NOVATECH INSTRUMENTS 7409B Manual
4.8 Example of Frequency Scaling. For an example of
scaling, suppose an external clock of 10.000 MHz is used
and an output of 1.544 MHz is desired. Also assume the
clock multiplier (Kp) is 15. Then:
(Fcommand) = (Fout ) (Kpi * Fint clk) / (Kpe*F ext clk)
(Fcommand) = (1.544) (429.4967296) / (15 * 10)
(Fcommand) = 4.4209530
To set output channel 0 to 1.544 send the command
“F0 4.4209539”. You should also provide an external
filter or you will see excessive distortion on you output
signal.
4.9 External Filters. The 409B default system clock has
a nominal value of 429.4967296 MHz. When using an
external clock, best performance is obtained when the
actual system clock (i.e. external clock frequency times
the cock multiplier setting) is close to or higher in fre-
quency than the default, since the 409B on-board filters
are optimized for the default. If the actual system clock
is lower than the default, then external filtering may be
needed to prevent output signal distortion. It is recom-
mended that the external filters have roll off frequencies
that are equal to or below 40% times the actual system
clock frequency.
4.10 Range Bit. It is possible to control the internal
range bit on the AD9959 DDS ASIC using the Kp com-
mand. For normal operation the Kp command is unmod-
ified. However, it may be desirable that the clock multi-
plier gain bit be set HIGH (for external clocks where the
system clock (Kp x External Clock Frequency) is from
255 to 500 MHz. To do this add hexadecimal 80 to the
Kp value to be set. For the bit to be forced LOW (for
external clocks where the system clock is 100 to 160
MHz), add hexadecimal 40 to the Kp value to be set.
4.11 Fractional Frequency Errors. When using the
default system clock there is no fractional frequency er-
ror. However, if you use an external clock that is not
identical to the internal clock then there will be a calcula-
tion round off error. The round off error as a fractional
frequency error (Δf/f) for output frequencies in the MHz
range will be less than 0.1ppm.
NOTE:
When using a 10 MHz external reference with Option /R
installed, there is no fractional frequency error.
4.12 Phase Alignment. Phase relationships are main-
tained by appropriate use of the “M” and “I” commands.
The “M” command has special modes “M a” and “M n”.
“M a” means automatically clear phase at the end of each
command. This will clear the phase register each time
any command is performed. This is important when all
outputs must be phase aligned. However, it will cause a
phase jump in the output.
4.13 Phase Synchronous. The “M n” command turns
off the automatic clearing of the phase register. This is
the default mode. In this mode, the phase register is left
intact when a command is performed. Use this mode if
you want frequency changes to remain phase synchro-
nous, with no phase discontinuities.
4.14 Command Execution. Further control of phase
relationships and timing of command execution can be
exercised by using the “I a”, “I m”, “I p” and “I e” com-
mands. The default mode is “I a” in which a command is
parsed and executed immediately following the end of
the serial input sequence. In the “I m” mode, an update
pulse will not be sent to the DDS chip until an “I p” com-
mand is sent. This is useful when it is important to
change all the outputs to new values simultaneously.
The “I e” command is used with the –AC Option.
4.15 Amplitude Matching. For applications which re-
quire precise amplitude matching between the channels,
the recommended method is to use the “Vn N” command
to adjust the channels to match. This command provides
10-bits of adjustment range.
4.16 Table Mode. The Model 409B contains on-board
static RAM capable of storing up to 32,768 profile
points. Each point contains phase, frequency, amplitude
and dwell time information. The on-board microcomput-
er reads this RAM and programs the AD9959 DDS ASIC
with the profile point data.
