1. v i 1 1 mm i m ^f a wr ^^^k ^r ^^^l ^^^h m ^r ^^r^ ^^^g ... · ^^f ^^v^^^i r i^h j f ^b ^b ^^b...
TRANSCRIPT
V
f*
H
■'-: *s s « •
'Ki 1 J L ■ ■
if" fc"-^
£ a£. ?-" ■ * : * >
:;. > ■ r IT-. ; * ■
'<*. I 1 ' f ' l w *- 1 ?* i" ■ q h j f
i f '#■>■ J»" «,,
.;-: ."> ! P ™"f ^ \ P'
F% ^^^^k i ^ t ^^k ^ ^ ^ ^ ^ L
^^k ^ ^ ^ ^ ^ k ^ T ^ ^ ^ ^ ^ ^ B ^ ^ __ ^ ^ ^ ^ ^
V^^^J^k ^^^^^B ^ ^ ^ ^ ^ ^ l^m
^^^^^^^^^^ ^^^^^^^^ ^^^^^^^^H ^^K ^^^^^^^^^B ^^^^^^^^^^^^^^^^^^H ^^m^^
^ ^ ^ ^ ^^^ ^^^MT *
^^M^ ^ ^ H ^H_^^^^^^ ^^H
^^^ _^^^H .^^F _« ^^H^^^^^T ^^MT
^K ^m i ^ L_ ^ F J ^ ^ ^ ^ H
^C^H HR?SSB1^--^ ^ ^ ^ ^ ^ M ] u j | ^ ^ « r f Q r j i t
'S^^^^sS^**?? 3^? ^ j — B s t r V 3' J f c . V p j A i t f j t - ! ■ ■
^ ^ ^ 1 ^H. V I \ 1 1 k m * l ^ B B ■
L ^^b ^ f k ^^^M ^^^L ^^^H
^[ ■
^^^^H ^ |
.^■H ^ ^ ^ r ^ ^ ^ ^ H ^ ^ F ^^v^^^i ^r I^H J f ^ B ^B ^^B ^K
^^^^^^f ^^H
W^M W^ i^B I i
r ^B ^i 1 i ^V B ^^^^^^J
ifJiiS^^KR "fi*T^Si
j ~ f J P J T V
' ^^H
■ : .
■ ^^^H 1 MM I m ^f A [ ^L^r J Wr £
^M ^^^d ^r ^^r^ ^^^ . ^^^g^^
K^^fl
^^^_ ^^V^^^^H ^^^^H
^P^L. ^V ^ ^ ^ b
p ^ ^ L ^ ^ ^ ^ ^ ^ -^ ^ L ^ ^ ^ ^ ^ L
^ L ^ ^ ^ ^ ^ ^ k ^ C ^^^H^k ^^^L
I ^T^fc V ^ L ^^^L i ^ h ^ ^ L
1 K ^ K . ft ^L ^^^^^^^H
^L^H . : S K S
■
^^H
&ft ■
.
r ^H ^^^k ^r ^^fl ^ ^ ^ ^ ^ H ^^^^^^ ^^^^H
M^^V ^ ^ ^ ^ ^
^^^^^H ^^^^B
^^^^^^^^^H ^^^^^^H
^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ■ ^ 1 ■ ^^^H ^ ^_ ^^^B ^ ^ ^ ^ ^ ^ ^ ^ H
^ ^^B ^ ^ t
L ^H ^^^^^^^H
^^Br*i3 ^^Hsdt^ r^jSP^SfiT^G i&t^&SSi^^^B H J ^ f i " * i * i ^ J irf^L * "''■"'■ ^^ff*S^*^y-HUH
M <
15-039
Contents
General in t roduct ion 3
Measuring methods and experiences w i t h the object ive measurement of reference equivalents in Norway
by Trond Ulset 3
Review of current te lephonometr ic measurements in Finland by I. Jantt i & T. Tuisku 7
Acoust ic feedback in telephone sets by Einar Laukli 10
Factors determining the reproducibi l i ty and accuracy requirements of te lephonometry measurements: Review of some practical experiences
by I. S a lama 13
Frequency response measurements on the telephone sets type 7 3 D , F68 and 47E connected via a telephone line
by S. A. Jager & P. Schnack 17
What should be required of a Telephone Measuring Equipment? Product ion control experiences
by 0 . Larsson 21
General Introduction On the 12th and 13th November
1974 a special Telephonometry Symposium was held in Copenhagen. Hosts for the symposium were KTAS (Copenhagen Telephone Company) and JTAS (Jutland Telephone Company), and initiative for the meeting was mainly due to the efforts of Mr. P. V. Arlev and S. A. Jager of JTAS.
and Sweden contributed to the success of the meeting.
The symposium consisted of lectures and discussions related to developments in the field of telephonometry, and participants from Eng-and, Denmark, Finland, Norway
Of the papers presented, two summaries and six full texts were later printed in the B & K Technical Review No. 1 , 1975 , copies of which are available on demand from B & K . This Application Note prints a further five papers plus an extra paper which although not presented at the symposium received some discussion there.
Papers printed in Technical Review No. 1, 1 975:
Problems in Telephone Measurements, by Norman Gleiss
Proposals for the Measurement of Loudness Ratings of Operators' Headsets, by R. B. Archbold Comparison of Results obtained by Subjective Measuring Methods, by lb Gilberg
Repeatabilities in Electro-Acoustic Measurements on Telephone Capsules, by R. E. Walford
Stable Subset Measurements wi th the 73D, by K. Damsgaard Vibration Testing of Telephone Equipment, by R. Fluhr
Measuring methods and experiences with the objective measurement of reference equivalents in Norway by Trond Uiset, Sectional Engineer.
Norwegian Telecommunications Administration, Central Laboratory Acoustics Group.
Summary Information on the use of objec
tive measuring of reference equivalents in Norway. Description of measuring methods and adjustment procedure for the measurement of telephone sets equipped wi th linear microphone and amplifier. Description of future adjustments and improvements. Description of additional measurements made in connection wi th the telephone transmission measuring set. Experiences wi th the use of objective reference equivalents.
1 . Introduction The purpose of this paper is to
give some information on the measuring methods for objective telephone measurements used by the Norwegian Telecommunications Administrat ion.
Our first objective electro-acoustic telephone transmission measuring set was purchased in 1959 . It
was a laboratory version of the Siemens Objektiver Bezugdampfungs-me&platz (OBDM). The system was used for equivalent measuring of telephone sets. Later on we became interested in measurements of objective reference equivalents for special equipment, automatic answering machines in particular, to ensure that this equipment had a transmission efficiency comparable w i th that of telephones.
In 1 9 7 1 , we decided to procure a Bruel & Kjaer Electroacoustic Telephone Transmission Measuring Set Type 3 3 5 2 . Our work itself had not been considerably changed, but the amount had increased. Relatively soon we were using only the new measuring set.
As from 1st January, 1974 , a new unit has been established at our Central Laboratory. One of the main tasks of the new unit is to study measuring methods for subjective and objective telephonometry and to participate in international cooperation. In the light of the constantly increasing amount of objective measurements and the prepara
tion of improved measuring methods, the Central Laboratory has purchased another measuring set Type 3 3 5 2 .
2. Measuring Requirements In 1967 , the Norwegian Telecom
munications Administration introduced a new telephone set. This set was equipped wi th an electrody-namic microphone and a two-step microphone amplifier. All special sets provided by the Norwegian Telecommunications Administrat ion today are also equipped wi th linear microphone and amplifier. This means that by far the majority of the objective measurements of reference equivalents are conducted wi th sets without carbon microphone. In the preparation of standard measuring methods it is consequently not necessary to consider factors such as non-linearity and packing. In the examination and control of carbon microphone sets we fol low the prescriptions corresponding to OREM C, described in the Instruction Manual for the Type 3 3 5 2 .
n Norway the objective reference
3
equivalents are presented as a function of inserted line length. Because of the linearity of the microphone 2 or 3 measurements are sufficient to describe the variation of the transmission/recept ion equivalent as a funct ion of inserted line length. Sidetone equivalents must be measured for several lengths of inserted line in order to describe the variat ion.
The feeding system of most new types of telephone exchanges in Norway is 4 8 V , 2 x 2 5 0 O , and this has usually been used during measurements. The artificial line equivalent generally used has been 0 , 4 m m , 2 8 0 0 / k m , 4 5 n F / k m .
Objective telephone measurements are conducted for the fol lowing purposes:
— Type control of special sets — Type control of headsets — Type control of privately owned
equipment — Production control of telephone
sets and handsets — Production control of micro
phones and earphone capsules — In connection wi th minor
changes on old models
4
3. Measur ing Methods 3.1. OREM A (NJ
Usually we measure the properties of a complete telephone set; not just the microphone and earphone capsules. That was the reason for originally selecting the OREM A procedure. This procedure prescribes an artificial mouth in the REF-position. This led to a mechanical contact between the labial ring
and the mouthpiece. We therefore had to use the AEN position. We also chose to use the new IEC audi-ometric coupler instead of the NBS 9A coupler. The sound pressure was adjusted wi th a SFERT-baffle to 93 ,6 dB re 2 x 1 o ~ 5 N / m 2
(—0,4dB re 1 Pa). The receiving reference level was also set at 93 ,6dB re 2 x 1 0 " 5 N / m 2 . The meter damping was OB DM. The rest of the procedure is identical wi th OREM A described in the Instruction Manual for Type 3 3 5 2 . We also chose to read the left deflection on the reference equivalent meter because we thought that this was simpler and safer.
3.2. Orem C (Nj In the beginning of 1973 we
changed the measuring procedure.
41 mm 30 mm h H \*~ —*\
750472
Fig. 1. Voice pressure measuring j ig used in OREM C (N)
The main reason for the change was that we wished to use an unweighted sound pressure from the artificial mouth. This meant that the SFERT-baffle for adjusting the sound pressure could no longer be used for calibration, so we designed a support for a 1 / 2 " microphone (Fig.1).
We chose a reference sound pressure of 94 dB re 2 x 1 0 - 5 N / m 2
(1 Pa). This sound pressure from the artificial mouth was measured 25 mm in front of the labial r ing.
The fol lowing parameters describe the measuring procedure:
— Artif icial ear: IEC Audiometr ic — Artif icial mouth: AEN position — Sweep: 0,2 — 4,0 kHz — Meter scale: y = kx ° ' 6 dB — Meter damping: OBDM — Weighting network: Off
Reading of the left deflection of the objective reference equivalent meter.
Modern l ightweight headsets are loaded wi th 2 cm3 artificial ear dur-
•
n ""
K .■
"£ £ k -J , ^ '1 3 * r
^
^ \
s
s
s
^
ing measuring of receiving and side-tone equivalents.
3.3. Modifications of the existing measuring method
The purpose of further preparation of objective measuring method is to
Describe the qualities of the telephone set in the most appropriate way w i th as few basic errors as possible Achieve the best possible correspondence between the subjective and the objective measuring results.
rable peak l imiting may occur in the amplifier. Apart from this, such a high pressure wi l l not be necessary. Experience has shown that a sound pressure of 84 dB re 2 x 1 0 ~ 5 N / m 2 , adjusted according to procedure OREM C (N), results in an objective transmission equiva-ent deviating relatively little from
subjective transmission equivalents rel. NOSFER.
Because it is not possible to carry out subjective tests in the Norwegian Telecommunications Admin i stration today, we have concentrated our effort on the preparation of the most suitable objective measuring procedures.
For sets w i th linear microphone and amplifier a sound pressure of 9 4 d B re 2 x 1 o ~ 5 N / m 2 (1 Pa) is too high for measurements of transmission and sidetone equivalents. When the conditions are unfavou-
We also re-considered the choice of sweep width and decided that a sweep of 3 0 0 — 3 3 0 0 Hz corresponds more to the bandwidth of a telephony channel.
The meter function and the meter attenuation for the objective reference equivalent meter also have to be considered. A meter function of y = kx ' dB corresponds more to the presented results. The question of using the Slow meter needle damping or whether it would be more correct to use the so-called OBDM damping, i. e. a damping corresponding to that of the meter needle in the Siemens OBDM, is also something to consider.
The practical execution of objective reference equivalent measurements is important and is given much consideration. In the preparation of measuring procedures for telephone sets w i th linear microphone and amplifier, we may disregard phenomena like packing. The measuring procedure itself w i l l accordingly be simpler, quicker and safer than a procedure for measuring carbon microphones. It is necessary to check that the handset is appropriately attached and that acoustic connection between the earphone and the coupler is good. The acoustic coupling is most simply checked by means of a Frequency Response Tracer.
We must take due account of the conditions under which the measurements are carried out. The essential factors are those influencing the amplifier of the telephone set, the line current and the sound pressure. Too high a line current may cause part or complete destruction of elements in the amplifier; unfavourable combinations of line current and sound pressure may cause peak l imiting of the amplifier.
For the subscribers' protection most modern telephone sets are equipped wi th a limiter in the earphone to prevent aural damage caused by acoustic shock. In measurements of reception equivalents we must take this into considerat ion when we decide the pressure eve I.
Finally, we must consider the surroundings. As far as possible reflections must be avoided during measurements of transmission and side-tone, and noise must be prevented from disturbing the measurements.
3.4. Additional measurements In addition to the objective refer
ence equivalent measurements we carry out the fol lowing measurements on the measured set:
(a)
(b) (c)
Characteristics of cu r ren t / voltage Harmonic distortion Adjustment attenuation (impedance)
(a) The characteristics of cu r ren t / voltage, i .e. voltage drop over the telephone set as a function of the line current, are found by varying
5
the voltage of the power-supply and reading off the line current and the voltage over the set connectors.
(The switch of the voltmeter is in position Line 2).
(b) We have considered different methods for the measurement of harmonic distortion:
Measuring of harmonic components by means of a spectrometer. Measuring of harmonic distortion by means of a slave filter. Measuring of harmonic distortion and products of intermodulation by means of sweep and highpass fi l ter.
The first two methods give information on harmonic distortion as a funct ion of the frequency. The result may be wr i t ten out in curves. The last method gives information on a mean value of the sweep area.
We are usually interested in information on the total harmonic distort ion, not the harmonic components, and f ind that the use of a spectrometer is not appropriate. In the opposite case, however, it is used.
The use of a slave filter requires a room wi th a good noise insulat ion, usually special rooms like ane-choic chambers. It is possible that a combination of a slave filter and a frequency analyzer Type 2 0 2 0 or 2021 may prove useful.
The use of sweep and high pass filter gives an expression of mean generation of harmonic distortion and intermodulation over the frequency range of the sweep. This is a fast measurement, and we have found it gives adequate information. Today we use two measuring techniques:
Sweep 2 0 0 4 0 0 0 Hz, 1 sweep per /s , fi lter w i th limit frequency 5 0 0 0 Hz, sound pressure 8 4 d B re 2 x 1 f j - 5 N / m 2 , (— 10dB re 1 Pa) 2 5 m m in front of the lip ring of the artificial mouth.
Sweep 300 — 8 0 0 , 2,5 sweep per /s , filter wi th limit frequency 9 0 0 Hz, sound pressure 94 dB re 2 x 1 0 ~ 5 N / m 2 , (1 Pa) 2 5 m m in front of the lip ring.
(c) The return loss is expressed by the ratio between the impedance of the equipment and a reference impedance given in the equation.
z z
+ ZR
-ZR
Z = equipment impedance. ZR = reference line impedance
The measuring shown in Fig.2.
schematic is
zg«zR
u
3.5. Use of plug-in units For use in the examination of cap
sules and as a reference, we have constructed two plug-in units. We have used two ordinary telephone sets, one of each of the most common models in the Norwegian Telecommunications system.
In order to increase the benefit of the use of plug-in units we are constructing a special reference set. From a transmission technical point of view, it wi l l be a perfect telephone set built w i th particularly strict tolerances in components and stability. Such a reference set constructed as a plug-in unit has great advantages when used in type controls and environment tests of capsules as wel l as of telephone sets.
750473
Fig.2. Arrangement for return loss measurement
The ratio between impressed and measured voltage when Z-j = Z 2 is
U 0 = 2 Z + ZR
U Z — Z R
and the result may be read off the frequency response tracer or the level recorder.
4. Experiences Our experiences from the use of
objective measuring of reference equivalents of telephone equipment in the Norwegian Telecommunications Administrat ion have been very satisfactory. An objective telephone transmission measuring set is indispensable in the production control of larger quantit ies. We have also
6
found that objective measurements of reference equivalents give us reliable information on the transmission qualities of the equipment. The transmission qualit ies of a telephone set may be described in a simple and wel l defined measurement presentation. The measure
ment precision seems to be satisfactory.
The determination of reference equivalents rel. NOSFER based on OREM measurements seems for the present to be somewhat insecure. It is therefore w i th great interest that
we are watching the work carried out by CCITT SG XII in this field because an improved conformity between the OREM and NOSFER measurements wi l l make the benefits of the objective measurements of reference equivalents even greater.
Review of current telephonometric measurements in Finland by / . Jantti and T. Tuisku
Equipment used in reference equivalent measurements
As a national standard for measuring objective reference equivalents a Siemens OBDM is used. It was supplied to Telephone Laboratory of General Direction of Posts and Telegraphs (PLH) in 1960 .
Measurements w i th OBDM are made as instructed by the manufacturer. Following points may, however, make an exception to the common practice:
the shape of the lip ring of the artif icial mouth has been slightly modified in order to al low measurements of sets w i th modern short handsets. when reading the OBDM meter the average position of the needle is observed. carbon microphones are condit ioned in principle according to IEEE standard 2 6 9 - 1 9 7 1 , but technicians show a tendency to have individual features in the method.
In addition to Telephone Laboratory of PLH objective reference equivalent measurements are made by fol lowing institutions:
Helsinki Telephone Company (HPY), research and on-receipt inspection. OY L . M . Ericsson AB (LME), production control. PLH/Division for Supply (HO), on-receipt inspection.
* General Direction of Posts and Telegraphs Telephone Laboratory, Helsinki.
#
(dB)
15 -
10
5
I- / / / Data points inside
these limits
5 10
LME 1/3/0.53 LME 2/3/0.62 LME 1/2/0.75 LME 1/1/0.84 LME 2/1/0.71 LME 2/2/0.74 PLH 1/3/0.95 PLH 1/1/0.92 PLH 1/2/0.82 PLH 2/3/0.95 PLH 2/1/0.96 PLH 2/2/0.67 HPY 1/3/0.93 HPY 2/3/0.94
15 OBDM (dB)
Telephones with carbon microphones, linear relations for sending
LME 1,2 : 2/LTM 357500 HPY 1,2 PLH 1,2
equipments based on OBDM
eg LME 1/2/.95 : LME 1, test 2, correlation coefficient .95
Correlation coefficient deviates significantly from other coefficients calculated to the same equipment 750474
F i g . l .
7
3
(dB)
"I
- 3
Data points inside these limits
OBDM (dB)
PLH 2/3/0.97 PLH 2/2/0.93 PLH 1/3/0.95 HPY 2/3/0.90 PLH 1/2/0.89 HPY 1/3/0.81 LME 1/3/0.86 LME 2/3/0.49 LME 2/2/0.48 LME 1/2/0.51
Telephones with carbon microphones, linear relations for receiving
LME 1,2 : 2/LTM 357500 HPY 1,2 PLH 1,2 equipments based on OBDM
eg LME 1/2/.95 : LME 1, test 2, correlation coefficient .95
Correlation coefficient deviates significantly from other coefficients calculated to the same equipment
750475
Fig.2.
I
LME 1/2/0.96 LME 2/2/0.97 LME 1/1/0.96 LME 2/1/0.96 LME 1/3/0.995 LME 2/3/0.995 HPY 2/1/0.99 HPY 1/1/0.97 PLH 2/1/0.99
PLH 1/3/0.96 PLH 1/1/0.99 PLH 1/2/0.99 PLH 2/3/0.96 HPY 2/2/0.996
HPY 1/2/0.97 PLH 2/2/0.996 HPY 1/3/0.97 HPY 2/3/0.94
Stable telephones, linear relations for sending
LME 1,2 : 2/LTM 357500 HPY 1,2 PLH 1,2 equipments based on OBDM
OBDM (dB)
eg LME 1/2/.95 : LME 1, test 2, correlation coefficient .95
* Correlation coefficient deviates significantly from other coefficients calculated to the same equipment 750476
Data points inside these limits
Fig.3.
8
J L OBDM <dB)
Data points inside these limites
PLH 2/2/0.99 PLH 1/2/0.99 HPY 1/2/0.998 HPY 2/2/0.99 HPY 1/3/0.95 PLH 2/3/0.95 HPY 2/3/0.90 PLH 1/3/0.95 LME 2/2/0.98 LME 1/2/0.98 LME 2/3/0.70 LME 1/3/0.89
Stable telephones, linear relations for receiving
LME 1,2 : 2/LTM 357500 HPY 1,2 PLH 1,2 equipments based on OBDM
eg LME 1/2/.95 : LME 1, test 2, correlation coefficient .95
correlation coefficient deviates significantly from other coefficients calculated to the same equipment
750477
Fig.4.
LME use their own design of transmission test equipment 2 /LTM 3 5 7 5 0 0 . The equipments used by HPY and PLH/HO are manufactured by HPY Research Institute. They are based on the principle developed by Dr. Braun and are calibrated against OBDM by means of stable telephone sets. Though the purpose has been to make these equipment essentially identical w i th OBDM, there are small differences in the frequency sweep, artificial mouths and time constants of the meters.
HPY also has a SRE working standard for subjective tests, but it has not been very much used recently.
A n Electroacoustic Telephone Transmission Measuring System B & K 3 3 5 2 has been ordered by the Telephone Laboratory of PLH, and was delivered in November 1974 .
Comparison measurements of reference equivalents
During the last three years, objective reference equivalent values obtained w i th OBDM and other equipments have been compared. Measurements have been made on both stable and carbon microphone sets, which have been circulated between all the institutions mentioned above. The main reason why these measurements were started, was the use of completely different objective R. E. measuring equipment at manufacturers production control and buyers' on-receipt inspections. In addition to that, correlation w i th the national standard was needed.
Some results of these comparisons are shown in f igures 1 and 2 (17 sets w i th carbon microphones) and 3 and 4 (3 stable sets). These results do not encourage one to think that the difference between values could be predicted, even though always same types of sets are measured.
A review of the network The telephone network of Finland
has been planned in accordance with recommendations of CCITT. The reference equivalent f igures used in planning agree w i th those in P. 2 1 , but the reference equivalents have to be measured w i th OBDM.
The publication "Puhelinverkkojen rakennemaaraykset" (Regulations for building telephone networks) essentially states:
"As reference equivalent values those values shall be used, which have been obtained by loudness comparison to an internationally accepted standard speech circuit (NOSFER). Until further notice an objective equipment for measuring reference equivalents (OBDM), which is located in Telephone Laboratory of PLH, shall be used as a national standard".
The regulations do not say anything about the difference in values obtained wi th subjective and objective methods, and the practice has been to use only values measured w i th OBDM. This means that instead of the target value (mean value of SRE's of several sets SRE = 4 , 5, . . . 7 dB at GO line) the telephones have 6 . . . 7 dB higher SRE's.
Measurements of speech voltages were started in October this year in our network. The first results make us believe that the send reference equivalents really are somewhat high. The speech volume at the line terminals of the set has a mean value — 2 7 , 6 vu and a standard deviation 4,1 vu at local calls.
An immediate question is, what , if anything, we should do to correct the send reference equivalents of the local telephone circuits. At present we feel, that it would not be wise to hurry, but to wait unti l the work on Question 1 5/XII is ready. After that we perhaps wi l l have a better view of the situation.
dB
0
- 10
- 2 0
- 3 0
- 4 0 x
Handset lifted 20 mm
Handset lifted 40 mm
kHz 750481
Fig.4. Ratio of receiver drive signal and microphone pick-up signal for various heights of hand set above subset
dB 0
- 10
- 2 0
30 kHz
dB + 10
0
- 1 0
Fig.5.
2 5 2 5 2 5 kHz 750482
Upper: ratio between receiver drive signal and microphone pick-up signal as the handset approaches a user Lower: Effect of a capacitance C on a frequency-dependent feedback path
difference varies according to frequency when the handset is kept in different positions: in free f ield, in normal positions on the telephone set, on table w i th holes parallel to table and on table w i th holes down.
In all examples shown, the set terminal is open and wi thout any line length, if not particularly ment ioned.
The last mentioned position wi l l give howling at about 3 kHz, at zero degrees phase difference.
The telephone sets are chosen to i l lustrate the different measuring methods, and they are not typical for the Norwegian sets produced today by A / S Elektrisk Bureau.
If the handset is put on the telephone set in its normal position and then gradually l ifted, we can see the peaks in the stability responses get lower and the resonances move to other frequencies (Fig.4).
We found that the phase difference between supplied and measured voltage change about three t imes 3 6 0 degrees during the first couple of cm the handset is lifted from the telephone set. The signal are in phase three t imes, and if the loop gain is sufficient, unstable condit ions can occur.
Handset position In the upper part of Fig.5 is dem
onstrated how the handset position influences the degree of stability: in free f ield, vertical in hand and held against the ear.
The lower part indicates ihe inf luence of a frequency dependent feedback in the microphone amplifier (a capacitor), C = 0 and C1 to C4 ( increasing capacity). The handset is here placed on a table w i th holes down.
Free field We have also kept the handset po
sition constant, in free f ield, and varied other factors.
On the upper left in Fig.6 is shown the result of two different types of capsules w i th the same sensitivity but different frequency responses. The middle curves show the effect of the capacitor in the feedback of the amplif ier, and on the right we have two different cord capacities (corresponding to the feedback capacitor).
In the middle response row, the left curves indicate how important the line matching (sidetone) is, here demonstrated w i th open line and terminated w i th 6 0 0 0 . The two responses on the right show two different line lengths, zero and one km (open line).
On the lower left, the caps are fastened wi th different t ightness, and on the right, the microphone cap holes are varied.
Studying the mechanical feedback wi l l clarify the need for elastic suspension of the capsules.
11
dB
0
10
- 2 0
- 3 0
- 4 0
Handset if ted 20 mm
Handset ifted 40 mm
777777777
kHz 750481
Fig.4. Ratio of receiver drive signal and microphone pick-up signal for various heights of hand set above subset
dB 0
- 10
- 2 0 -
- 3 0 kHz
dB + 10
0
10 5 kHz
750482
Fig.5. Upper: ratio between receiver drive signal and microphone pick-up signal as the handset approaches a user Lower: Effect of a capacitance C on a frequency-dependent feedback path
difference varies according to frequency when the handset is kept in different positions: in free f ield, in normal positions on the telephone set, on table w i th holes parallel to table and on table wi th holes down.
In all examples shown, the set terminal is open and wi thout any line length, if not particularly mentioned.
The telephone sets are chosen to il lustrate the different
The last mentioned position wi l l give howling at about 3 kHz, at zero degrees phase difference.
measuring methods, and they are not typical for the Norwegian sets produced today by A / S Elektrisk Bureau.
If the handset is put on the telephone set in its normal position and then gradually l ifted, we can see the peaks in the stability responses get lower and the resonances move to other frequencies (Fig.4).
We found that the phase difference between supplied and measured voltage change about three times 3 6 0 degrees during the first couple of cm the handset is lifted from the telephone set. The signal are in phase three t imes, and if the loop gain is sufficient, unstable conditions can occur.
Handset position In the upper part of Fig.5 is dem
onstrated how the handset position influences the degree of stability: in free f ield, vertical in hand and held against the ear.
The lower part indicates che inf luence of a frequency dependent feedback in the microphone amplifier (a capacitor), C = 0 and C1 to C4 (increasing capacity). The handset is here placed on a table w i th holes down.
Free field We have also kept the handset po
sition constant, in free f ield, and varied other factors.
On the upper left in Fig.6 is shown the result of two different types of capsules w i th the same sensitivity but different frequency responses. The middle curves show the effect of the capacitor in the feedback of the amplif ier, and on the right we have two different cord capacities (corresponding to the feedback capacitor).
In the middle response row, the left curves indicate how important the line matching (sidetone) is, here demonstrated w i th open line and terminated wi th 6 0 0 O. The two responses on the right show two different line lengths, zero and one km (open line).
On the lower left, the caps are fastened wi th different t ightness, and on the right, the microphone cap holes are varied.
Studying the mechanical feedback wi l l clarify the need for elastic suspension of the capsules.
11
dB
0 -
- 10
- 2 0 L
C = 0
Cord capacity
kHz
dB
0
10
- 2 0
30
Matching
open
60012
Line length
kHz
dB
0
- 10
- 2 0
i
P
Caps fastening
>— normal
I \t< /Til \
5
Jht
Micr. cap (holes)
nortma arger
kHz 750483
Fig.6. Effects of various parameters on howlback response
0,1 V u
dBSPL
70
60
mV
1
0,3
Fig.7.
2 b 2 5 2 5 kHz 750484
Direct measurement of voltage induced in a microphone output connection when a re ceiver is driven by a fixed amplitude signal
Measuring method 2 We can directly measure the in
duced voltage in the microphone when the receiver is driven by a signal (Fig.7). This method does not tell us if there is any chance of howling in the system, but for relative measurements it can be of great interest to use such a simple method. Critical frequencies can be studied in this way. We can also measure the frequency response of the receiver at the microphone entrance.
Below are shown examples of such measurements. On the left, the induced microphone voltage is measured when the handset is placed on a table w i th holes down. In the middle, the microphone voltage and receiver (loudspeaker) responses are shown when the handset is in free f ield, and on the right we have the same responses when the handset is placed on a table w i th holes parallel to the table.
Measuring method 3 A very simple method that can be
used e .g . in production control is shown in Fig.8. Varying the line
Tef. set Artif icial line 48 V 2 x 250 H
750485
Fig.8. Adjustment of line length to find minimum required to give a howlback-free system
length, this measurement tells us the min imum length that is necessary to give a stable system when the handset is lifted off the telephone set, when it is placed on a table in different positions or held in other ways.
12
Factors determining the reproducibility and accuracy requirements of telephonometry measurements. Review of some practical experiences by /. Sal am a
1 . Introduction The intention of this presentation
is to discuss the factors, which from the telephone company's point of view determine the accuracy and reproducibility requirements of the reference equivalent measurements.
Let me start by putting a simple question. Why do we have to measure telephone sets? Two answers can be given: we have to check the ful f i lment of the transmission plan and we have to check the quality of the telephone sets.
To discuss how these factors determine the accuracy requirements of the measuring equipment, I have to say a few words about the transmission plan of the Finnish telephone network, and the quality control methods used in the Helsinki Telephone Company.
2. The transmission plan In 1970 a new transmission plan
was adopted in Finland. It is based on the CCITT plan a), Recommendation G 121 (p. 21). In this plan the maximum value of the nominal sending reference equivalent of the subscriber system of the local exchange is 9 ,6dB (1,1 Np) and the corresponding receiving reference equivalent is 0 ,9dB (0,1 Np).
Based on these values there has been a tendency to draw up the reference equivalent requirements of the telephone set, so that less than 3% of the incoming and outgoing calls in a local exchange area wi l l exceed the given values. There is further a desire to limit the lowest reference equivalent of the local connection to a mean value of 2,6 dB (0,2 Np). Based on these facts the reference equivalent requirements for the telephone set as seen in Figure 1 are derived.
The transmission characteristics of a set are determined by connecting the set to a subscriber cable of 0,5 mm and 3 7 n F / k m , and to a
feeding-bridge of 60 V and 2 x 5 0 0 O in such a way that the mean value of a suitably large sample of a consignment must fulf i l the specifications in Figure 1. The length of the cable is 0 to 5 km. The standard deviation shall not exceed at any length the value 1,3dB (1,5dNp) on sending and the value 0,9 dB (1 dNp) on receiving. The measurements are made wi th an objective reference equivalent measuring equipment (OBDM).
These requirements imply that telephone sets w i th automatic level control have to be used. On the receiving side the cable attenuation
must be compensated completely, whereas on the sending side feeding loss and only a part of the cable attenuation must be compensated.
The experience has shown that the combined reference equivalent of the telephone set and the local line is approaching the allowed maximum value mainly on long subscriber lines. If the main objective is that only 3% of the calls of the entire exchange area wi l l exceed the limits 9,6 and 0 , 9 d B , a bigger exceeding percent value for subscribers far from the exchange may be allowed.
Reference equivalent
10 _ dB
9 -I
8 -
3 -
2
1 -
0
- 1 -
600
The mean value of the receiving reference equivalent
0 . . . 5 km
0 0,5 mm, 37 n F/km
I 2 x 500 n
60 V / / / /
^ j [ _
1 2 3 The mean value of the sending reference equivalent
Length of the cable 1 T
4 5 km
The specified mean values of the sending and receiving reference equivalents of a sample of telephone sets
Standard deviation at sending, 5 max = 1,3 dB Standard deviation at receiving, 5 max = 0,9 dB
750486
F i g . l .
1
SRE A (dB)
10 -
3. 2,5% 2. 34%
5 -
Standard deviation 5 = 1,3 dB
1 2 3 4 T
5 L/km
1 . no. of subscribers in 1 km region 2. no. of telephone sets exceeding 9,6 dB 3. no. of subscribers exceeding 9,6 dB
750487
Fig.2.
SRE dB
12 -
10 -
8 -
9,6 dB
1. 2 . 3.
8,5% 4,7% 65%
3,8% 7 , 3 ° / o ^^T 35% \^T 11% ^ f
1 1 1 3 4 5
5 = 1,3 dB
no. of subscribers in 1 km region no. of telephone sets exceeding 9,6 dB no. of subscribers exceeding 9,6 dB
3. 2. 1.
750488
Fig.3.
In Figure 2 a typical mean value curve of a telephone set is shown. It is supposed that the mean value at a distance of 5 km is exactly 9,6 dB. It is further supposed that the standard deviation is exactly the allowed 1,3dB. The upper end of the curve has been substituted by steps of 1 km. Over both steps there are three numbers. The numbers shown come from a large investigation made in Finland some years ago. The lowest number indicates the percentage of subscribers in the corresponding 1 km region.
The middle number indicates percentage of telephone sets exceeding 9 ,6dB (assuming normal distr ibution). The uppermost number indicates the percentage of the subscribers exceeding the allowed limit 9 ,6dB . When adding the last two numbers it is found that the limit of 9,6 dB is exceeded by about 3,8% of the subscribers.
In reality the situation is, however, not so simple. The mean value curve and standard deviation are varying from one manufacturing lot to the other. However, w i th this method we can roughly estimate the percentage of the subscribers who wi l l exceed the allowed l imit.
3. Quality control Helsinki Telephone Company buys
about 5 0 0 0 0 telephone sets per year. The telephone sets are delivered in lots of about 8 0 0 pieces. Our Quality Control performs a sampling inspection of the telephone sets, where the most important functions of the telephone sets are tested and the reference equivalents are measured objectively w i th measuring equipment designed according to K. Braun's principle. The method of inspection is mainly the inspection by attributes whereas the reference equivalent is tested "by variables". The reference equivalents are measured w i th cable lengths of zero, 2 and 5 km (diameter 0,5 mm). Mean value and standard deviation is calculated for each line length. However, in addition to this, if an individual reference equivalent differs more than 4 dB on sending and more than 3 dB on receiving from the mean value requirement, the telephone set is considered defective and the inspection by attributes is applied. These values are excluded in calculations of mean value and standard deviation.
14
4. The accuracy requirement of a measuring equipment
The influence of the accuracy of a measuring equipment can be examined according to the same method as above.
Let us assume that the measuring equipment reads incorrectly 1 dB in the direction that the real reference equivalent is 1 dB higher than the measured value. In Figure 3 the same situation as in
Figure 2 is shown, but now 1 dB higher. Again, it is easy to calculate from the reference equivalents and subscribers distribution given in Figure 3 that the percentage of the subscribers exceeding 9,6 dB is now 4 ,7 + 3,8 = 8,5%. If we further assume that the error of the measuring equipment is 2 d B , an exceeding percentage of 14% is obtained. In addition to this the subscribers located nearer than 3 km must now be taken into account.
In practice the mean value of the incoming telephone sets is varying between the allowed limits from one lot to another. However, at 5 km, the mean value is more often approaching the upper than lower-limit. It can thus be estimated that the exceeding percentage wi l l be moderate, if the inaccuracy is about 1 dB. An inaccuracy of 2 dB begins to be critical from the standpoint of ful f i lment of the transmission plan.
Let's compare the result w i th the measuring accuracy of the other parts of telephone connection. The attenuation is usually measured using level meters and signal generators. Nowadays, very accurate level meters and stable generators are available, but very often the accuracy of the measuring equipment in use is only about 0,5 dB. In addition to this we must notice other varying factors, for example on carrier frequency connections about 1 dB must be reserved for variation w i th t ime of the loss. Thus we can conclude that the other parts of the telephone connection cannot either be determined w i th greater accuracy than 1 dB.
5. Other factors affecting the accuracy of the measured results — the reproducibility of the results
As mentioned above, the mean value and standard deviation of the incoming lots of telephone sets are varying. This is partly caused by component parameters and produc-tional deviation. Total deviation is also increased by variations of environmental factors, especially temperature and humidity.
To enable the comparison of the results of the different lots w i th each other, the reproducibility should be as good as possible. To investigate the reproducibility of the
Fig.4.
Fig.5.
reference equivalent values measured objectively, we made the fo l lowing test. 10 telephone sets equipped w i th carbon microphone and 5 telephone sets equipped wi th linear microphone were chosen. The sending reference equivalent of
them was measured on 20 successive days and the mean value was calculated every day. The results can be seen in Figures 4 and 5. From these figures it is immediately noticed that the mean values for carbon microphones may vary widely.
15
Telephone sets equipped wi th l inear microphones vary considerably less. It is diff icult to say, how this variation should be divided between telephone sets and measuring equipment. However, it is evident that for stable telephone sets the reproduci-bility of the results is acceptable if the deviation is of the magnitude 0,5 dB.
it should be possible to achieve the same reproducibil ity w i th different measuring equipment. Our experience has shown that if the difference between the mean values obtained by the manufacturer and the customer is more than 1 dB, it w i l l be difficult to decide who has got the correct value. Wi thout any further examination it can be concluded that results obtained using different measuring equipment must be w i th in the accuracy l imits.
6. Practical quality control As we know that reproducibility
of the results of the carbon microphone telephone sets is bad, we cannot immediately return a lot of telephone sets, if the mean value slightly exceeds the permitted limit. Our Quality Control logs cont inuously the mean value and standard deviation of the incoming lots in the manner seen in Figures 6 and 7. Steps are taken against the supplier, if the mean value remains continuously outside the limit or if the mean value deviates considerably (1 dB) from the allowed limit.
The standard deviations (maximum 1,3dB on sending and 0 ,9dB on receiving) are also very good quality criteria. If the standard deviat ion suddenly increases, it is usually a sign of some production error, usually when assembling the carbon microphone or receiver.
It happens, however, sometimes that we have to discard the lot because of our inspection. Our inspection system enables us to prove to the manufacturer that the returning is just i f ied.
7. Conclusions I have tried to describe the accu
racy requirements of the objective reference equivalent measuring equipment implied by transmission plan and quality control. The conclusion is that the accuracy ought to be about 1 dB (maximum 1,5dB) regardless of the principle of the measuring equipment. A different subscriber distribution and different requirements for reference equivalents might lead to a different accuracy requirement, but I think that our final result is quite realistic as it is based on both examination above and practical experiences.
1974
date of insp.
Jan. February March
S R E a t 5 k m
Mean value and standard deviation
April May June July August September October
30 5 7 12 13 14 15 18 25 1 7 7 14 18 27 1 5 16 18 22 25 30 7 13 22 23 28 29 5 10 13 12 20 26 28 2 6 5 15 12 16 22 29 3 5 ^3 16 18 27 3 14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
The inspected tots
Fig.6.
dB
RRE at 9 km Mean value and standard deviation
750491
Fig.7.
16
Frequency response measurement on the telephone sets 73D, F68 and 47E connected via a telephone line by S. A. Jager and P. Schnack, Jydsk Telefon A / S .
en
en i -
n c
>
CO =3
<
o \— o
"O C H3
^ CO
1 —
> C/l t o 3 C"" L—
<
CD u , CD
" Q c CO
1-
CTR
N /
\ /
. /
Fig. 1. The measured chain 750160
Introduction From the State Audiology Centre
in Arhus, Denmark, the Danish telephone company J.T.A.S. received a request regarding the frequency response and sound pressure levels to be expected from normal telephone connections in Denmark. To answer this question experiments were made wi th commonly used telephone sets and an arbitrarily chosen trunk connection.
The Measuring Arrangement The trunk connection was set-up
over Tranebjerg — Arhus South — Skanderborg — Arhus South and back to Tranebjerg, involving five telephone exchanges. Measurements were made wi th the B & K equipment shown in Fig. 1.
Various separate calls were made using three different types of telephone set, and for each call frequency response curves were recorded. A comparison of the results shows that the connection itself was almost identical in response from call to call. The curves shown in this report may therefore be regarded as typical for the particular trunk connection chosen for the experiments.
Important Note At frequencies below 2 0 0 Hz the
response shown on the diagrams are not correct due to excessive distortion from the Art i f icial Voice signal.
Transmission Line Measurements Fig.2 shows the frequency re
sponse of the "pu re " telephone-connection Tranebjerg — Arhus South — Skanderborg — Arhus South — Tranebjerg, loaded by a 6 0 0 O resistance in each end. As the connection between Skanderborg and Arhus South is of the carrier frequency type bandwidth l imitations are present. These can be clearly seen from Figure 2 (bandwidth: 2 5 0 H z — 3400Hz) . The peaks and valleys occuring in the 1 0 0 0 H z to 3 4 0 0 Hz region indicate artificia (Pupin) loading of the line.
Measurements with Different Telephone Sets
In the fol lowing f igures, various frequency responses are shown, including sender (A) set and receiver (B) set connected via the above de
scribed line connection, see also Fig. 1. A brief outl ine of the characteristics of the telephone sets used in the experiments is given in Appendix A.
Typical response using the telephone set 73D as transmitter (A) as wel l as receiver (B) i. e. A : 7 3 D — B: 7 3 D is shown in Fig.3.
Note that over the complete frequency range 3 0 0 H z — 3 4 0 0 H z the received sound pressure level is of the order of 80 dB, wh ich , according to several investigations is the preferred mean sound pressure evel for normal connections.
Typical response using the telephone set F68 as transmitter (A)
] n □ Q a n a n n n a D D n n a D n D n n n D n a a a a D a n a □ D n n o n n n D a D a Q n a D n n c BKiei&KjEBr Briiel & Kjaer
Cop4nhiapan 50r25
dB Measuring Object:
Transmissions-linie
40
Tranbjerg-Ar, syd -Skanderborg 3 0
dB
■20
-15 BF t i l Ar. syd Tranbjerg
Rec No.: D a t e : _ 2 M 2 - 7 3 Sign.: ?S.,
20
RMS
-10
Rectifier.:. Zero Level: 10 m V . I Lower Lim. Freq. 20 Hz 5 Potentiometer:25 dB[_ Writing Speed:80 m/m s e k ~ Paper Speed:_10 m/m s e k "
Multiply Frequency L t ^ _ oL 0 5 =
1 0 V -Scale by- 1
4+r
I Oi75
d&
8 - 6 0
dB
6 -45
4 - 3 0
200 500 1000 2000 5000 10000
Q P 1 1 4 2 T50161
Fig.2. Characteristic frequency response of the plain transmission line
17
8L
8 9 d o\ i t * 9"6!d SSiOSl ZfrLL dD
0000 L
09 -8
s^kj
0009 oooz
: :
001 0 10 I ■ i "Aq eieos
>tas LU/UJ o i~;pe»ds j a d e j Has UJ/LU o g : p ^ d s S U S ^ M
~|8P S3^Joiaujoiiueiod q.*HOZb9ii rLun ««<n
■s-d "u6!S
01- 03
SL-OE
03-
8P
93 ^
89d — 3Z.t
Otr iDefqO Suunsee^j
SP H IN U
t a n n n n G n n n a n D D a n n n n D D n n n a n a n a a a a D D a n a a D D n n n n a a n D D n a JSef^ 19 | 9 n j a
Q£L o i 3Z.fr 9"6|d
9 £ K U
D a D a D n n a a D a a D D n n n n n n n a D a a n n n D a n D n a D a a a D a a a a a a a a n n D t l
S Z -* 0 S LI*. JB^U *dC3
aez_ 01 89=i "fr'Bu £91091
0000 L ro
3frU dD
0009 ' \' "Aq f>\*3$ AouenbeJj A|di»|nyy
>|&S a i / U J O L _ ; p o a d s J o d o d
>|3S UJ/LU 0 8 : P & e d S Su'lHM "lap 92" j^SLuo^uoi°d
^ z H 0 2 " b o j d LU ! l JO WOT ndsapgg^*^ ° i a
siAiy -^*!*3e
g*-Zl-OZ :eiB(
~ : O N 30]
Q£Z - 89 d
8P ^ ^
QZ.KH
t j a n a n n a n G D D D a n n u G a D G D n n D a D n a a a G a n n n a Q D n D n n n n n D D a a a c
Q £ £ 01 BujuiLUSUBJj. a e z . 9uoqda|sx '£"^!d Z9tO$£
0 0 0 0 1 ro
SI -3
oe - *
9t>
0 9 - 8
gp
000s =F 0003 000 L 009
- 9
8P
9 i k l L
J L 10 n o
"I Aouer
ZfrU dD 00 L
I :Aq Q|BOS lenbei j A|d<i|nftj
fas LU/UJ QL-:peeds jaded as UU/UJ Q8 :P s e dS 6 U I IMAA
"|gp g2:j&lsajo!lUB1°d 9 ZH 0 £ ' b e J d *"!"! JB^o i
" I d S 8P 99 : l0Aei OJOZ
~ ^ $ 3 — " u 6 | S
nL , "E*-ZL-oT~wa
QL-
03-
SP
oe aci - ati
ov : ioefqo Buunsea^j
92-*
gp
OS
*&rf*
JSe[>l 'S l 8 P j a jae!» * ia™a j n n a a a D n a n n D a a a a n n D n n a n n D D i n c n D a a D n D D n n n D n a a a a n Q n n D c
A|JOOC1 S|eu6fs Aouanbajj. M O | saAjao -aj 11 JEU,J pue ' jauj iusuBJi poo6 A|8A!ie|9j e SUJBIUOD 89=1 i9S auoqd - a p l BLJJ jeq i papnpuoo aq snqi Aew if a jnb j j 6ujo6ajo^ pue SILJI LUOJJ
"8P ZB J° uoiBaj aqj uj s| ZHOO^C 01 ZHOOOt a6uej Aouanb -aj^ UJ |3Aa| ajnssajd punos paAiao -aj aqj UJB68 ' jaAaMon zj-iOOOL °* Z H 0 0 £ 86UBJ Aouanbaj^ agj u ! 8 P 0 8 AAO|aq ieqAABiuos si |8A8| a jnssajd punos paAjaoaj aqj a jan
7_"6ij uj UMOtjs S| 3 9 J :g 89d : V 9 '! '(8) JQAjaoaj se ||3AA
se (v) J8UJLUSU9JJ se 89-j ias auoqd -a|aj a i j j 6ujsn asuodsaj |eoidAi
■;as 89d 8M1 UJ jaAiaoaj aq j p Ajueaun -uou Aouanbajj. a|qejapisuoo aq j pug jas 3/_-t? aqi ui j au j iusue j i /(>)BaMj(
jeqAAaiuos aq j p uoneujqujoo aqi 01 anp sj smj_ pauieiqo s\ aPS8 aixios p jaAa| a jnssajd punos paAiao -9J 8 ZHOOt^e 01 ZHOOOL s6uej aq i ui aj iqM 'gp g/_ p jap jo aqi p SJ |9A3| a jnssajd punos paA|ao -aj a iu 'ZHOOOI — Z H 0 0 £ ^^ue j Aouanbejj ei\\ u\ IBLJJ uaas SJ 1
•9-61-1 ui UAAOMS Sj 89d :a — 3Zlt7 : V a j '(g) jaAjaoaj se 89=1 J^s au,i pue
(V) jau iu jsue j j se ^p ias auoqd -a|ai aqj 6uisn asuodsaj leojdAi
"E 'Bld
si 3i_p jas auoLjdaiaj au,j u; j a i -lILUsueJi aqi se 'zHOOt^C — Z H 0 0 £ a6uej Aouanbaj^ aiajdiuoo au,i jaAO g p 0 8 ueMJ J9AAO| ;eu,Maiuos si |9Aa| a jnssajd punos paAiaoaj aqj_
■g-6!j u; UMoqs s; Q££_ :a — 3ZLfr ^V a 'j '(g) jaAjaoaj se Q£/_ jas aqj pue
(V) j au i i usue j ; se 3/_t7 l Q S ouou,d -a|ai aqi 6ujsn asuodsaj |60idAi
las Q£/_ au,i UBLJJ
j au ju j sue j ; ^ ja^eaM^ e seq 89d *8 S
auoqdajai aqj asneoaq si Sjq_L 'C'Bj-j UJ UMoqs ieq j ueq ; J3MO| jeqAAaiuos 'j3A3Moq 'SJ |aAa| s jnsss jd a6e -jaAe a q i Z H O O ^ C — zHOOC 96UBJ
Aouanbajj, aja|dujoo aqi JSAO g p 0 8 p jsp jo sq j p si |3A3| s jnssajd punos paAjaoaj aqi 'U IB6B ajaH
■t7'6|-j ui UMOL|S si a£Z. B — 89d : V a I '(g) jaAjaoaj SB Q£/_ jas aq ; pus
Typical response using the telephone set 73D as transmitter (A) and the set F68 as receiver (B), i. e. A: 7 3 D — B: F68 is shown in Fig.8.
Once again, the poor receive characteristics of the F68 apparatus results in a response with a relatively large frequency nonlinearity. In the region 300 Hz — 1000 Hz the received sound pressure level is of the order of 77 dB, while in region 1000 Hz to 3 4 0 0 Hz it is approximately 84dB. The overall received sound pressure level is, however, somewhat larger than that indicated in Figs.6 and 7 due to the fact that the transmitter in the 73D set is stronger than those used in the 47E as well as in the F68 sets.
Typical response using the telephone set 73D as transmitter (A) and the set 47E as receiver (B), i. e. A: 73D — B: 47E is shown in Fig.9.
The received sound pressure level is here somewhat above 80 dB over the complete frequency range 300 Hz to 3 4 0 0 Hz, as the set 73D contains a relatively strong transmitter, and the 47E set receives wel l . The receiver in the 47E apparatus does, although not too pronounced, show reception properties better in the frequency range above 1000 Hz than below 1000 Hz.
Typical response using the telephone set F68 as transmitter (A) and the set 47E as receiver (B), i. e. A: F68 — B: 47E is shown in Fig.10.
It is seen that the received sound pressure level is around 80 dB over the complete frequency range 300 Hz to 3400 Hz, although considerable frequency nonlinearity is present.
Typical response using the telephone set 47E as transmitter (A) as well as receiver (B), i .e. A: 47E — B: 47E is shown in Fig. 1 1.
Here the received sound pressure level in the frequency range 300 Hz — 1000 Hz is approximately 76 dB, while in the frequency range
] □ a u a o □ [ Bri ie l & Kjaer
Copenhagen c Q
• ^ dB
Measuring Object: — . A ("1
4 U
F68-F68 30
Rpr N n __ Jl IJ
Date :_ -20 - l 2 -73_ Sign.: P.S Rectifier.. RMS
■ i n D n a n a D D D n n D n n D D
Zb
"dB
-20
-15
-10
Zero L e v e l : 6 5 d B S P L . L o w e r U r n . F r e q . 2 0 H i 5 P o t e n t i o m e t e r : _ _ 2 5 d B
W r i t i n g S p e e d . 8 0 m / m St P a p e r S p e e d ; _ l 0 m / m se
1 j
Multiply Frequency L Scale by - 1 c\\ nl
OP 1142
T
^ 1
i
, .
a a a D [ Bri iel I
f
L j
■ I I J\
■k I / >k r^
1
[ I
1
j 1
1
u n □ Q i Kjaer
D n n n D n a a n D L
TT m m
" r t ' L 1
r 1 i I f 1
|
1 ■" F 1
■■ , ■ ■ ■ ■ ■ ■ ■ ■ ■ — ■ ■ ■ ■■■
■ m
J □ D □ □ D □ □ I
f i
|
[ f
■■ 1
1 I
1
\
1 I 1 1 [ _. . . .. . H ^
l | I 1 1
i= I
I
1
, - * - ^
| | [ ' | 1 ,—,
|
j _ , — i ■■ i
t 1
dB
8 -
1 ._..
[ E
1 L 1 ■
t T I
! i 1 1 j I I I
i o -
4 -
2-
n
/ 3
dB
60
45
30
15
100 Hz 200 500 1000 2000 L 5000 10000
750166
Fig.7. F 6 8 t o F 6 8
i n n n n n n D D D D D a n n D D D a D D a a n n D D n n n a n D a a D D D D D a D D n D D a n D D t Briiei &Kjt»r Briiel & Kjaer
copflnrmoBr 50 r25
dB
Measuring Object: 40
dB
■20
73D - F68 30-15
lee. No.: 20-1 D a t e : _ 2 0 : l ? : Z 3 _ [ Sign.: P S - ^ ^ectifier.:_RMS_ Zero L e v e l : 6 5 d B S P L . Lower Lim. Freq,20 Hz Potentiometer:-^ 25 dB Writing Speed: 80 m /m Paper Speed:- 1 0 m/m Multiply Frequency L Scale by.: . 1 Q L
K = T Oi75
II m
dB
8-
6 - 4 5
dB
60
4 - 3 0
- 1 5
100 200 500 1000 2000 5000 10000 0
QP 1142 750167
Fig.8. 73D to F68
3 □ □ □ D 0 D [ Bri ie l & Kjaer
ri " ■—■ PVafj n r i m n^tWi lrJ" ■ 'b
i . □ p4rini gtrn C n
^ ^ d B
M e a s u r i n g O b j e c t :
4 U
73D-47E ■ — ^ n
o u
R n r N n ^ n
Date: 20-12-73 Sig r . P .S .
^ e c t i f i e r . : _ R M S _
■ D D Q o n a n D n a D D D f :
25
dB
20
-15
-10
Zero L e v e l : 6 5 6 B S P L ^
L o w e r L i m . F r e q . 2 0 H z 5
P o t e n t i o m e t e r : - 2 5 d B L W r i t i n g S p e e d : 8 0 m / m 5E P a p e r S p e e d : _ J 0 m / m se
^^^m a^^ ^^ ^ ™ a^ ^^
_ _ L ^BB 1 1
M u l t m l v F r e n u e n r v 1 1
f
1 -^ - f
J I
!K ™ — p
,
— , — , , . 1 . . . .
■K: ^ 1
-1 1 E 1 / Sea e b y . : 1 nL n l ■ w -w -m— -m- j ' " ™ ■ ■■"■" %J "
QP 1142
11 30 F Jm t
— ■
: □ i : □ n n a n a n a a a n n n a a Bruel & Kjaer
1 r i T
— ^ —| |
— — - «
p - - " - T W - | ' V j i
1 ■ " t '
D D L
'
/ 1 1 /1 1 V i
i
— i — i — i
■
,
i
■
T I 1 j - - ■ . i.
.
i i - I \ 1 - —
/ ■■ '§]}lm ^f—
1
J 1
J 1 1 T
j D a a D n o n n n r
i i
~\ I I I I
I i
1 V J 1 — 1
1
1 1 1
1 I 1 LJ 1
. 1 1 F 1 I I j r \ \ r
'
dB
8-
1 |
1 1 | |
_ _
|
1 1 1
I i
6-
4 -
2-
i
n.
dB
60
45
30
15
\z** 200 500 1000 2000 5000 10000 v
750768
Fig.9. 73D to 47E
l a a a D D □ □ a n n n n n n a D D a Q D a D n a n a D n D n a a D n n D n a n a a D D D D D D D D r Bmei RKj r Briiel & KjaBr
Cop#Plhagan
Measuring Object:
F 6 8 - 4 7 E
Eec No.: a ie:^-20-12-73
Sign.: F \ S , _ Rectifier.: RMS' Zero Level: 65 dB SPL, Lower Lim. Freq. 20 H z ^ Potentiometer: 25 dB[_ Writing Speed:80 m/m sek Paper Speed:-10 m/m sek Multiply Frequency L Scale b y . : _ 1 . Q L 0
[QP 1142
Fig.10. F 6 8 t o 4 7 E
19
1000Hz to 3 4 0 0 Hz it is in the region of 82 dB. This is due to a combination of the 47E sets relatively weak transmitter and the tendency in its receiver to receive low frequencies insufficiently.
The major results of the described experiments are summarized in the table.
Conclusion By looking at the figures 3 to 1 1,
as well as the summarizing table, it seems that the best and most pleasant sound picture, wi th highintel l igi-bility, is obtained when the telephone set 73D is used both as transmitter (A) and receiver (B).
Furthermore, in all combinations where the set 73D is used as re-
Fig.11. 47E to 47E
3 □ □ D D D D [ Bruel & Kjaer
^r3> dB
Measuring Object: , . — A C \
4 7 £ - 47E 4 0
■-■% #"%
— — — 30
■
"dB
■20
-15
P^i ft r
Ht*r. N o - I Q I * r t D a t e : „ 2 0 - 1 2 - 7 3 _ * " w
^ign P-S. Rectifier.:, RMS. Zero Level: 65 dB SPL. Lower Lim. Freq. 20 Hz 5 Potentiometer: 25 dB_ ■ B. ■ .■*-& C J f l _ ' _ _ J B-M-TM. . H .
T
J ,
1 | — T j .* i — ■ — — ■ — ^ " ™ - ■
n
P ^^™
J ^ r -
— 1
Wr i t i ng £p* f tH" OU m / m StfK \ 4 A /
Paper Speed: 10 rn/m seK |— Multiply Frequency L 1 1
Rc^nlft hy 1 n! nl J ■ ■ \ f
Q P 1 1 4 2
TOO y
~m ~ i
3 □ t 1 a
™ | 1
|
| 1 _
1 . — f —
J r - ~ M 1 J
1
^_^__.—„—
□ n a n C D n a n n n n D n n n n a r Bruel & Kjaer
■ I ■ j '
1 1 -i h*~^ ^ . f- .
£ 1 _ F — ■
f |
1 H — 4 - H ^ ■—■
yf
|
1
-"
(
' * '
r r r ^ B
a .
L_
■ ■ ■ ■ m J 1
I 1 1
] D 0 D D D D D C
^ 1
f ! 1
T - * * " ~ i
f ! 1
1
^ p
1
1 1 J
1
■-■ 1 ■■ 1 rwmj
^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ™ T
1 _ ^
"~ '
[ 1 —
[ 1 I
J ^ ~ — -
| | 1 J \ !
■ ■ ■■—# h
1
T
j 1
| — |
1
r L
1 ' 1
\ "
* r ■ * 1 N f "*" 1 ■—1 ^ X T V J
I r I
rT"1 1
- W I H J 1 _
J 1 \ 1
1
J
F
1
_ ^
. f .
** V_I_M_^H
1
._._. ^
| | |
J L 1
1
1
1 u -
8-
/a
dB
60
6 -
4 -
2^
n
45
30
15
[ T ^ ^ O O 500 1000 2000 ' 5000 10000
75G170
ceiver (B), good results are achieved, and the sound pressure level vs. frequency characteristics is
reasonably constant around the preferred mean sound pressure level of 8 0 d B .
Transmitt. (A)
73D
F68
— ■ — —
47E
47E
1
i 1
F68
73D
73D
F68
47E
Receiver (B)
1 ■ - ■ " ^ ■ ■ n v
73D
! 1
1 1
1
i
73D
73D
F68
■ .
F68
F68
47E
47E
47E
Receiver Sound Pressure Level 3 0 0 H z — 1 0 0 0 Hz
81 dB SPL
80 dB SPL
i
78,5 dB SPL j 1
7 4 , 5 d B SPL 1 1
1
77 ,5 dB SPL
1
77,5 dB SPL
1 1
j 1 i
81,5 dB SPL
1 m_j _ 1 1
1 1 1 1
80 dB SPL
76 ,5 dB SPL
1 0 0 0 H z — 3 4 0 0 Hz
81 dB SPL
8 0 dB SPL I 1
■
78 ,5 dB SPL
8 2 dB SPL
8 2 dB SPL
8 4 dB SPL
1
i
8 5 d B S P L
1
1 T
[ 85 dB SPL 1 1 >
L 1
■J ■ ■ ■
1
i
1 81 r5 dB SPL
Remarks 1 1 h-rf 1 1 1 t „ J ■ ■ \ " i _ "
Fig.3 Both high frequency and low frequency ends of the passband are somewhat pronounced, resulting in a pleasant and highly intell igible sound picture.
Fig,4. As Fig. 3 but w i th a somewhat lower overall level.
Fig.5 As Fig,3, The overall level is, however, still lower than that obtained in Fig.4,
Fig^6. Fairly high attenuation in the low frequency region. This gives a sort of "meta l l ic" sound picture which is less pleasant for the listener.
Fig.7. As Fig.6. The attenuation in the low
i frequency region is less than in ! Fig.6, which results in a slightly
more pleasant sound picture. 1
Fig.8. As Fig,6, but w i th a higher average sound pressure level, result ing in a
; better intell igibil ity. i . „ . - . —_____ t ■ " ■ I - - - - - - . . . - _ . . . .-■■
I Fig.SL Fairly high overall sound pressure level and only slight attenuation in the low frequency region resulting in a pleasant sound picture and good intell igibil ity.
Fig.10. As Fig,9, the overall received sound pressure level isr however, slightly lower.
F ig .11 . As Fig.6, w i th less pronounced attenuation in the low frequency region.
Appendix A
Brief Classification of the Telephone Sets used in the Experiments
The Telephone Set 47E is a conventional set containing a carbon microphone and an electromagnetic receiver cartridge. It was introduced into J.T.A.S. in 1 947 .
The Telephone Set F68 is again a conventional set w i th carbon microphone and electromagnetic receiver. It is, however, supplied wi th a varistor in the microphone circuit to compensate for differences in feed-currents between different lengths of lines. This should al low for relatively constant sending (transmitting) level. The classification No. F68 referes to the year of introduction into Denmark, 1 9 6 8 .
The Telephone Set 7 3 D is a set all-electronic trans-contaming an
mission system, and electrodynamic microphone and receiver cartridges. The apparatus was ready for production in 1 973 .
20
What should be required of a telephone measuring equipment? production control experiences by O. Larsson
Requirement
Accuracy + 0,5 dB Easy to calibrate Easy to handle Fast measurements Insensitive to low frequency room noise
Purpose
Laboratory
X
Manufacturing
X X
X X
X
Table 1
1 . Introduction Normally we require that tele
phone transmission measuring equipments both for laboratory testing and production control shall measure characteristics close to true Reference Equivalents (RE). By the definit ion true RE's can only be measured subjectively (Recommendation P72 CCITT Green Book Vol. V). However, as loudness (in this case represented by RE's) is a main factor contributing to the transmission quality, we want to make a measurement that brings us as close as possible to the true RE's.
Other important transmission characteristics are for instance frequency responses, impedance and distort ion. At manufacturing it is both unpractical and uneconomic to measure all characteristics at the f i nal testing and we have to keep to the most important ones.
2. Measuring Equipment for Different Purposes
The requirements on laboratory equipment for objective measurements do not necessarily have to be the same as those for manufacturing control. In our opinion the differences can be described by table 1.
The specific requirements for manufacturing control caused us to develop a special transmission tester, 2 /LTM 3 5 7 5 0 0 , which is used in all LME factories w i th telephone instrument production. Sending, Receiving and Sidetone w i th and w i th out local cable are tested.
In the laboratory we use both Bruel & Kjaer 3 3 5 0 - 5 2 and 2 /LTM 3 5 7 5 0 0 .
When the microphones and receivers are manufactured as components, they are tested regarding-level and frequency response. Together w i th the f inal control of sending, receiving and sidetone on all telephone instruments we think we have a good check of the transmission quality. (Other characteristics are checked by random sample tests).
3. Transmission Testing Equipment 2 / L T M 3 5 7 5 0 0 (comparison with B & K 3 3 5 2 and NOSFER)
The 2 /LTM 357 5 0 0 Transmission Testing Equipment is described in LME/G 1 5 5 1 - 9 9 8 Ue (submitted to the participants in the meeting). In principle it works wi th whi te noise, speech weight ing network, artificial ear, artificial mouth, VU meter, amplifiers and attenuators.
As mentioned under 2 above we use both B & K 3 3 5 2 and 2 /LTM 357 5 0 0 in our laboratory work. The theoretics for loudness are not so wel l simulated in 2 /LTM 3 5 7 5 0 0 as in B & K 3 3 5 2 but
some confusing results have been obtained and are compared and commented below.
In table 2 are listed sending measuring results at 0 0 line for two types of telephones w i th carbon microphones.
As can be seen 2 /LTM 357 500 gives a good estimation of the true SRE's for both sets whi le the OREM-A equipment shows a difference of 2,2 dB for set A and 5,6 dB for set B showing that a correlation figure must be related to the specific telephone sets concerned.
Another example of sending measurements is shown in table 3.
Method
Estimation by 2/LTM 357 500 Subjective tests rel NOSFER
Estimation by OREM-A
Measurements of Sending Reference Equivalents Tel set A
+ 4,4
+ 4 , 0 ( 9 5 % ±0,5)
Diff 2,2
1 + 1,8
Tel set B
+ 4,5
+ 4 , 6 ( 9 5 % cont. int. ±0,4)
Diff 5,6 t
- 1 , 0
Table 2
Method
Estimation by 2/LTM 357 500
Subjective measurements rel NOSFER
Estimation by OREM-A
Measurements of Sending Reference Equivalents Tel set B
0 km + 4,5
Diff 0,1
+ 4,6
Diff 5,6
- 1 , 0
3 km 0,4 mm line + 12,1
Diff 1,1
+ 13,2
Diff—4,1
+ 9,1
Table 3
21
Here we can see that the difference between objective and subjective measurements is not the same for 3 km local line as a Okm. Again 2 /LTM 3 5 7 5 0 0 is closer than B & K 3 3 5 2 . This is confusing because the more accurate realisation of the loudness theory should in theory make the B & K 3 3 5 2 show more accurate values (except for a calibration constant). However, the t ime variation of the noise signal is closer to that of speech than the si-noidal, so we have a feeling that this can be an explanation. There might also be others.
4. The Customer -The Manufacturer
CCITT LAB
The discussion above does not prove that 2 /LTM 3 5 7 5 0 0 is a better laboratory equipment than B & K 3 3 5 2 . In fact we at LME use both. On the contrary it may happen that telephone instruments can be found, for wh ich B & K 3 3 5 2 gives the closest correlation to true SRE's. But the results imply that the importance of simulat ing the t ime-ampli tude characteristics of normal speech should be included in the study to make a good equipment for loudness rating measurements.
Very often the customer (the administration) requires tolerances of SRE and RRE of about ± 3 dB.
Manufacturer
In the CCITT documentation is shown that differences between laboratories for subjective measurements can be expected to be ± 3 d B . If we then add a transfer factor from subjective to objective values, we may land in a very bad accuracy. Therefore, when an agreement is reached on the acceptance of a certain type of telephone instrument (of course based on true reference equivalents) there must also fol low an agreement of objectively
measured values related to a specific measuring equipment. Otherwise it wi l l be almost impossible to keep the tolerances wi th in the ± 3 dB wanted by the acceptance control. Thus the acceptance control and the manufacturer shall preferably use the same type of testing equipment to minimize the measuring problems. It is not said that the best laboratory equipment is the best one for this purpose too. On the contrary the requirements for manufacturing is likely to agree w i th those for acceptance control for instance for economical reasons.
The fol lowing example shows the practical arrangements used by L. M. Ericsson and the Swedish Administration (Tvt). In the laboratory LME and Tvt use subjective measurements to determine loudness ratings and other transmission characteristics. The laboratories are also equipped wi th OBDM and 2 /LTM 357 5 0 0 for supplementary measurements. A t manufacturing and for acceptance control Transmission Tester 2 /LTM 3 5 7 5 0 0 is used both at LME and Tvt.
C 1 9 7 9 B & K Ins t ruments , I n c . , C l e v e l a n d , 0 H P r i n t e d in U . S . A .
X$8cJ\ Instruments, Ina Bruel & Kjaer Precision Instruments 5111 W. 164th Street, Cleveiand, OH 44142 / Phone: (216) 267-4800