PDSP16116MC Zarlink Semiconductor, PDSP16116MC Datasheet - Page 9

no-image

PDSP16116MC

Manufacturer Part Number
PDSP16116MC
Description
16 by 16 Bit Complex Multiplier
Manufacturer
Zarlink Semiconductor
Datasheet
ii) WTB1:0 = 00 No shift applied giving a shifter output format:
The effective weighting of the shift bit is -2
iii) WTB1:0 = 01 Shift complex product one place to the right
giving a shifter output format:
The effective weighting of the sign bit is -2
iv) WTB1:0 = 10 Shift complex product two places to the right
giving a shifter output format:
The effective weighting of the sign bit is -2
Overflow
contain a 32 bit word, then an invalid result will be passed to
the output. An invalid output arising from this combination of
events will be flagged by the SFTA0 flag output. The SFTA0
Flag will go high if either the real or imaginary reslut is invalid.
Output Select
Mux, which is controlled via the OSEL inputs. These inputs
are not registered and hence allow the output combination to
be changed within each cycle. The full complex 64 bit result
from the multiplier may therefore be output within a single
cycle.
combinations as as summarised in Table 4.
(Where MSR and LSR are the most and least siginificant 16 bit
words of the Real Shifter output, MSI and LSI are the most and
least significant 16 bit words of the imaginary Shifter output).
extracting the full 32 bit result from the PDSP16116. The first
mode treats the two 16 bit outputs as real and imaginary ports
allowing the real and imaginary results to be output in two
halves on the real and imaginary output ports. The second
mode treats the two 16 bit outputs as one 32 bit output and
allows the real and imaginary results to be output as 32 bit
words.
Bit Number
Weighting
Bit Number
Weighting
Bit Number
Weighting
If the left shift option is selected and the Adder/Subtractor
The output from the Shifters is passed to the Output Select
The output select options allow two different modes for
The OSEL control selects four different output
OSEL1
0
0
1
1
31
31
31
S
S
S
30
30
30
2
2
2
0
1
2
29 28 27 26
2
29 28 27 26
29 28 27 26
2
2
Table 3 - Output Selection
0
1
1
2
2
2
OSEL0
0
1
0
1
0
2
1
2
2
2
3
2
1
2
2
2
4
3
2
25
2
25 24
2
4
3
PR
MSR
LSR
MSR
MSI
2
24
2
2
8
22
5
4
2
7
23
2
2
2
6
6
6
24
23
22
2
2
2
1
2
3
5
5
5
PI
MSI
LSI
LSR
LSI
.
.
.
25
24
23
2
2
2
4
4
4
25
26
24
2
2
2
3
3
3
27
26
25
2
2
2
2
2
2
28
27
26
2
2
2
1
1
1
29
28
27
2
2
2
0
0
0
30
29
28
PIN DESCRIPTIONS
XR, XI, YR, YI
from these ports on the rising edge of CLK. The data format
is Twos Complement Fractional, where the MSB (sign bit) is
bit 15. In normal mode the weighting of the MSB is -20 ie -1.
PR, PI
registers and passed to the PR and PI outputs on the rising
edge of CLK.
Fractional. The field of the internal result selected for output
via PR and PI is controlled by signals OSEL1:0 (see Table 4).
CLK
CEX, CEY
inputs enable the CLK signal to the X or Y input registers
allowing new data to be clocked into the Multiplier.
CONX, CONY
then the data in the associated input has its imaginary
component inverted (multiplied by -1), see Table 3. CONX
and CONY affect data input on the same clock rising edge.
ROUND
16 bits of the Adder/Subtractor result prior to being passed to
the output register. The rounding operation takes place one
cycle after the ROUND input is taken high. The ROUND input
is not latched and is intended to be tied high or low depending
upon the application.
MBFP
is selected. This allows the device to maintain the dynamic
range of the data using a series of word tags. This is especially
useful in FFT appllications. When low, the chip operates in
normal mode for more general applications. This pin is
intended to be tied high or low, depending on application.
Data inputs 16 bits: Data is loaded into the input registers
Data outputs 16 bits: Data is clocked into the output
Common Clock to all internal register.
Clock enables for X and Y input ports: When low these
If either of these inputs are high on the rising edge of CLK,
The ROUND control is used to round the most siginficant
Mode select: When high, Block Floating Point (BFP) mode
The data format is Twos Complement
PDSP16116/A/MC
9

Related parts for PDSP16116MC