Skip to main content

Full text of "DTIC ADA054704: Microelectronic Transceiver Development Program. Phase I."

See other formats


AO-A054 704 

unclassified 

I I 

AO 

AO&47D4 


ARINC RESEARCH CORP SANTA ANA CALIF WESTERN DIV F/G 9/5 

microelectronic transceiver development program, phase I.(U) 

AUG 68 H ROSENBERG. T D PRICE. R A MAMMANO N00123-68-C-2520 
467-01-1-904 Nt 







AO No. 

DOC FILE COPY AOA054704 



MICROELECTRONIC TRANSCEIVER DEVELOPMENT 
PROGRAM, PHASE I 


August 1968 


Prepared for 

U. S. NAVY ELECTRONICS LABORATORY CENTER 
San Diego, California 



RESEARCH CORPORATION 






UNCLASSIFIED 


SECURITY CLASSIFICATION OF THIS PAGF. Dan, Entfttyfli 


REPORT DOCUMEHTATION PAGf 

KKAD INSTRUCTIONS I 

BEFORF. COMPLETING I'OKM j 

1. REPORT number ^ 

U67-01-1-904 

2. GOVT ACCESSION NO. 

3. RECIPIENT'S CATALOG NUMBER 

4 . TITLK (and SubllUcJ 

MICROELECTRONIC TRANSCEIVER DEVELOR<ENT PROGRAM, 
PHASE I 

5. TYPE OF REPORT & PERIOD COVERED 



6. PERFORMING ORG. REPORT NUMBER 

467-01-1-904 

7. AUTHORfs; 

H. RosenLerg 

T.D. Price 

R.A. Memmano 

8. CONTRACT OR GRANT NUMBERfsJ 

Ku. 

N00123-68-C-2520 

9. PERFORMING organization NAME AND ADDRESS 

ARINC Research Corporation /' 

2551 Riva Road 

Annapolis, Maryland 2lh01 

10. PROGRAM ELEMENT. PROJECT. TASK 
AREA & WORK UNIT NUMBERS 

11. CONTROLLING OFFICE NAME AND ADDRESS 

U.S. NAVY ELECTRONICS LABORATORY 

CENTER 

tz. REPORT DATE 

August 1968 

San Diego, California 


13. NUMBER OF PAGES 

09 

14 . monitoring AGENCY NAME 4 ADDRESSfi/ ditlerent from Conlrotttne Office) 

U.S. NAVY ELECTRONICS LABORATORY CENTER 

San Diego, California 

IS. SECURITY CLASS, (of this report) 

UNCLASSIFIED 



15a. DECLASSIFICATION/ down GRADING 
SCHEDULE 

16. distribution statement (of this Report) 



UNCLASSIFIED AJNLIMITED 



17. distribution statement (of the abstract entered tn Btock 30, ft different from Report) 

18. SUPPLEMENTARY NOTES 

19. KEY WORDS (Continue on reverse side i/ necessary and /denf//y by block number) 

20. abstract (Continue on reverse s/de // nece^aary and Identify by btock number) 


DD 1 jam « 1473 -^niTioN OF I NOvesisoBsoLt Ti: UNCLASSIFIED 


SECURITY <L ASSIFiqAI ION OF THIS PAGE Dnlx Kiilrrxli 








Prepared for 



U. S. NAVY ELECTRONICS LABORATORY CENTER 



ARINC RESEARCH CORPORATION 
Western Division 
P. O. Box 1375 







CONTENTS 


1. INTRODUCTION 


2. LABORATORY EVALUATION 


LABORATORY DEVELOPMENT 


4. VOICE CODING TECHNIQUES 


5. VOICE SCRAMBLER DESIGN 


LIST OF ILLUSTRATIONS 


Title 


Digital Coding Technique 


Analog Time Permutation 


Frequency Permutation 


Block Diagram, Frequency Inverting Voice Scrambler 
Active Low-Pass Filter (One Section, 12 db/octave Rolloff) 
Double- Balanced Modulator Configurations 


Ac: «•' 

”• ' ■ 


H'l. 


CTi 

m 


0(1 O 

J-T 

j'jsu'iCA io>: 

□ 





1 BBTRIBaTIWIWMUBIllTI WSJ 

Dist. 

(AIL. arrf/oi 

bKLUIfU. 1 



467-01-1-904 


1. INTRODUCTION 


ARINC Research Corporation has completed Phase I of the Microelectronic 
Transceiver Development Program, performed under Contract N00123-68-C-2520 
for the U. S. Navy Electronics Laboratory Center, San Diego, California. This 
effort was conducted under the technical direction of Code S-240, and consisted of 
the following tasks ^ 

a. A complete evaluation of the equipment, materials, procedures, and 
goner£d capability of the NELC Hybrid Microelectronics Laboratory; 

b. Expansion of this capability to fulfill the requirements imposed by proto- 
type fabrication of the microelectronic transceiver hardware; 

c. A review of voice-coding techniques applicable toward providing secure 
communications, and the definition of an optimum approach for use with 
the transceiver; r\d 

d. The detailed design of the voice scrambler portion of the microelectronic 
transceiver. 

At the request of Code S-240, development of the remaining portions of the 
transceiver was delayed imtil completion of the voice scrambler section, since this 
represented the area of greatest risk in terms of compatibility with state-of-the-art 
electronics. 


tasks: 


The remaining (Phase II) effort of this program consists of the following 


a. Establishment of procedures and personnel training as required to fabri- 
cate prototype transceiver hardware in microelectronic form; 

b. Final development and evaluation of the scrambler portion of the 
transceiverj 

c. Design and development of the remaining transceiver circuitry. 



467-01-1-904 


2. LABORATORY EVALUATION 


At the beginning of this program, the Hybrid Microelectronic Laboratory 
contained most of the equipment necessary for thin-film deposition and hybrid compo 
nent assembly. Available items of equipment included; 

a. Lead bonding machine 

b. Die bonding machine 

c. Vacuum evaporator 

d. Film thickness monitor 

e. Film resistance monitor 

f. Photo resist system 

g. Mask alignment fixture 

h. Chemical sink 

i. Etching system 

j. Cleaning system 

k. Laminar flow hood 

l. Microprobe 

m . Ultrasonic cleaner 

n. Microscope 

o. Coordinatograph 

While this equipment provides a significant capability, the laboratory had 
several major operational deficiencies . Some of these deficiencies were the 
insufficiency of trained personnel to operate the equipment, the inadequacy of elec- 
tronic test instrumentation to support circuit development, the lack of documented 
sources of supply for supporting materials and chemicals, the lack of established 
procedures and man-machine-material interfaces , and the fact that some of the 
equipment had never been properly adjusted and calibrated. 

In addition to these problems, there were minor difficulties caused by 
inefficient placement of equipment, inadequate maintenance and repair of the 
facilities, and the fact that some of the general purpose equipment had not been 
optimized for the specific requirements of the transceiver circuitry. 


f 

\ * 467-01-1-904 

3. LABORATORY DEVELOPMENT 


During Phase I, ARINC Research activity within the Hybrid Microelectronic 
Laboratory was concentrated on getting all equipment, materials, and processes 
ready for initiating prototype work on the transceiver circuitry. As a result of this 
effort, the Lab is now able to process substrates with thin-film component mounting, 
conductor, and resistor patterns. These substrates are processed in accordance 
with a previously prepared laboratory flow chart, and represent state-of-the-art 
techniques in hybrid technology. 

As a first step in developing the laboratory capability, all equipment was 
thoroughly inspected, adjusted, repaired as necessary, and completely evaluated. 

As each item was put into working order, precautionary steps and operating proce- 
dures were established, and critic J parts were so labeled. 

With the intent of interfacing suitable processing operations at each station, 
the following conditions were investigated and set up as formal procedure: 

a. Lead tending: Bonding time, power settings, node pressures, gas flows, 
wire sizes and compositions, substrate temperatures, mechanical 
resonances, etc. 

b. Die bonder: Gas types and flow rates; temperatures of die, collet, 
substrate heater, and hydrogen heater; and the location of each with 
respect to the die used. 

c. Photolithography ; Resist composition, viscosity, and baking temperatures; 
developing materials and times; spinner speeds and times, etc. 

d. Evaporation: Fixture- evapo rant pressures, temperatures, volumes, 
etc. ; sensing procedures for film thickness and resistivity; substrate 

holding facilities. 

•• ••• •. 

e. Microprobing: Probe location, pressures, electrical wiring to test 
equipment, microscope positioning. 

f. Etchii^: Chemical solutions; etch times and temperatures; and cleaning 
procedures for all materials, including alloys, evaporants, and 
substrates . 

In addition to preparing the equipment for use, ARINC Research determined 
the most suitable sources for the materials and supplies necessary to make the lab 
functional in its output of hybrid microelectronic circuits. Items were ordered and 
inspected upon receipt to determine their ability to meet specified requirements. No 
formal specifications have been constructed for these materials; however, this is an 
objective of the next phase, time permitting. 

Personnel training was also initiated. Serving as a training aid was a substrate 
pattern useful for constructing a hybrid audio amplifier. This pattern was produced 
and applied successfully in the processing of substrates on glass and alumina. The 
initial film layer of the substrates was nichrome, overlaid with nickel and gold films. 


3 


467 - 01 - 1-904 



Photolithographic operations were performed utilizing KMER and related solutions. 
Once the pattern was etched chemically, die bonding and lead bonding were carried 
out to completion. 

Although these structures are very simple, they represent early efforts of 
lab personnel trained over a short time span only. Significant improvements toward 
more complex and useful circuitry are the primary objectives of ^e next phase. 


4 


467-01-1-904 


4. VOICE CODING TECHNIQUES 


Prior to the design of the microelectronic transceiver, a survey was imdertaken 
of applicable voice coding techniques. While extensive efforts over several decades 
have been devoted to the techniques and practices of securing voice commimications, 
no general-purpose approach has been established. In all probability, this is because 
no one technique satisfies all requirements for this type of device. Some of these 
requirements are listed below in what is considered to be their order of importance 
for the transceiver application: 

a. Small size 

b. Compatibility with existing equipment 

c. Intelligibility 

d. Security 

e. Low cost 

There are two basic approaches to voice encryption — analog and digital; and 
of course there are several possible versions of each approach. In general the 
greatest security is offered by utilizii^ digital techniques, but at the expense of size, 
cost, and transmission bandwidth. As a result, various digital approaches have 
been utilized almost exclusively to provide strategic security, while analog techniques 
are usually applied to tactical levels where size and cost are more important. 

To code a voice signal digitally, the usual approach is to sample repetitively 
the analog signal voltage, generate a binary word for each sample representing the 
average or peak voltage during the sample, code the digital words by the application 
of an arbitrary code key, and transmit the resultant coded digital bit stream. Upon 
receipt, the signal is decoded and converted back to an analog voltage. This process 
is illustrated in Figure 1. 

The fundamental reason why this approach could not be applied to the trans- 
ceiver program is that the bandwidth requirements are beyond the capability of 
existing equipment. For adequate intelligibility, the sampling rate must be at least 
twice the highest audio frequency desired, or about 6000 samples per second. A 
minimum of four bits are needed for reasonable accuracy in measuring the value of 
each sample, which means that the transmitted bandwidth must accomodate 24,000 bits 
per second — significantly above the 3-kilohertz maximum bandwidth of existing 
transceivers. 

Some bandwidth compression is possible with specialized digital techniques 
(i. e. , use of a vocoder), but at the expense of degraded quality and/or greatly 
increased complexity. It was therefore concluded that some form of analog tech- 
nique was indicated for this application. 

Analog voice coding, or scrambling, can be accomplished by two main 
techniques: time and frequency permutation. (Voice coding should be distinguished 
from mere masking of the signal with either noise or tones, which are later filtered 
out. The security afforded by masking was deemed insufficient for further 
consideration. ) 


5 


467-01-1-904 



Sampling i 


1 

1 

1 

i 

1 

Time 

Interval ' 

1 

1 


1 

1 

1 

1 

1 


1 

Digital Signal — 

1110 

1 

0010 1 

1 

1 

0110 !•••• 


plus 

+ 

+ 

+ 


Arbitrary Code Key — ►•••• 

0100 

1010 

1101 •••• 


equals 





Transmitted Signal — ►•••• 

1010 

1000 

1011 •••• 



Figure 1. Digital Codir^ Technique 


Time permutation is accomplished by breaking the audio signal into time 
increments and rearranging the increments accordii^ to some arbitrary code. A 
block diagram of this approach is shown in Figure 2, together with representative 
waveforms. 

As would be expected, frequency permutation is accomplished by dividing the 
voice spectrum into separate frequency bands, and through mixing and filtering, 
shifting the bands to other places within the audio spectrum. Representative wave- 
forms with this technique are shown in Figure 3 . 

For each of these techniques there is no theoretical limit to the number of codes 
that can be utilized. In practice, however, it is difficult to achieve more than 10 to 
20 unique codes . Thus , security is relatively low in both cases as compared to a 
digital approach. 

The main difficulty with time permutation is that at the frequencies of concern 
the only practical means for temporary storage is the use of a mechanical 
recording device. Since this is incompatible with the requirement for small size, 
frequency permutation was selected as the approach most suited for the 


6 


467-01-1-904 





Figure 2. Analog Time Permutation 


7 












467 - 01 - 1-904 


microelectronic transceiver application. This approach was further restricted to 
frequency inversion for the following reasons : 

a. It is the only approach that offers a high degree of confidence that the size 
limitations can be met. 

b. It represents a basic technique that can be expanded at a later date to 
provide a greater degree of security. 

c. The development effort could be performed within the allocated schedule. 


9 


467-01-1-904 


5. VOICE SCRAMBLER DESIGN 


Referring to Figure 4, the requirements for mechanizing a frequency-inverting 
voice scrambler can be broken into three basic functions: 

a. An input low- pass filter for limiting the input high-frequency response; 

b. An oscillator-modulator for sideband generation; 

c. An output low-pass filter for removing the upper sideband. 

The voice scrambler designed by ARINC Research performs these functions. It 
has unity gain, operates with about 1.0 Vac input, and consumes less than 100 mW 
power. The band of frequencies at the output (500-2500 Hz) is approximately equal 
to the band of frequencies at the input. These characteristics make this device fully 
compatible with the transceiver bandwidths in general, and the Motorola HT series 
specifically. 

Block A in Figure 4, an emitter follower with approximately imity gain, isolates 
the transceiver output from the scrambler input. Block B is an active low-pass 
filter with unity gain in the passband, a cutoff frequency of 2500 Hz, and a slope of 
48 db/octave. This filter is an optimally flat Butterworth type with four identical 
sections, one of which is shown in Figure 5. The poles of the filter must be 
equidistantly spaced on the perimeter of a imit semicircle plotted on the S-plane, 
which requires that each section have a different damping factor and thus different 
values of resistance and capacitance. The resulting cutoff characteristics are closer 
to theoretical with less stringent component tolerances, and eases the requirements 
imposed on the integrated version. 

Block C of Figure 4 is a passive network that adjusts the modulator input level 
to 0. 1 Vac. Block D is the double-balanced modulator, for which Figure 6 shows 
two versions being considered. Each version is transformerless and performs 
adequately in discrete form with hand-selected components. Selection of the final 
circuit will be based on the configuration best suited to integration. 

The circuit of Figure 6A is similar to a transformer-coupled modulator. 
Matched resistors Ra. Rb. l^c ^d replace the respective transformer windings, 
and if points 1 and 2 are driven equally but 180 degrees out of phase, the difference 
between points 3 and 4 is the double-balanced output. The performance of the circuit 
depends not only on the resistor and diode match but also on the common mode 
rejection of the difference amplifier (Ql and Q2). Considering these requirements, 
a practical integrated version is possible. 

In Figure 6B the local oscillator signal modulates Q5 and Q6 and hence the 
current to Q1-Q2 and Q3-Q4, dual difference amplifiers. The input signal is coupled 
equally to Ql and Q4, and by virtue of the inherent logarithmic transfer characteristic 
of the transistor, the two signals are intermodulated. Although this circuit has 
fewer components than that of Figure 6 A, component matching is a greater problem. 
The desired level of harmonic rejection depends heavily on the match of the transfer 
characteristics of Ql, Q2, Q3, and Q4, and to a *.^sser degree on Q5 and Q6. 
Extensive computerized analyses of matched logarithmic characteristics indicate 
that experimentation with an integrated version is warranted. 


10 


0 



Figure 4. Block Diagram, Frequency Inverting Voice Scrambler 




467-01-1-904 



1 


] 



Figure 5. Active Low- Pass Filter 
(One Section, 12 db/octave Rolloff) 


Block E is the local oscillator, a free- running multivibrator of conventional 
design. By eliminating the microminiature transformers in the modulator, and hence 
the unbalanced nonlinearities of the devices, a sine-wave oscillator to minimize the 
generation of harmonics is not needed. 


Block F is a linear amplifier of conventional design that compensates for other 
circuit losses so that an overall unity gain is achieved. Block G is similar to Block B, 
giving a 48 db/octave rolloff above 2500 Hz. Block H is identical to Block A and 
provides a low Impedance output to the transceiver. 


I