REV. B

–12–

AD724

With the crystal selection circuit described above, the unse-

lected crystal and diode provide additional shunt capacitance

across the selected crystal. The evaluation board tested actually

required no additional capacitance in order to run at the

proper frequency for each video standard. However, depending

on the layout, some circuits might require a small capacitor

from FIN (Pin 3) to ground to operate with the chrominance

at the proper frequency.

SUBCARRIER FREQUENCY MEASUREMENT

It is extremely difficult to measure the oscillation frequency of

the AD724 when operating with a crystal. The only place where

a CW oscillation is present is at the FIN pin. However, probing

with any type of probe (even a low capacitance FET probe) at

this node will either kill the oscillation or change the frequency

of oscillation, so the unprobed oscillating frequency cannot be

discerned. Neither the composite video nor chroma signals have

the subcarrier represented in a CW fashion. (The LUMA signal

does not contain any of the subcarrier.) This makes it virtually

impossible to accurately measure the subcarrier frequency of

these signals with any oscilloscope technique.

Two methods have been found to accurately measure the sub-

carrier oscillating frequency. The first uses a spectrum analyzer

like the HP3585A that has an accurate frequency counter built

in. By looking at either the COMP or CHROMA output of the

AD724, a spectrum can be observed that displays the tone of

the subcarrier frequency as the largest lobe.

The CHROMA or COMP output of the AD724 should be

input into the spectrum analyzer either by means of a scope

probe into the 1 M

Ω input port or a 75 Ω cable that can be

directly terminated by the 75

Ω input termination selection of

the HP3585A. Each of these signals has present at least the

color burst signal on almost every line, which will be the domi-

nant tone in the frequency band near its nominal frequency.

Sidelobes will be observed on either side of the central lobe

spaced at 50 Hz (PAL) or 60 Hz (NTSC) intervals due to the

vertical scanning rate of the video signals. There will also be

sidelobes on either side at about 15.75 kHz intervals, but these

will not be observable with the span set to only a few kHz.

The center frequency of the spectrum analyzer should be set to

the subcarrier frequency of the standard that is to be observed.

The span should be set to 1 kHz–3 kHz and the resolution

bandwidth (RBW) set to between 10 Hz to 100 Hz. A combina-

tion of wider frequency span and narrower RBW will require a

long time for sweeping the entire range. Increasing the RBW

will speed up the sweep at the expense of widening the “humps”

in the subcarrier tone and the sideband tones.

Once the subcarrier is located, it can be moved to the center of

the display and the span can be narrowed to cover only that range

necessary to see it. The RBW can then be narrowed to produce

an acceptably fast sweep with good resolution.

The marker can now be placed at the location of the subcarrier

tone and the frequency counter turned on. The next scan across

the location of the marker will measure and display the subcarrier

frequency to better than 1 Hz resolution.

A second means for measuring the subcarrier frequency of an

AD724 operating from a crystal involves equipment more spe-

cialized than a spectrum analyzer. The technique requires a

Tektronix VM700A video system measurement instrument.

R1

10k

Y1

Y2

CR1

IN4148

CR2

IN4148

STND

AD724

FIN

HC04

U1

OPTIONAL

NOTES: Y1 = 3.579545MHz

Y2 = 4.433620MHz

R2

10k

PAL

NTSC

0-5pF

Figure 17. Crystal Selection Circuit

Pin 1 (STND) of the AD724 is used to program the internal

operation for either NTSC (HIGH) or PAL (LOW). For NTSC

operation in this application the HIGH signal is also used to

drive R1 and the input of inverter U1. This creates a LOW

signal at the output of U1.

The HIGH (+5 V) signal applied to R1 forward biases CR1 with

approximately 450

µA of current. This turns the diode “on” (low

impedance with a forward voltage of approximately 0.6 V) and

selects Y1 as the crystal to run the oscillator on the AD724. The

bias across the diode does not affect the operation of the

oscillator.

The LOW (0 V) output of the inverter U1 is applied to R2. This

creates a 0 V bias condition across CR2 because its cathode is

also at ground potential. This diode is now in the “off” (high

impedance) state, because it takes approximately 600 mV of

forward bias to turn a diode “on” to any significant degree. The

“off” condition of the diode does, however, look like a capacitor

of a few pF.

For PAL operation, the STND signal that drives Pin 1 is set LOW

(0 V). This programs the AD724 for PAL operation, deselects the

NTSC crystal (Y1), because CR1 has no bias voltage across it and

selects the PAL crystal (Y2) by forward biasing CR2.

In order to ensure that the circuits described above operate

under the same conditions with either crystal selected, it is im-

portant to use a logic signal from a CMOS type logic family

whose output swings fully from ground to +5 V when operating

on a +5 V supply. Other TTL type logic families don’t swing

this far and might cause problems as a result of variations in the

diode bias voltages between the two different crystal selection

modes.

FREQUENCY TUNING

A parallel resonant crystal, is the type required for the AD724

oscillator, will work at its operating frequency when it has a

specified capacitance in parallel with its terminals. For the

AD724 evaluation board, it was found that approximately 10 pF

was required across either the PAL or NTSC crystal for proper

tuning. The parallel capacitance specified for these crystals is

17 pF for the NTSC crystal and 20 pF for the PAL crystal. The

parasitic capacitance of the PC board, packaging and the internal

circuitry of the AD724 appear to be contributing 7 pF–10 pF in

shunt with the crystal. A direct measurement of this was not

made, but the value is inferred from the measured results.