oximat ely one division

to th e right of th e left edge of th e

display (in a 400 mm span). This

means t hat under n ominal condi-

tions, the reference extension cable

should be less tha n 10 millimeter s

equivalent air length (7 mm actual

fiber length) longer th an t he DUT

fiber. Due to the 400 mm allowable

measur ement span , the extensionshould be no more t ha n 360 milli-

meter s (250 mm actua l fiber length)

shorter t han t he DUT fiber. In math -

emat ical term s, this is given by:

–7 mm <L2–L1<250 mm

where: L2 is the DUT fiber length

an d L1 is the r eference extension

length.

The ideal condition is to have t he

two cables be of equal length.

The H P 8504B option 001 cont ains

refer ence exten sion cables of 40, 50,

75, 100, 125, 150, and 175 cm lengths.Optimum reference extension length

depends upon th e path length to

th e DUT.

If the DUT ha s no pigta il, L1 and

L2 are any two cables of equal length.

A condensed summ ar y of the mea-

surement procedure is found in the

front of th is document. In operating

the instrum ent there ar e “har dkeys”an d “softkeys”. Har dkeys ar e th ose

keys whose function is print ed dir-

ectly on t he ph ysical keypad. Th ese

keys ar e noted with bold type such

as PRESET. Softkeys are those tha t

ar e located to the right of the inst ru -

ment display and whose function is

displayed on th e instru ment display.

They are noted in brackets such as

[MKR ZOOM].

There ar e seven steps to perform ing

a measurement.

1 Warm up the instrument

2 Clean fiber ends andinstrument test ports

3 Select operating wavelength

4 Select reference extension

5 Perform a calibration

6 Connect the DUT

7 Optimize the measurement

The HP 8504B ha s a “Guided

Setu p” featur e. Guided Setup is a

powerful user inter face tha t leads

users th rough th e steps required to

make measu rements. Guided Setup

is implemented by pressing SYSTEM,

[Guided Setup]. The inst ru ment willthen display each step required to

setup and calibrate the instrum ent.

The following text describes the

steps u sed in Guided Setup, as well

as some discussion on wh y each

step is perform ed.

Warm-upthe ins trumen t

The HP 8504B is capable of making

measur ements of extremely small

reflections. Even very sma ll spuri-

ous signals within t he instru mentcan degrade dyna mic range. To

ensure ma ximum dynam ic ran ge

and measurement stability, turn on

the H P 8504B and a llow it to warm

up for two hours pr ior to ma king



9

Option 001 Cable connection guide

DUT path Ref ext

L2 (cm ) cable s L140<L2<65 40

50<L2<75 50

75<L2<100 75

90<L2<115 40+50

100<L2<125 100

115<L2<140 40+75

125<L2<150 125

140<L2<165 40+100

150<L2<175 150

165<L2<190 40+125

175<L2<200 175

190<L2<215 40+150

200<L2<225 50+150

215<L2<240 40+175225<L2<250 50+175

240<L2<265 40+50+150

250<L2<275 75+175

265<L2<290 40+50+175

275<L2<300 100+175

290<L2<315 40+75+175

300<L2<325 175+125

315<L2<340 40+175+100

325<L2<350 175+150

340<L2<365 40+175+125

350<L2<375 75+100+175

365<L2<390 40+150+175

375<L2<400 50+150+175

Measurementcal ibrat ion

Once the reference extension h as

been selected an d att ached to the

instrument, the instrument must

go thr ough a measurem ent calibra-

tion. The measurement calibration

consists of balancing the polarization-

diversity receiver an d calibra ting

th e instr umen t for a ccur ate r eflec-

tion ma gnitude measurements.

Balance receiver

As light t ra vels th rough single mode

fiber, its polarization characteristics

vary. The m agnitu de of the detector

response is potent ially a function

of the polarization of the reflected

waveform r elative to the light in

the reference arm . Ideally, the detec-

tor r esponse is only a function of

th e reflection ma gnitude. To ensure

that the r eflection m easurement is

insensitive to polarization tr ans for-

mation, a m easurement calibrat ion

process is used. The instr um entreceiver consist s of two photodiodes

which r espond to orthogonal st at es

of polarization. During the Balance

Receiver calibration, the instrument

test port is terminat ed with a high

retu rn loss optical load (great er th an

40 dB) which is supplied with the

instrument. Therefore the light t hat

hits t he two detector diodes is only

from th e r eference mirror. The polar-

ization of this light is adjus ted with

the polarization adjustment knobs

at the instrument front panel in

such a way t hat the r esponses fromeach detector a re equal or balan ced.

Note: Polarization of the light in the

reference path must not be a l tered

once the Balance Receiver calibration

has been perform ed. T he reference

extension cable and polarization

adjustment k nobs mu s t no t be

moved to ensu re optim um perfor-

m ance. If th e reference extension is

m oved, t he receiver will no longer be

balanced an d su bsequent m easure-

m ents m ay be in error.

To perform the Receiver Balance

step, simply follow th e inst ru ctionsgiven by the instrum ent.

Magnitude cal ibration

The magnit ude calibrat ion is a sim-

ple process consisting of measur ing

a kn own reflection. The instr umen t

then automatically scales th e mea-

sur ed response so the t ru e value is

displayed.

For every reference extension cable

supplied in th e HP 8504B option 001,

there is a corr esponding fiber of equal

length which may be used as a cali-bration standar d. The return loss of

th e fiber end (a super P C ferr ule) is

15 dB or 3.16% reflection a t 1300 nm

an d 14.7 dB or 3.37% at 1550 nm .

As you continue with the “Guided

Setup” procedure, the instrum ent

will display the inst ru ctions t o per-

form the m agnitu de calibrat ion. The

HP 8504B scans t hrough the entire

measurement span and determines

the value of the peak response, which

should be the r eflection generat ed

at th e end of the cable. If you u se a

reflection st andar d other tha n aFr esnel reflection, it mu st ha ve a

retur n loss greater tha n 14 dB to

prevent sa tur at ion of th e receiver.

The HP 8504B compares t he mea-

sur ed value to the value ent ered by

th e user. Any differences ar e due to

systematic errors in th e measure-

ment system. This error term is sub-

sequent ly removed from all furth er

measurement s unt il another calibra-

tion is performed or t he instr ument

is preset to the default settings.

As a check, the calibrat ed measu re-ment of the r eflection sta nda rd can

be compared t o the known value.

It is importa nt to note that th ere

is no length calibra tion process an d

consequent ly the H P 8504B does

not make a bsolute length m easure-

ment s directly. It is a comm on mis-

conception t hat the m easurement

calibration process will offset the

position of the reflection sta ndar d

to the 0 length position. Recall tha t

when t he length of the DUT fiber

is identical to that of the r eference

extension, the first event appears

at a bout t he 40 mm point a nd not

0 mm. The instrument has n o know-

ledge of the length of the reference

extension cable used, nor th e length

of fiber t o the device un der t est.

Therefore, all distan ce accur acy is

in ter ms of relative distance to other

reflections in the measurement span.



10

Connect and measurethe tes t device

The calibrated instrum ent can n ow

be used t o measur e devices. Once

th e device is connected, it mu st be

locat ed on th e screen . If a device is

connected while th e instru ment is

in a m easurement sweep, a response

ma y be genera ted in th e process of

ma king the connection tha t is not

related t o the a ctual device response.

Consequen tly, it is a good measu re-

ment pr actice to restart the mea-

sur ement as soon as t he device is

atta ched. Press MENU , [Full span],

MEAS , [Meas restar t]. The instru -

ment will then scan over its full

measur ement ran ge and display

all detectable r eflections within a

400 mm (air) span .

Optimizing th emeas urement

For measurement s at 1300 nm, the

reference mirror in the HP 8504B

tr avels at a constan t velocity of

18 mm /sec (21 mm/sec at 1550 nm ).

A 0 to 400 mm sweep will tak e over

22 seconds . The sweep-to-sweep time

can only be reduced by reducing th e

measurement span. Reducing the

measurement span will also increase

the spatia l resolution of multiple

reflections. It is recomm ended th at

once th e DUT responses h ave been

located, to reduce the mea sur ement

span to the narr owest range that

includes th e events of inter est.

Nar rowing of the s pan can be

achieved using th e MKR FCTN

men u a nd [MKR ZOOM] key, or by

using the SPAN , CENTER, STOP ,

and START keys.

Once the measurement span ha s

been optimized to include a ll events

of int erest , the n oise floor can be

reduced through data averaging.

Avera ging redu ces the effects of

ra ndom noise, and can increase th e

measur ement ra nge by 6 or 7 dB.

(Note: the nar rower th e measur e-

ment span , the more effective aver-

aging becomes). Press AVG and

[Avg on]. The default aver aging fac-

tor is 16, meaning that each new

meas urem ent is weighted by a fac-

tor of 1/16 and will contribute this

value to the current m easurement

tr ace. The averaging factor can beset to any integer value between 2

and 999.

Measurement example:connector pair

In t his measur ement example, you

will measur e the r etur n loss of a

simple connector pa ir. In a ddition,

you will see the connector charac-

teristics as it m akes the t ran sition

to a fully torqued connection.

Measurement procedureUsing the same general measure-

ment process described above, select

the reference extension cable. The

device to be measur ed is a simple

conn ector pa ir. One of th e conn ec-

tors is at the en d of a 75 cm pat ch

cord. Since the pat h length t o the

DUT is simply the length of this

pat ch cord, the ideal length for t he

referen ce exten sion is 75 cm. The

reference extension could be as short

as 51 cm, which would place the

reflection at th e end of the m easur e-

ment r ange, but sh ould not be anylonger th an 75 cm.

To calibra te, follow th e Guided S etup

or press CA L, [GUIDED CAL],

atta ch the supplied high retur n loss

terminat ion at the test port, adjust

the polarization adjustment knobs

for a ba lanced display as inst ru cted

by the instrum ent an d press [DONE].

For the magnitude calibration, the

calibrat ion sta nda rd can be one of

the 75 cm cables supplied with th e

instrum ent, or the DUT patchcord

connector end (if the r etur n loss isknown). Connect t he calibrat ion

standa rd to the test port . Pr ess

[FRESNE L 3.16%] when u sing the

supplied cable or [USER STD] and

enter th e value for t he reflection.

Pr ess [MEASURE]. The ana lyzer

will then measure the standar d and

adjust its m easur ement t o coincide

with th e actua l reflection value.The pat ch cord to be tested can

now be connected to the t est port .

Once it is connected, press MENU ,

[FULL SPAN]. The instr um ent will

then bring the reference mirror to

its sta rting position an d begin a

new measur ement t race. The posi-

tion of the mirr or is indicat ed by a

sma ll red dot a t th e bottom of the

instr umen t display. The location of

the dot on th e instru ment display is

proportional to the location of the

mirror on the tran slation st age. For

insta nce if the st ar t position of th emeasurement is set at 0 mm and

th e stop position is set at 100 mm,

the mirror will then be traveling

back and fort h over th e first 25%

of the a vailable mirr or movement .

The m irror position indicator dot

will th en m ove back an d fort h from

th e left edge of th e screen t o a point

25% or 2.5 divisions away from the

left screen edge. The actual dat a

meas ur ed will be displayed across

the ent ire screen.

As the mirr or tra vels across its full400 mm ra nge, the r eflection from

the end of the 75 cm pat ch cord

should be seen at appr oximately

th e 40 mm point. Becaus e the con-

nector is not ter minated, a retu rn

loss of about –15 dB should be seen.



12

In this measurement, there are

several r eflections, some which ar e

very small. Through averaging, the

measurement sensitivity can be

increased. Press AVG, [Averaging

On], and let the instru ment ta ke

several sweeps. This allows us to

see reflections th at m ay not have

been visible before. In this mea-

sur ement , a fifth reflection is now

clearly visible.

Measureme nt exam ple: Highreturn loss air-gap co nnec tor

An example of this is a very high

return loss beveled-edge connector.The connector r esponse is easily seen.

For t his pa rt icular connector, you

not only measure th e retur n loss,

by decreasing the measurement

span t o 1 mm, you measu re th e gap

between th e connector ferru les.

Press MKR FCTN , and use the

knob to move the ma rker slightly

to the left of the first event on the

screen. Press [Mkr–> Star t]. Thenuse the kn ob to move the mar ker to

the r ight of the last event on the

display. Press [Mkr–> Stop].



13

Applicat ions

The following sections demon-

str at e how th e HP 8504B precision

reflectometer can be used t o mea-

sure a variety of lightwave compo-nents. Although the measu rement

examples are just a small sample

of potential measurements, they

do provide a diverse survey of the

instrum ents capabilities and t he

types of measurements t hat can

be made.

In virt ua lly all cases, th e processes

for performing t he m easurements

are t he sam e as t hose described in

the pr evious sections for measu ring

the ph otodiode assembly and t he

connector pair. When m easur ement

techniques n ot previously intr o-duced are en count ered, these will

be described in appropriat e detail.

Measu ring theref lect ions froman op t ical i solator

Isolators are us ed to minimize the

impact of reflected light on other

components, in particular narrow

linewidth high-speed lasers. Not

only mus t th e isolator perform the

function of att enua ting backreflec-

ted light, it mu st do so without gen-

era ting r eflections of its own.

In gen era l, isolators consist of a

variety of components including

lenses, crysta ls and Fa ra day ele-

ments. Each individual component

ma y genera te a r eflection.

The procedure to measur e the

isolator is similar to tha t u sed for

th e photodiode in t he pr evious s ec-

tion under “Measurement example:

Char acterizing a photodiode assem -

bly.” Once the instr um ent ha s been

configured an d calibrated, t he isolatoris connected an d t he display opti-

mized.

This device ha s several int erfaces,

so there a re severa l reflections. The

measurem ent span is 30 mm. The

largest reflection, generat ed from

the angled fiber end, is approxi-

mat ely 60 dB. Beyond th e Far aday

element, th ere ar e no reflections,

indicatin g that t he isolator is per-

form ing its int ended function.

Characterizingref lect ions w ithina laser assembly

As ment ioned earlier, lightwa ve

source perform an ce can be degradedby backreflected light. Measuring

th e reflections by probing back into

the laser can give us insight into

how light r eflects inter na lly as well

as how back-reflected light ma y be

re-reflected from th e laser.

The d evice sh own below will first

be measur ed according to the pre-

vious procedures for measu ring

components.

This par ticular laser m odule con-

sist s of a pr otective window, a ball

lens, and th e laser chip. During

norma l opera tion of th e laser, lightpropagates from the chip and is

focused and a ligned th rough th e

ball lens before leaving th rough th e

window into a n att ached fiber. Light

tr aveling back into th e laser will

follow a similar pat h.

The resulting measurement shows

the r eflections generated a t each

inter face. The lar gest r eflection is

th e fiber end. Mar ker 1 shows the

reflection from th e front of the win-

dow, 2, the back of the window, 3,

th e front of th e ball lens, an d 4, the

back of the laser chip.

The response between m arker 3

an d 4 indicat es where th e back of

th e ball lens an d front of the laser

chip meet . Zooming in, we will

examine th is region.

LensLaser chip

Window



14

Here we see that t here are actu-

ally two reflections since th ere is

a sma ll gap between the lens and

the laser chip. The size of the gapis 43 microns.

Par t of the tran smission pat h of this

device is the semiconductor m ate-

rial of the las er. This ma terial is

more transparent at 1550 nm than

at 1300 nm. By making the mea-

suremen t at 1550 nm (this requires

a new measurement calibration),

the r eflected energy from the end

of th e laser chip experiences less

at tenu at ion an d is then effectively

larger.

Typically, the transparency of

the semiconductor ma terial willvary with t he bias curr ent th rough

th e laser. However, if the laser is

biased above threshold, the trans-

mitt ed energy ma y be sufficient

enough t o satur ate t he r eceiver of

the HP 8504B. Therefore, care

must be taken in setting the laser

curren t t o optimize the t ra deoffs

between material tr ansparency and

instrum ent sat ura tion level.

Characterizingdevices pigtai ledwith mu lt imode f iber

The H P 8504B pr ecision reflectome-

ter measur ement system detects

reflections when the pa th length

to the reflection is identical to the

pat h length to the reference mirror.

Because the light traveling through

mu ltimode fiber will propagate over

man y different pat hs, the measur e-

ment of reflected ener gy thr ough

mu ltimode fiber will be different

tha n if the device were pigtailed

with single mode fiber.

In essence, the r eflection responses

tend to be broadened. An exam pleof this would be to measure t he

Fr esnel reflection at the en d of a

1.25m length of 62.5/125 multi-

mode fiber compared to a similar

measu remen t with 9/125 single

mode fiber.

Two things are appar ent in th e

above plot. Fir st, th e response is

significant ly broadened. The 3 dB

width is about 75 microns, compared

to less than 15 microns for the sin-

gle mode case. Second, the amplitude

of the reflection is reduced. This is

due t o two factors. The t otal energy

of the reflection is distributed over

a wider ran ge, which decrea ses the

peak a mplitude. In a ddition, a sig-

nificant a mount of th e reflected

energy is lost at th e multimode to

single mode core mismat ch at th e

instrum ent test port.

The above meas ur ement is for a

specific length of mu ltimode fiber.

As the length of fiber increas es, so

does the effective pulse sprea ding.In addition, for multimode fiber

lengths m uch beyond one met er,

th e mult imode effect will genera te

multiple r esponses for a single

event.

It is interest ing to examine the

impact that this spreading has on

measur ements. Consider the photo-

diode measu rem ent int roduced on

page 11.

The a bove plot is a composite

meas ur ement of the same device.

One measur ement is with single

mode fiber, while th e other is witha 1.25m length of multimode fiber.

Each reflection is still visible, and

the relative magnitude informa tion

is still valid.

Although t he measur ement capa-

bility of the HP 8504B degrades

when u sing multim ode fiber, the

dynamic range and two-event reso-

lution provide very useful inform a-

tion in locating and identifying

reflections.



15

The function of the coupler is to

divide the inpu t power into two

pat hs. In s ome applications, it is

desirable to have the two path

lengths a s closely mat ched as pos-

sible. By monitoring the positions

of the reflections at the coupler out-

put s, we can determ ine the differ-

ential path length.

The measurement is straightfor-

ward. After set up an d calibrat ion,

th e coupler is connected t o the sys-tem. The outpu t ports of the coupler

are left unterm inated, which gener-

at es two large r eflections.

Determining the differential

path length is a simple matt er of

placing a m ar ker on ea ch reflection.

Press MKR FCTN , [Max Search],

[Mkr–>F ixed Mkr], [MKR 2],

MKR FCTN , [Peak search], [Next

highest pea k]. The analyzer will

th en display the one-way distan ce

between mark er 1 and mark er 2.

The measur ed distance value is the

equivalent distan ce tr aveled in air,

which in this case is t he different ial

path length to the t wo output ports.

To determine t he physical path

length differen ce, th e index of refra c-

tion value mu st be entered into the

instr umen t. In the case of glass

fiber, th e index of refraction is 1.46.

Press MEAS , [N VALUE], and ent er

1.46 x1. The displayed distan ce is

now the tr ue one-way physical

length difference.

In order to identify each pat h, we

simply termin ate one of the ports .

As expected, one of the m easur ed

reflections will decrea se in ma gni-tude th us indicat ing which response

coincides with the ter mina ted port.

This pr ocedure can be u sed for

virtua lly any 1xN coupler. The mea-

surement limitations are sensitivity

and two-event r esolution. As t he

power is divided into more and

more paths, th e magnitude of the

reflected ener gy from each outpu t

port will decreas e. However, the

sensitivity of the H P 8504B is such

tha t t he power could be divided into

over 1000 ports and the reflections

still be detectable.

Thus t he HP 8504B is sensitive

enough to measu re virtually any

1xN coupler. Obviously, th e pr acti-

cal measurement limitation then

becomes th e t wo-event resolution.

As the differential length of two

paths get smaller a nd sma ller, the

reflected signa ls displayed by the

instr umen t will eventua lly over-

lap. The two-event r esolution is

25 microns at 1300 nm an d can be

65 microns a t 1550 nm. Ther efore,

differential path lengths as sm allas 25 m icrons (air dista nce) can be

measured.

Precis ion measurementsof di f feren t ial length:1xN coupler

The HP 8504B is capable of making

relative length measu rements with

very high r esolution. This capability

can be used to measu re relat ive dif-

ferences in pat h length. An exam-

ple of this would be to measu re a

simp le 1x2 coupler.

Out 1

Out 2

In



16

The differential distance measure-

ments can also be interpreted in a

“time” mode format . This is u sed t o

indicate t he differentia l time of flightor propagation time between events.

Time format can be activated by

pressing FORMAT , [TIME].

Using the t ime form at, we can now

determ ine the effective delay that

one signa l path experiences relat ive

to others. With 25 micron two-event

resolution, pr opagat ion delays as

sma ll as 80 femtoseconds can be

characterized.

Characterizing highreturn loss terminations:index matching gel

Index m atching gel is comm only

used to term inat e an optical fiber

in an a ttem pt to minimize reflected

light genera ted at t he fiber end.

Once the gel has been applied to

the fiber end, the comm on assum p-

tion is tha t virtu ally all of the light

is scattered or “absorbed” and essen-

tially no light is reflected back.

The high sensitivity of the HP 8504B

allows us t o see the r eal perform ance

of ma tching gel. The r esults can be

surprising.

The measur ement is quite simple.The Fr esnel reflection at t he end of

a cable is located and t he instr ument

span is r educed to 1 mm. Then th e

end of th e fiber is dipped in th e

matching gel. Typically, the result

is tha t th e single reflection is reduced

to a retur n loss value better than

50 dB.

However, there ar e cases when t wo

reflections exist, one at th e glass/gel

inter face and a second r eflection at

th e gel/air inter face.

In m ost applications, the gel termi-na tion provides an adequa te fiber

ter mina tion. However, it is not sa fe

to assum e tha t it completely elimi-

na tes ba ckreflections.



17

Common ques t ionsand ans wers

The following section deals wit h a

variety of commonly asked questions

about the HP 8504B precision reflect-

ometer a nd th e techniques used tomake m easurements.

Q. Can th e HP 8504B detect fiber

backscatter?

The coherence length of the

HP 8504B sour ce is appr oximat ely

10 to 15 um . The ener gy reflected

from the DUT must h ave been gen-

erated within t his length in order

to be detected. Alth ough ther e will

be backscatt er genera ted over th is

length, the am ount of reflected

energy is t oo sma ll to be detected

by the HP 8504B.Q. Can the HP 8504B measure

fusion sp lices?

The measu rement sensitivity of

HP 8504B allows reflections with

retu rn loss values up t o 80 dB to be

measu red. In general, this is not

sufficient to det ect good fusion

splices.

Q. Th e HP 8504B d isplays 401

m easurem ent points in any m ea-

surement span. Is it possible for an

event to be between th ese points an d

therefore not be detected?The HP 8504B is designed to mak e

a measurement every 2.5 microns

along the measu rement pa th, inde-

pendent of the measurement spa n.

The coheren ce length of the inter-

nal source is 10 to 15 microns. Thus

several points will be measu red

within t his coheren ce length. Con-

sequent ly a r eflection will always

be detected, even if it falls between

th e actua l points of detection.

For wide measurement spa ns,

there are still only 401 data points

displayed, yet measurements arema de every 2.5 microns. For a 10 cm

measur ement span, th ere will be

40,000 measurements made. A

point will be displayed every

250 microns (401 point s over a

10 cm s pan). Thus each displayed

point will represen t th e largestresponse detected over 100 actua l

measurements. This also indicates

why the h ighest two-event resolu-

tion is achieved in th e nar rowest

measurement spans.

Q. Can th e HP 8504B be used to

measure devices pigtailed with

m ultim ode fiber?

The HP 8504B can be us ed to make

measurements in multimode fiber.

However, amplitude accur acy and

spat ial resolution will be degraded.

See “Characterizing devices pig-

tailed with multimode fiber” on

page 14.

Q. What happens wh en devices

with polarization maint aining fiber

(PMF) are measured?

Depending upon how light is

launched int o PMF, it will tr avel at

different velocities. Light traveling

on one principal axis will travel at

a different r ate t han light tr aveling

on the other principal axis. The light

emitt ed from th e HP 8504B sour ce

is only partially polarized, thu s

light will typically be tra veling onboth axes. Thus for a single reflection,

two extra responses may appear.

Q. Will the moving m echanical

components wear out?

The gears a nd motors in t he

HP 8504B are all designed to work

with much heavier loads an d stresses

tha n what is actually required. How-

ever, to avoid un necessa ry wear,

the HP 8504B will aut omatically

go into t rigger hold mode and stop

sweeping whenever t he instr ument

ma kes 500 sweeps (or th e specifiednum ber of averages, whichever is

greater) without the operator press-

ing a front pa nel key. Simply press

MEAS , [Contin uous], to resum e

sweeping.

Q. Wh y are the interferom eter

fringe patterns n ot seen?

An envelope detector is us ed to

smooth out the fringe pattern s, soonly one r esponse with in th e source

coheren ce length is displayed.

Q. Wh at is t he difference between

m easurem ents made with a power

m eter versus the HP 8504B and can

I determine total return loss using

the HP 8504B?

Total retur n loss and spa tially

resolved retu rn loss were discussed

in “Sur vey of ret ur n loss measur e-

men t t echniques” on pa ge 5. Total

retu rn loss is a function of the source

chara cteristics as well as t he r eflec-

tion m agnitudes and their spacing.

The 8504A can yield insight into

the component s of total ret ur n loss,

but n ot give an actua l value.

Q. What effects do connectors or

other losses ha ve when placed in

front of the DUT ?

Losses will reduce the m easu re-

ment sensitivity of the HP 8504. If

these losses are n ot included in the

ma gnitude calibration, they will

also decrease the m easured value

of reflections.

Q. What h appens when I m easure a

narrowband device such as a W DM

filter?

The high spat ial resolution of the

HP 8504B is based upon the wide

spectral width and subsequent

short coher ence length of its LED

sour ce. If a t est d evice’s res ponse

varies versus wavelength over th e

spectrum of the source, amplitude

accur acy and spat ial resolut ion

ma y be degraded.



18

Unders tandingMeasurement Accuracy

There ar e severa l element s which

affect t he a ccura cy of any m easur e-

ment m ade with the HP 8504B

Pr ecision Reflectometer. In general,various contribut ions to measu re-

ment error include systematic

errors tha t are inherent in the

Pr ecision Reflectometer measu re-

ment technique and i tems tha t

degrade measurement accuracy,

yet can be minimized by the user

thr ough good m easurement tech-

niques.

Measurement errors can also be

classified into those that affect

amplitude measur ement a ccuracy

an d those th at affect positional or

spat ial a ccur acy.

Amplitudeaccuracy i s sues

The uncertainty in th e HP 8504B

measur ement of retur n loss magni-

tu de comes from s everal different

factors including:

• Dynamic accuracy

• Connector repeatability

• Transmission path losses

• Measurement flatness vs. position

• Sweep-to-sweep repeatability

• Polarization sensitivity

• Calibration accuracy

• Chromatic dispersion effects

• Source spectral width

Dynamic accuracy

Dynamic accur acy refers to the

ability of the HP 8504B to accu-

rat ely measur e over a wide ran ge

of retu rn loss values. The HP 8504B

is typically calibrat ed at the level

of a F resn el reflection. Ideally the

response of the r eceiver block in th e

HP 8504B will be linear versus returnloss. However, as the reflected sig-

nals from t he DUT become very

sma ll, th e response from t he detec-

tor can deviate from th is ideal lin-

ear relat ionsh ip. Optimizat ion of

the receiver has minimized this

effect. Th e r esulting contribut ion

to amplitude measur ement a ccuracy

is dependent upon th e reflection

ma gnitude. It is typically less tha n

±1.5 dB, but degrades a s th e reflec-

tion magnitu de approaches 80 dB.

Refer t o the cur ves in th e specifica-tion tables of the H P 8504B operat -

ing man ual for detailed informat ion.

For mea sur ement s of sma ll reflec-

tions, dynamic accur acy can be a

significant uncertainty term.

Connector repeatabil i ty

There will always be some insert ion

loss in a ny fiber-optic connection.

When the HP 8504B ma gnitude cal-

ibrat ion is executed, th e loss at the

test port connection is effectively

removed from the measurement of

the calibration stan dar d. However,

when t he DUT is connected to th e

test port, the inser tion loss of this

connection may be different tha n

the connection made during calibra-

tion. If the insert ion loss is higher

by 0.5 dB, subsequent ret ur n loss

meas urem ents will be twice this or

1 dB larger th an t he actual value.

This is because both t he stimu lus

signal an d th e reflected signal will

experience a 0.5 dB atten uat ion.

Similar ly, when t he DUT connection

to the test port is better tha n theconnection at calibrat ion, ret urn

loss values will be worse t han actual

by two times the insertion loss

improvement.

It is th erefore importa nt that proper

connector care and usage be observed

to min imize th e effects of conn ector

repeatability.

Transmiss ion path losses

Transmission path losses in the

DUT affect m easurements in a way

similar to connector ins ertion loss

repeatability.

Large r eflections can degra de the

meas urem ent of other r eflections

furth er int o the DUT. An exam ple

would be if a simple Fr esnel air ga p

existed in front of other reflections.

3.5% of th e ener gy is reflected a t

the glass-to-air interface, 96.5%

tr avels to the air-to-glass int erface

where again 3.5% of th e ener gy is

reflected. The net r esult is tha tonly 93% of the stimulus signal

(an 0.3 dB loss) reaches reflections

beyond t he a ir gap. Sim ilarly,

only 93% of th e reflected ener gy

will reach the instrument detector.

The resu lt is similar t o ha ving an

0.3 dB lossy splice, which would

lead to 0.6 dB measur ement err ors.

Measurementaccu racy vs. posit ion

A port ion of the int erna l light

path to the reference mirror in t he

HP 8504B is an open beam en viron-ment . As the r eference mirror is

moved and th e length of th e open

beam pa th is increased, ther e will

be some beam divergence. This beam

divergence can resu lt in power var i-

ation in the reference beam at the

detector. This phenomenon is very

repeatable. Consequently, most of

th is effect is removed from th e mea-

suremen t a s par t of the factory cali-

bra tion process.

The HP 8504B detection scheme

requires that the r eference mirrorideally moves at a constant velocity.

However, ther e will be some velocity

jitter.

The err or due to the combination

of the above two err or sources is

appr oximat ely ±1 dB.

Sweep-to-sweeprepeatabil i ty

Sweep-to-sweep repeat ability is the

sweep-to-sweep am plitude va riat ion

seen when measur ing a known stable

reflection. This err or source is also

relat ed to the mechanical movement

of the int erferometer m irror. It does

not include t he effects of noise when

measur ing reflections nea r t he instru -

ment noise noise floor. It is specified

to be less th an ±0.5 dB.



Polarization sensit ivity

The stimulus signal from t he

HP 8504B is ra ndomly or only par-

tially polarized. As th e light tr avelsto and from th e DUT, its sta te of

polar ization will vary. Alth ough th e

HP 8504B receiver is designed to be

insensitive to the polarization char-

acterist ics of the r eflected light,

th ere will be some un certaint y in

amplitude measur ement accura cy

due t o polarizat ion, even if the source

stimulus is ra ndomly polarized. This

uncert aint y is typically less th an

±0.75 dB.

Calibration unce rtainty

When a ma gnitude calibra tion is

perform ed with t he HP 8504B, the

instrum ent uses the known return

loss value of a calibrat ion st an dar d

to remove man y system at ic err ors

in subsequent measu rements. The

cables supplied with t he HP 8504B

can be used as a reflection sta ndar d.

The retur n loss value at the open

end is consist ent with in ±0.1 dB. This

uncertainty in the a ctual retur n

loss of th e calibra tion sta nda rd will

result in uncertainty in DUT mea-

sur ement s of appr oximat ely ±0.1 dB.

To mainta in th is low uncerta inty,

the calibration stan dar ds (typically

the end of a fiber pa tch cord) mu st

be clean and undama ged.

Chromatic dispersion effects

Par t of the propagation path of the

reference signal is in a n open beam.

In most cases, light traveling in the

DUT pat h is all in fiber. When m ak-

ing measur ements at 1550 nm, the

amount of chr omatic dispersion

experienced by the light t ra veling

in th e reference path will be less

than that in the DUT path. This

mismatch in dispersion results in

a broadening and su bsequent drop

in th e peak value of the displayedreflection “impu lse”.

The peak value will decreas e mono-

tonically as a function of the length

of dispersion mismatch. This effect

is consistent and has been corrected

out by the H P 8504B. The instrument

assumes a dispersion coefficient of

17 ps/(nm*km). The result of this

corr ection leaves a r esidual err or

on th e order of ±0.3 dB.

The pr oblem becomes d ifficult when

the pa th t o the DUT is both in fiber

and a n open beam. The effects a rethen very difficult to remove from

the measur ement, and subsequent

uncert aint ies due to chromatic dis-

persion can approach 5 dB. The user

has the option of disabling the inter -

nal dispersion correction to facilitate

his own correction meth ods.

Effects of sourcespectral width

The spectral width of the HP 8504B

source is approximately 55 nm.

Another uncertaint y component will

exist if th e DUT r eflection charac-

teristics vary over th is spectral ra nge.

The level of uncerta inty is dependent

on the DUT characteristics.

Bibl iography

S. Newton, “Technology tren ds in

optical reflectom etry”, Photonics

Spectr a, November 1991, pp. 118-126

H. Booster, H . Chou, “High er R eso-

ution for Backscatter Measurements”,

Laser s an d Optr onics, October 1991,

pp. 27-30

Pos i t ionalAccuracy Issue s

The a ccur acy of the H P 8504B in

determ ining the r elative location

of reflections is based on its ability

to cont rol and monitor th e position

an d velocity of the reference mirr or.

This uncert aint y is less tha n 2% of

the measu rement spa n. To have the

highest a ccuracy, the nar rowest span

th at includes th e two events of

interest should be used.

Summary

In genera l, the individual error

components a re un corr elated. The

total measurement u ncerta inty is

determ ined with an "RSS" (Root

Sum Square) analysis, and n ot a

linear su mma tion.

Return loss (dB)

14.614.6514.714.7514.814.8514.914.95

6

5

4

3

2

1 R e l a t i v e o c c u r a n c e s

λ =1550 nmAvg =14.73 dBStd =±0.085 dB

Calibration Standard Repeatability

Return loss (dB)

14.8514.914.9515.015.0515.115.1515.2

6

5

4

3

2

1 R e l a t i v e o c c u r a n c e s

λ =1300 nmAvg =15.003 dBStd =±0.065 dB

19



For more information, callyour local HP sale s officel i sted in your te lephonedirectory or an HP regional

office l isted below for thelocation of your nearest salesoffice.

United States:Hewlett-Packard CompanyTest and Measurement Organization5301 Stevens Creek Blvd.Bldg. 51L-SCSant a Clara, CA 95052-80591 800 452 4844

Canada:Hewlett-Packard Canada Ltd.5150 Spectru m WayMississauga, Onta rio L4W 5G1(905) 206 4725

Europe:

Hewlett-PackardEuropean Marketing CentreP.O. Box 9991180 AZ AmstelveenThe Netherlands

Japan:Yokogawa -Hewlett-Pa ckar d Ltd.Measurement Assistance Center9-1, Takakura-Cho, Hachioji-Shi,Tokyo 192, J apa n(81) 426 48 3860

Latin America:Hewlett-PackardLatin American Region Headquar ters5200 Blue La goon Dr ive, 9th F loorMiam i, Florida 33126, U.S.A.

(305) 267 4245/4220

Australia/New Zealand:Hewlett-Packard Australia Ltd.31-41 Joseph Str eetBlackburn, Victoria 3130, Australia131 347 Ext. 2902

Asia Pacific:Hewlett-Packard Asia Pacific Ltd.17-21/F Sh ell Tower, Time Squ ar e,1 Mather son Street, Causeway Bay,Hong Kong(852) 2599 7070

Data Subject to ChangeCopyright © 1992Hewlett-Packard CompanyPrinte d in U.S.A. 3/955963-7191E

HP-PN8504-1_Measurements of Lightwave Component Reflections

Documents

Transcript of HP-PN8504-1_Measurements of Lightwave Component Reflections