AD9957SLASHPCBZ datasheet(23/61 Pages) AD

23 / 61 page

Data Sheet

AD9957

Rev. F | Page 23 of 61

Encoding and pulse shaping of symbols must be implemented

before the data is presented to the input of the AD9957. Data

delivered to the input of the AD9957 may be formatted as either

twos complement or offset binary (see the Data Format bit in

Table 13). In BFI mode, the bit sequence order can be set to

either MSB-first or LSB-first (via the Blackfin Bit Order bit).

INVERSE CCI FILTER

The inverse cascaded comb integrator (CCI) filter predistorts

the data, compensating for the slight attenuation gradient imposed

by the CCI filter (see the Programmable Interpolating Filter

section). Data entering the first half-band filter occupies a maxi-

mum bandwidth of ½ fIQ as defined by Nyquist (where fIQ is the

sample rate at the input of the first half-band filter); see Figure 33.

If the CCI filter is used, the in-band attenuation gradient can pose a

problem for applications requiring an extremely flat pass band.

For example, if the spectrum of the data supplied to the AD9957

occupies a significant portion of the ½ fDATA region, the higher

frequencies of the data spectrum are slightly more attenuated

than the lower frequencies (the worst-case overall droop from

f = 0 to ½ fDATA is <0.8 dB). The inverse CCI filter has a response

characteristic that is the inverse of the CCI filter response over

the ½ fIQ region.

INBAND

ATTENUATION

GRADIENT

CCI FILTER RESPONSE

½fIQ

fIQ

Figure 33. CCI Filter Response

The product of the two responses yields an extremely flat

pass band (±0.05 dB over the baseband Nyquist bandwidth)

eliminating the in-band attenuation gradient introduced by the

CCI filter. The cost is a slight attenuation of the input signal

(approximately 0.5 dB for a CCI interpolation rate of 2, and

0.8 dB for higher interpolation rates).

The inverse CCI filter can be bypassed using the appropriate bit

in the register map; it is automatically bypassed if the CCI inter-

polation rate is 1×. When bypassed, power to the stage turns off

to reduce power consumption.

FIXED INTERPOLATOR (4×)

This block is a fixed 4× rate interpolator, implemented as a

cascade of two half-band filters. Together, the sampling rate

of these two filters increases by a factor of four while preserving

the spectrum of the baseband signal applied at the input. Both

are linear phase filters; virtually no phase distortion is intro-

duced within their pass bands. Their combined insertion loss

is 0.01 dB, preserving the relative amplitude of the input signal.

The filters are designed to deliver a composite performance that

yields a usable pass band of 40% of the input sample rate. Within

that pass band, ripple does not exceed 0.002 dB peak-to-peak.

The stop band extends from 60% to 340% of the input sample

rate and offers a minimum of 85 dB attenuation. Figure 34 and

Figure 35 show the composite response of the two half-band filters.

–120

–110

–100

–90

–80

–70

–60

–50

–40

–30

–20

–10

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Figure 34. Half-Band 1 and Half-Band 2 Composite Response

(Frequency Scaled to Input Sample Rate of Half-Band 1)

–0.010

–0.008

0.008

–0.006

0.006

–0.004

0.004

–0.002

0.002

0.010

0.1

0.2

0.3

0.4

0.5

Figure 35. Composite Pass-Band Detail

(Frequency Scaled to Input Sample Rate of Half-Band 1)

In BFI mode, there are two additional half-band filters resident,

yielding a total fixed interpolation factor of 16×. The extra BFI

filters use the same filter tap coefficient values as the QDUC

half-band filters, but their data pathway is 16 bits (instead of

18 bits as with the QDUC half-band filters). As such, baseband

quantization noise is higher in BFI mode.

AD9957/PCBZ 数据表(PDF) 23 Page - Analog Devices

AD9957/PCBZ 数据表(HTML) 23 Page - Analog Devices

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61

数据表下载