Under normal conditions HPWM would be run at
1.058MHz as this gives the best switching
performance, however, by selecting the 2.116MHz high
frequency mode and with a high quality, high speed,
output stage it is possible to take advantage of the
better definition available through the superior noise
shaper performance. Care has to be taken however, as
an indifferent output stage can cause the higher
switching speed to give a higher level of switching
distortion that can mask the noise shaper performance
improvements.
HPWM provides a lower level of overall RF noise in the
AM band with that noise concentrated on multiples of
the switching frequency. With very careful OP stage
design it is capable of the best measured dynamic
range.
RPWM is a proprietary PWM scheme that applies
different data to both sides of a conventional H bridge
output stage. The primary purpose of this is to
maximise the resolution of the internal noise shaper.
This doubles the noise shaper resolution. The other
advantage of RPWM is that the FET switching
frequency is halved. RPM therefore has two speeds of
529kHz and 1.058MHz. This lower speed does not
reduce the noise shaper performance.
The nature of RPWM also ensures that it is less
susceptible to correlated jitter on the master clock
resulting in a better signal to noise ratio with jittery
clocks.
The recommended mode of operation for new users is
HPWM with a modulation frequency of 1MHz.
Chip select
Within the SPI 8 bit control word is a soft code for the
device chip select. Each device can be programmed,
via external pins, to match to any one of four chip
addresses. In this way up to 4 ZXCW8100 devices can
be combined together to form a multi-channel system.
Mute
A hard mute facility is provided for the device when
used in direct drive mode. Active low, this control will
shut off the output drive. Once released, the output will
remain disabled for approximately 0.5 seconds.
Additionally this facility can be used to provide thermal
and current overload protection.
Mode
The direct drive ZXCW8100 device can be operated in
different modes. In the configuration provided, the
device can operate with or without dead time
depending on the selection of output drive
characteristics. The direct drive ZXCW8100 device can
be selected though external input pins to operate with
or without dead time. It is recommended to run the
device without dead time.
Full details about the use of these modes is available in
the associated Application document. Contact your
nearest Zetex office for full details.
Supplies
The device is provided with several power supply
connections. A nominal 3.3 volt supply is required with
the supply pairs being de-coupled separately and as
close to the device as possible.
NOVALOAD™
The NOVALOAD™ system provides a mechanism for
overload control. It can operate in several different
ways as it acts to back off the gain of the volume or bass
blocks. Once activated there are two modes of
operation available for the user to select. With the
NOVALOAD™ Mode register bit NOVLM=1any
overload in the volume circuit block will result in the
gain of the volume block being reduced and if any
overload takes place in the bass circuit block then the
volume is again reduced keeping the bass boost
unaffected. With the NOVLM = 0 any overload in the
volume block will result in the gain of the volume block
being reduced, however, if any overload is present due
to bass boost the bass boost is removed completely.
When no bass clipping has occurred for a period of
time greater than that set by the limiter release rate
register (but not greater than two times that period) the
bass boost is fully restored to its previous level as
determined by the bass gain register.
The rate at which the volume is reduced in response to
clipping can be programmed through the SPI interface.
The gain reduces by 0.5dB over a selected number of
word clock periods. Typically this will be set to 0.5dB in
4 word clocks. Once the overload condition is removed
then the gain is released to increase again. This is also
programmable as 0.5dB per selected number of word
clocks. Typically this will be set to 0.5dB per 16 word
clock periods.
The release rate coming out of NOVALOAD™ is critical
to a good sound. Typically 2 seconds is suggested for
pop and rock, 4 to 8 seconds for classical.
De-emphasis
De-emphasis is activated when older audio recordings
are used in the system. These will have used
pre-emphasis to achieve noise reduction. The
de-emphasis frequency response curve is selected
versus the system sampling rate. De-emphasis only
applies in single speed mode.
ZXCW8100S28
ISSUE 1 - JULY 2002
11