欢迎访问ic37.com |
会员登录 免费注册
发布采购

PDSP16510AMAGCPR 参数 Datasheet PDF下载

PDSP16510AMAGCPR图片预览
型号: PDSP16510AMAGCPR
PDF下载: 下载PDF文件 查看货源
内容描述: 单机FFT处理器 [Stand Alone FFT Processor]
分类和应用: 外围集成电路时钟
文件页数/大小: 23 页 / 97 K
品牌: MITEL [ MITEL NETWORKS CORPORATION ]
 浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第12页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第13页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第14页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第15页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第17页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第18页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第19页浏览型号PDSP16510AMAGCPR的Datasheet PDF文件第20页  
PDSP16510A MA  
NnS > 2 X [nS + PK + D] for 50% overlapping  
NnS > 4 X [nS + PK + D] for 75% overlapping  
OPERATING MODES  
The operating mode of the PDSP16510 is determined by  
the condition of 16 bits in an internal Control Register. The  
status of these bits is defined by the inputs present on the  
AUX15:0pinswhentheDEFinputisactive.TheDEFinputcan  
be a simple power on reset if the operating mode is fixed once  
power is supplied. The AUX pins are also used to provide the  
imaginary component of the complex input data. Thus, if  
complex inputs are needed, the mode definition must be  
implemented through a tri-state buffer which is only enabled  
when DEF is active. The imaginary input data must be  
disabled during this time.  
N is the number of devices, n is the transform size, S is the DIS  
strobe period, P is the number of system clock periods given  
inTable4, Kisthesystemclockperiod, andDisthetotaldump  
time including 4 extra DOS periods as discussed previously.  
The DIS and DOS periods are any value defined by the user,  
down to the system clock period with the A grade part.  
In this mode increasing the output clock frequency will  
allow a greater continuous input rate. The provision of  
separateDISandDOS pinsallowsthistobemechanized,and  
the DOS frequency can be increased to that of the system  
clock used internally. When the sum of the dump time (  
including four extra DOS periods for output priming ) plus 12  
system clock periods (the transform time variation caused by  
input synchronization) is less than the load time, one device  
will be guaranteed to have finished dumping before the next  
one starts. The inverted DAV to DEN connection between  
devices is then not needed, and all DEN inputs can be  
grounded.  
Table 6 lists the functionality of each of the bits in the  
mode control register, and further explanations are as fol-  
lows:-  
BITS 2:0  
These bits define one of 7 options for the sample size and  
type of data. In the 1024 point options the device will assume  
the non concurrent operating mode, regardless of whether a  
single or multiple device system is specified. The internal  
control logic will then ensure that data is loaded, transformed,  
and dumped in sequential operations.  
For other data set sizes, loading, transforming, and  
dumping, can all occur simultaneously with a single device;  
the actual overlap will be dependent on the relative occur-  
rences of the INEN input. Only in Mode 1 can concurrent  
operations be done with multiple devices.  
The LFLG transitions occur at the same times as Mode 1,  
except that the double transition does not occur with multiple  
concurrent transforms. Fig. 10 illustrates a timing sequence  
with three devices. Real transforms still only use the real  
inputs regardless of the amount of block overlapping.  
MODE 3 (BITS 10:9 = 11)  
MultipledeviceMode3isprovidedinordertoimprovethe  
performancewhenblockoverlappingisneeded,andseparate  
output processors are provided. In this mode transfers in and  
out of the device are never concurrent with transform opera-  
tions. The device will actually load extra data such that the  
required data to perform two overlapped transforms is stored  
internally. TheamountofinternalRAMprohibitstheuseofthis  
mode when performing overlapped 1024 point transforms.  
LFLG will go in-active after a normal data block have been  
loaded, regardless of the overlap selected. The device, how-  
ever, continues to load more data. Thus, for example, in the 4  
x 64 mode, five 64 point blocks will be loaded. This technique  
allows each device in the system to complete two or four  
overlapped transforms (depending on the amount of overlap)  
before any new data is needed. When doing a straightforward  
256 point transform the device will load 256 + 128 data points.  
The full benefits are only obtained if more than one output  
processor is provided, but an extra processor is not always  
necessary for every device. Sampling rates up to the system  
clock rate are possible. The equations defining the sampling  
rates become:  
BIT 3  
This bit determines the number of right shifts built into the  
data path. In either condition only two right shifts occur during  
thefirstpass.Ifthebitisreset,threeshiftsoccurinsubsequent  
passesandtheblockfloatingpointschemeallowsuptofifteen  
compensating left shifts. If it is set, two shifts occur in every  
pass and overflow is possible. This is indicated by reducing  
the number of compensating left shifts to fourteen, and using  
scale tag value fifteen to indicate that overflow has occurred.  
BITS 5:4  
These bits define the choice of window operator. If other  
windows are needed they must be applied externally. The  
fourth option is used to specify the inverse transform, which  
does not require the use of a window operator. When 16 x 16  
complex transforms are specified by Bits 2:0, only the rectan-  
gular window can be used. The use of any of the other options  
will cause the device to enter an internal test mode.  
BITS 8:6  
These bits define 0%, 50%, or 75% data block overlap-  
ping, and the division factor on the DIS input. Overlapping  
must not be specified with 16 x 16 complex transforms.  
Two decodes allow the DIS input to be divided by two or four,  
when 50% and 75% overlapping is respectively needed.  
These options allow the DOS and DIS input pins to be still  
supplied from a common source, even though the output rate  
mustbefasterthantheinputrate.Thefrequencyofthissource  
would be dictated by the output rate requirement, with the  
input rate internally reduced by the correct amount.  
Special decodes are provided to support real only trans-  
forms from dual sources, using both the real and auxiliary  
(N - 1)L > 2PK + 2D for 50% overlaps  
(N - 1)L > 4PK + 4D for 75% overlaps  
where L is the time needed to load a normal block of data but  
not including the extra data, P is the number of system clock  
periods given in Table 4, K is the system clock period, and D  
is the total dump time including 4 extra DOS periods.  
When real transforms are to be performed on single  
sourced data, an external FIFO is needed to provide pairs of  
data blocks. These are loaded simultaneously into the real  
and imaginary inputs. See the section on real transforms.  
16