Skip to main content

Full text of "NASA Technical Reports Server (NTRS) 19790023016: Digital phase-locked loop development and application to Loran-C"

See other formats

General Disclaimer 

One or more of the Following Statements may affect this Document 

• This document has been reproduced from the best copy furnished by the 
organizational source. It is being released in the interest of making available as 
much information as possible. 

• This document may contain data, which exceeds the sheet parameters. It was 
furnished in this condition by the organizational source and is the best copy 

• This document may contain tone-on-tone or color graphs, charts and/or pictures, 
which have been reproduced in black and white. 

• This document is paginated as submitted by the original source. 

• Portions of this document are not fully legible due to the historical nature of some 
of the material. However, it is the best reproduction available from the original 

Produced by the NASA Center for Aerospace Information (CASI) 




A digitol ohase-locked loop has been developed 
and implemented for use in a low-cost Loron-C 
receiver. This paper documents the DPLL design 
and application to Loran-C. 


Daryl L. McCall 

Avionics Engineering Center 
Department of Electrical Engineering 
Ohio Universih' 

Athens, Ohio 45701 

(NXSA-(;R-162283) digital phase-locked loop N79-31187 

Dniv,) 14 i> HC A02/*1P AO 1 CSCL 17G 

'Incl as 

G3/04 35726 

September 1979 

Supported by 

National Aeronoutics and Space Administration 
Langley Research Cente' 

Langley Field, Virginia 
Grant NGR 36-009-017 


A digital tracking filter for o Loran-C envelope manipulation receiver is presented, 

A first-order, increment or decrement, CMOS, counting-type of circuit is arranged to generate 
an 8-pulse replica of the average phase of the Loran-C signal compared to a deloy and add 
type of envelope detector output, A common offset cloi ’< oscillator of 1 MHz drives three 
similar loops locked to a master and two slave stations, A slow-fast circuit allows manual 
or computer-controlled positioning of the loops for Initial signal acquisition. Tracking 
precision is limited by the RF front-end envelope detector used to about ^ 1 microsecond 
for S/N ratios above 0 dB in a 20 KHz Input bandwidth, 


A. Implementation of the 4029 and 4013 , The basic loop Is a simple circuit (see 
Figure 1), the 4013 (type D flip-flop) functions as a phase detector and the 4029 (pre- 
settable, up/down, BCD/binary counter) as an increment counter or decrement counter, 
respectively, depending If the 1 MHz clock Is offset negative or positive. 


100 KHz O 



— 1 

L 4029 

Q4 Ci 


1 MHz 

Figure 1. Basic Digital Phase-Locked Loop. 

The present enable, soii’retimes labeled as the parallel load input, can be used to add or 
eliminate pulses of the offset 1 MHz clock until the offset is eliminated and exactly 1 MHz 
is obtained and counted by the 4029, As o result, the output Q4 is exactly 100 KHz and 
is m phase with the desired 100 KHz because the 4013 only generates a correction pulse 
when the two frequencies are in phase. The circuit of Figure 1 will track input frequencies 
on either side of the reference clock depending on the programming of the 4029, 


A. Working with the 4029 . To understand how the 4029 is capable of deleting or 
adding pulses to correct for the offset of the 1 MHz clock, one must study the Internal 
logic of the integrated circuit (see Figure 2). 


Pi. (Pa>«li«i Lo«(J Input) - A«vnchronout<v Lo«di P >niu Q. gli Othur Inpuii 

P {Puiallgl Input) - Dma on tht» P.n li A*vnchronou»ty LocdAcJ mto O, when FL t« tOW Owflfriiiino all Oth®f Inputs 
{Toggle Input) Forcii thu Q Output to SvnchiOOOuily Toggle v. hgn a LOW I| Plnc«P on this Input- 

CP ICIock Putff Input! 

U, Q I Tru* and Compinttontat y Ou' uuli) 

Figure 2. 

Logic Diagram of a 4029 

The specific poinf of inferesf is the fjnetion of the three input NAND gate 
driven by the inverted inputs labeled PL (preset enable or parallel load input), CP 
(clock pulse), and CE (clock enable). Because the preset enable, clock pulse, and 
clock enable are nanded together, the preset enable can function as a clock enable 
as well as a preset enable when the clock pulse and clock enable are driven by the 
same source. It is by operating in this manner that pulses can be added ar deleted, 

A pulse can be injected if in the course of time the clock and clock enable 
are low and the preset enable goes from a low to a high, then to a low again. The 
output of the NAND gate produces a complement of the PL, this in turn presets the 
programmed number of the counter and advances the counter by one count, (See 
Figure 3.) 

Similarly, a pulse can be deleted if,durlng the time the PL is high, the clock 
pulse and clock enable go from a high to □ low and then to a high again. The deletion 
occurs due to the fact that the internal flip-flops are continuously preset during the 
changes from their programmed number and, therefore, the incoming clock pulse is 
effectively ignored, (See Figure 4,) Because of the logic involved with the preset 
enable, clock pulse, and clock enable, the designer has the option to develop and 
use several variations. Some workable variations will be discussed In Section IV. 

- 2 - 

Clock Pulse 

Clock Pulse 

Clock Enable 

Preset- Enable 

NAND Output 

Clock Enable 

Preset Enable 

NAND output 

Figures, Pulse Injection, 

Figure 4. Pulse Deletion 

B, Obtaining More Control with the 4013. Because the length of the output of 
the 4013 is very small (i.e,, only one propagation delay), the preset enable of the 4029 
Can have problems recognizing a correction pulse. In order to lengthen and significantly 
improve the square shape of the correction pulse, a network of a resistor, diode, and 
capacitor Is added to the output of the 4013 The network allows the 4013 to function 
very much like a monostable (see Figure 5), This simple addition of circuitry gives a 
high degree of control over the deletion and injection of pulses and also creates a 
pulse the 4029 can easily recognize. 

Desired 100 KHz 
Derived 100 KHz 


Figure 5. 4013 with Monostable Output, 

If the deletion of pulses is desired, then the correction pulse should be slightly 
greater than the period of the 1 MHz clock. In this case, the correction pulse should 
be greater than one microsecond in widf}i. 

If the injection of pulses is desired, then the correction pulse should be smaller 
than one-half microsecond so the chances of Injecting a pulse are greater. The length 
of the pulse must not be made too small, otherwise problems will arise with the 4029. 

A correction pulse that is too short would nor allow enough charge or discharge time 


for Hie reset capacitor. The pulse width is also limited by the relatively slow speed 
of the CMOS 4029 logic. 


A, Using the Clock Enable . One of the most obvious variations would be to 
ground the preset enable, feed tlie correction pulses into the clock enable, and feed 
the 1 MHz clock into the clock pulse input. This appears to be the most logical 
approach, in that pulses are Injected or deleted by enabling and disabling the clock 
input. However, as stated before, upon studying the internal logic of the 4029 one 
finds that the preset enable plays just as important a part in enabling the clock as 

the clock enable. The only difference In operating with the clock enable is that the 

counter continuously counts up or down depending upon the programming of the UP /DOWN 
input. Figure 6 works under the principle of pulse injection and, as a result, has the 
1 MHz clock offset one to five cycles low. 

Phase Detector 

Offset 1 
1 MHz 

B. 1 MHz Clock Offset. Clock offset is another point of consideration. Is 
it better to offset the clock high or low? The answer con be found only by considering 
the application for which the loop is to be used and experimenting with the loop until 
the proper design has been found. Parameters t hg t may be varied in conjunction with 
the offset include the programming of the UP/DOWN input, programming of the parallel 
data inputs, use of the monostable network (which is highly advisable in any case), and 
varying the RC parameters, or pulsewidth. 


^ Wou 














1 Vf,f, 





3 ” 

4029 4510 

Figure 7. Pin Diagram of 4029 and 4510 1-^^. 

_ 4 _ 

C, Using Hie 4510. Another option is to use a different counter chip. The 
4510 is cin excellent chip to use to replace the 4029 due to the fact that pin-for-pin 
the two chips are almost exactly the some. The differenc e lies in the fact that the 
4510 has a master reset in place of the BINA RY/DE CADF input. Fortunately, If the 
identical programming for the 4029 is used for the 4510, the 4510 will function exactly 
as the 4029. The loop might function slightly better because the 4510 is a newer chip 
than the 4029 with buffered outputs. Using the 4510 also offers another option for an 
input for the correction pulses. It appears to be possible to use the MR (master reset) 
to receive the correction pulses as the MR also enables and disables both the clock 
and the preset enable, thus the MR also has control of injecting or deleting pulses. 


One variation of this digital phase locked-loop has been applied to developing 
a Loron-C prototype receiver at Ohio University. The loop is designed to track a 
synthesized eight pulse train to a train of eight IRQ pulses. The IRQ pulses are 
derived from the actual Loran signal at the receiver's RF front end'-'^-*. 

The Loran-C DPLL can be broken down into five interrelated systems. In 
addition to the phose detector and counter of the actual loop, a GRI generator, eight 
pulse generator, and Integrator are required. 


A. GRI Generator The Group Repetition Interval (GRI) generator 
consists of four 4029's cascaded together (see Figure 8). The 4029's ore programmed 
to produce a single pulse (seven to eight microseconds in length) with the same period 
as the desired Loran chain. 

The programming of the first stage of the GRI generator allows the operator of 
the receiver to speed up or slow down the loops by dividing the incoming 100 KHz by 
either 9 (to speed up), 10 (to cause no change), or 11 (to slow down). This allows 
positioning of each loop to agree with the incoming Loran signal. The remaining three 
stages are programmed to count down from the Loran GRI number set at the GRI switch. 
If the switch Is set for 996, the final three stages will count down from 996 using the 
10 KHz signal and effectively count the desired period of 99600 microseconds. 

The Terminal Count (labeled Co in Figure 8) of the MSD of the GRI generator 
is fed into the tame type of monostable network used with the phase detector of the 
original DPLL. The network is used to obtain control over the length of the GRI 
pulse for external microprocessor applications. The pulse from the monostable is 
then inverted and fed to the preset enables of the 40?9*s controlled by the BCD 
switches. The pulse is then inverted again, so that a short time delay exists between 
the GRI pulse and the preset enable pulse, and is fed into the eight pulse generator. 

- 5 - 

B. Eight Pulse Generot’or. The GRI pulse, fed to the clock input of the 4013, 
li used to activote the eight pulse generator (see Figure 9), When the flip-flop is set, 
the 4001 is enabled and allows the I KHz signal (generated by the second stage of the 
GRI generator) to pass through the 4001 and at the same time be counted by the 4017, 
The 4017 counts up to eight and resets the entire eight pulse generator. As a result, 
the system feeds a synthesized eight pulse train to the phase detector everytime the 
generator receives a GRI pulse. Thf eight pulse train will be delayed, after the GRI 
pulse, 100 microseconds times the LSD of the GRI switch. For example, if the GRI 
switch is set for 993, the eight pulse train will occur 300 microseconds after the GRI 
pulse. This is due to the GRI pulse resetting the GRI generator. 

Figure 9. Eight Pulse Generator. 

It is very Important to note that if the output of the 4001 (the eight pulse 
generator output) is viewed with an oscilloscope, the train will appear to have only 
seven pulses. The eighth pulse does exist, but it is only one to two gate propagation 
delays long. The short pulse is a result of the output of the 4017 directly resetting 
the eight pulse generator. The short eighth pulse does not create any problems because 
the clock output of the 4013 (phase detector) is positive edge triggered and effectively 
ignores the length of the pulse. 

C. Correction Pulse Integrator . Using the DPLL to track desired frequencies 
involved phase comparison and correction of every pulse for every time period. Only 
one correction pulse could be received for every period of the desired frequency. In 
dealing with Loran, data is derived from the phase of two eight pulse trains every GRI 
period. As o result, one could receive as many as eight corrections over the desired 
GRI period. Since eight corrections every GRI could be too much, and indeed is too 
much in practice, on Integrator is implemented. A 4024 counter, fed directly by 
the phase detector and driving the monostable network for the phase counter, Is used 
as an integrator (see Figure 10). 

By experiment. It has been found that an averaging number (or divide-by number) 
of 64 works quite well. This allows a maximum of one correction pulse for every eight 
GRI periods. 

- 7 - 


Figure 10, Bosic DPLL with Integrator. 

The monostable network was added to the output of the 4024 because the PL 
input of the 4029 phase counter is pulse width sensitive. The network was eliminated 
from the output of the detector because the input of the 4024 is positive edge triggered. 

D. Comments on the Whole System. The Loran DPLL was designed to operate 
with a 1 MHz clock offset four to five cycles high. The positive offset allows the loop 
to lock up as early as possible on the signol and, as a result, closer to the desired third 
cycle of the Loran pulse 

The lock indicator circuit, fed by any of the outputs of the 4024 (see Figure 1 1), 
can be used ta indicate the 1 MHz offset. The recommended output of the 4024 is Q5 
(the sixth output stage of the counter) and the recommended offset is approximately 5 

The integrator (4024) gives a correction pulse for every 64 phase counts; there- 
fore, the LED will flash twice for every correction pulse. In other words, the averaging 
number for the LED is half the averaging number for the correction pulse. If the clock 
is offset five cycles high, the nominal 100 KHz will be exactly 100 KHz plus j cycle. 
This comes to five extra cycles every ten seconds or five correction pulses every ten 
seconds. Because the averaging number provides two flashes for every correction, there 
will be approximately one flash per second for an offset of five cycles. A similar method 
could be implemented for determining the offset of a 1 MHz clock if offset negatively. 


The new DPLL has eliminated many of the problems associated with the VCXO 
type analog loops. Some of the advantages as applied to Loran-C are: 

- 8 - 




(1) A simple clrcuU that requires only nine chips. 

(2) The DPLL boards require no adjustments. 

(3) The three orv-board 100 KHz oscillators and the components and 
adjustments required with them were eliminated, 

(4) One central 1 MHz oscillator provides for a common drift rate among the 
three DPLL boards. 

(5) The new DPLL provides for possible auto-syr.c controls. 

(a) up/down of the 4029. 

(b) Monitoring the outputs of the 4024 (integrator), 

(6) The DPLL is capable of locking up on low-levol signals and remaining 
locked on standard level signals during periods of severe skywave and spheric interference, 

(7) The basic circuit of Figure 1 has application to other types of digital 
phase-locked loops with longer counting chains capable of tracking offsets as high as 
- 5% of the input signal frequency, A very similar method was applied to one of the 
early experimental On.ega receivers developed at Ohio University 

For Loran-C, further improvements could be realized with the addition of a 
temperature compensated oscillator to use as the offset 1 MHz source. 

The Loran-C DPLL described by this report is currently used in a Loran-C 
protoh/pe receiver at Ohio University. Reports of the receiver's performances will 
be published later as test flights and data collection allow. 


The DPLL is a continuation of the development of a low-cost, general aviation, 
navigation :eceiver funded by the NASA Tri-University Program. The tutoring by 
Edwin Jones in the initial phasesof the DPLL'i design is greatly appreciated. Special 
thanks goes to Ralph W. Burhans, who has been an endless source of information, 
advice, encouragement, and without whom the development of the DPLL would not 
have been possible. 


Cl] "CMOS Data Book", Fairchild Camera and Instrument Corp. , Mountain View, 
California, 1977. 

[2] Lancaster, Don, "CMOS Cookbook", Howard W. Sams and Company, Inc., 
Indianapolis, Indiana, 1977. 

- 10 - 

[ 3 ] 

Burlians, R. W., "A Low-Cosi- Loran-C Envelope Processor (The MtnI-L 
Loran-C Receiver)", NASA TM-57, Avionics Engineering Confer, 
Doporfmenf of Elecfricol Engineering, Ohio University, Athens, Ohio, 

April 1978. 

[ 4 ] Buihons, R, W., "Phase- Locked Tracking Loops for Loran-C", NASA TM- 
60, Avionics Engineering Center, Department of Electrical Engineering, 
Ohio University, Athens, Ohio, August 1978. 

[5] Burhans, R, W., "Narrow Band Binary Phase Locked Loops", NASA TM-25, 
Avionics Engineering Center, Department of Electrical Engineering, Ohio 
University, Athens, Ohio, April 1976. 

- 11 - 


12 - 


Figure A-1. Mini-L DPLL Board without Monostoble Outputs.