HP 416A - Page 48

of the S ET T O FULL SCALE a dj us t ment pr ovid e s s u ffi cie nt va r iation in ga in i :n. the
reflect e d chann el to co m pen s a t e fo r a b ou t 6 db of di ff erenc e in th e r ec tifica tion ef-
fici e nci e s of the two dete c tors.
Sinc e, for a g ive n level of i np u t, t he d e t ec ted output also varies with the
frequency of the inpu t, in b r oadband ope r ation t he adjustment of gain in the re -
fleeted c hannel c annot f ully co n1 pensate fo r diff e r e nce s in re c tific a tio n ef fi c ie n cy
between fo .e d e te c tors. Tw o f a c t ors contr i bu t e t o th e po s s ib ilit y of err or : (1) lac k
of tracking betwe e n th e va r ia t ions wit h fre qu en cy in the level of the outputs from
the two dete c tors and { 2. ) gain in se rted b y SET TO FULL SCALE to compens a te for
difference s in r ect i fic a t io n e ff iciency at on e fr equen c y m ay b e t oo m u ch or too
little at anothe r fr e q ue n cy .
Error Due to Va riat ion s in C h ar a cteri sti c s of Cou pl~~ --
The c ou pling eff icie n c y of a d irect ion al co upler varies with frequency , there
being a d r op in th e leve l of d el i vere d p ow er at the edges of the band. These v ari a -
tions wi th f r eq ue n c y in th e level of po wer out of the s e co ndar y a r rn of a di r ecti o nal
coupler a l so va ry fr orn coupler to c o uple r . If th e c haract eri sti c s of th e reflected
and incident c ou pler s va rie d in exa c t ly the sam e manner , c oupler deviations wou l d
track , and a r at io i nd epende nt of f re q ue n cy woul d b e m a int ai ne d between th e inci de n t
and refle cted v oltages. S in ce th e c h arac t er i stics of tw o c o uple rs do no t dev iat e in
exactly the s a:;.ne mann er, in broa db an d operatio n lack of t racki ng in co up le r c ha ra c-
teristi c s co nt r ibu t es to th e p os sibilit y of err or .
Loss Be ca u se of Power Split Off by I ncid e nt C oup ler
Th e inc id e nt wa ve su s ta ins a s li g ht p ow er l oss because of the power split
off by the s econ d ary arm. of the inc iden t co u pler. Wh en a 20 - db c o up ler i s used in
the incid e nt c h annel, o nly 1% of t he in c id e nt p o wer is split off by th e coup l er b ut
as a r esult of the s pli t ti ng , th e le v el of powe r a t th e input to t he incident cou p le r
will be 1 % high er t han th e level of po wer a t t he l oad and , the :r ef o re , t h e powe r r e -
fleeted fr om th e lo ad wi ll be ~lig h tly le s s tha n if th e in c id e nt pow er ha d no t b een
samp l ed . E rr o r whi c h mi gh t re s ult fron1 th is f ac to r is compensated f or dur in g
calibration b y adj ust m ent of th e SE T TO FU LL SC ALE c ontro l.
Spurious Si g na ls --
Gen er al - In a refle c to meter an y sig n al th at a rr ives at e ither th e for war d
(incident ) o r reverse (reflected ) de tector • other than t he de s i r e d i nci de n t and re --
fleeted sam ple s , c a n b e tern 1e d a spurious si gnal. S pu rious s ign a ls a :r ise f ro rn
three sou rce s : lac k of perfec t di re c tivity in the di rec t ion al co u pl e rs , in ev itab le
small mismat che s of c om po nen t s of the reflect om e t er, a n d d i s c ontinui tie s ar i si n g
from to lerances in dirnensions, waveg u ide r n.isalignrnen t, e tc,, T he in a nn er in
which spur i ous sign al s are gen er ate d a nd p r i ncip a l c i rcuit pa th s a :re in dica t ed in
Fig. 15 .
At In c id ent D etector - A s descri bed b el o w, thr e e spurious signals of c on -
sequence can ar r iv e a t the forw a rd d etector. A ppr ox im ate v alues giv e n be low a r e
referred to r f vo lt ag e in the ma in gu i de.
~ 33 -