The information contained herein is subject to change without notice.
The information contained herein is presented only as a guide for the applications of our products. No
responsibility is assumed by Hyundai for any infringements of patents or other rights of the third
parties which may result from its use. No license is granted by implication or otherwise under any
patent or patent rights of Hyundai or others.
These Hyundai products are intended for usage in general electronic equipment (office equipment,
communication equipment, measuring equipment, domestic electrification, etc.).
Please make sure that you consult with us before you use these Hyundai products in equipment which
require high quality and / or reliability, and in equipment which could have major impact to the welfare
of human life (atomic energy control, airplane, spaceship, traffic signal, combustion control, all types
of safety devices, etc.). Hyundai cannot accept liability to any damage which may occur in case these
Hyundai products were used in the mentioned equipment without prior consultation with Hyundai.
Copyright 1999 Hyundai Micro Electronics Co.,Ltd.
All Rights Reserved
3
Page 3
GDC21D401B
TABLE OF CONTENTS
1. General Description............................................................................................................5
The Video Decoder(VD) decodes video
elementary stream of MPEG-2(ISO/ICE
13818-2)MP@HL. It supports the ATSC
digital TV video standard, and can be used for
the video part of the ATSC digital TV with the
Transport Decoder and the VDP(Video
Display Processor). Picture decoding timing
can be controlled internally for A/V lip
synchronization, and externally for Video
Trick Mode by host microprocessor via I2C bus.
The Video Decoder can extract video user data
including caption from video elementary
stream, and host microprocessor can read the
video user data from the Video Decoder(VD)
via I2C. It uses four 16x1M SDRAMs and can
support up to 81 MHz memory clock speed.
2. Features
• Supports MPEG-2 (ISO/ICE 13818-2)
MP@HL
• Supports all video input formats of ATSC
digital TV standard
• Supports picture decoding capability up to
1920x1088 30 Frame/Sec
• Supports all kinds of motion compensation
methods of MPEG-2
50 % duty cycle (the same clock as MCLK)
SDRAM interface clock through clock buffer for delay effect.
This signal input is MCLK_IN.External system time clock. - 27 MHz
RESET
8IPower on reset(active low). At least 3 VDCLKs.
Decoding starts after 128 VDCLKs from the last reset low state.
I2C-BUS INTERFACE
26I
28I/O
I2C-bus serial clock. - 400 KHz(max)
I2C-bus serial data
TRANSPORT INTERFACE
I
Transport Decoder data bus
18,16,15,14
29O
13I
Transport data request(active low)
Transport data strobe.
VID_DATA[7:0] is latched on the rising edge.
9I
PTS & DTS data enable(active high).
In LG DTV chipset, this signal is connected to the
PTS_DTS_STRB pin of GDC21D301A.
10I
11I
Video bitstream data enable(active low)
STC data enable(active high)
HOST INTERRUPT
31O
30O
Video decoder interrupt(active low)
User data FIFO is full(active low).
When it happens, host microprocessor must read the user data
from user data FIFO.
Otherwise video decoder suspends decoding.
This indicates bank address, and low value selects bank ‘0’ .
103,102,101,
99,92,89,86,
84,87,91,98
163,162,161,
158,157,154,
152,150,43,
42,40,39,38,
36, 34, 33
O
I/O
SDRAM address
SDRAM data bus
GDC21D401B
7
Page 7
Pin Description (continued)
NAMEPINTYPEDESCRIPTION
VDP INTERFACE - SYNC & PICTURE FORMAT
DIS_INFO
D_INFO_WIN
PIC_DIS_SYNC
P_WAIT
PDWIN
PSTR[1:0]
\FFPN
SCLK
MBCLK
DEC_ERROR
FP_FD
174O
175O
230I
231I
177O
179,178O
171O
166O
168O
233O
235O
GDC21D401B
Serialized picture format data
Serialized picture format data enable(active high)
Picture display sync. - 30 Hz or 29.97 Hz, 50% duty
VDP INTERFACE – PICTURE DATA
PDATA wait(active high).
This signal makes PDATA output to be suspended after 50
VDCLKs from the last high value.
This signal is the output of the VDP.
Picture data window(active high).
During 1 picture data decoding, this signal is high.
Picture structure.
This indicates the structure of output picture.
If this is equal to ‘1’ , the output picture is top field picture.
If this is equal to ‘2’ , the output picture is bottom field
picture.
If this is equal to ‘3’ , the output picture is frame picture.
First field parity(active low).
This signal is the first_field_parity flag of output picture.
When output picture is interlaced frame picture, the field of
output frame is the first output by the VDP.
Slice decoding window(active high).
This signal has high value when a macroblock with the same
vertical position is decoded. There are at least 2-clock low
value periods between each slice decoding window.
Macroblock decoding window(active high).
This signal has high value when a macroblock data is
decoded. The width of high value is always 96 VDCLKs.
There are at least 2-clock low value periods between each
macroblock decoding window.
Decoding Error (active high)
This is a multiplexed output signal.
It is used to inform VDP R1.2 (GDC21D701B) of an error in
Picture or Macro Block.
Frame_Pred_Frame_Dct (active high)
This signal is explained in ISO/IEC 13818-2
(Information technology – Generic coding of moving
pictures and associated audio information : Video)
If this flag is set to ‘1’ , only frame_DCT and frame prediction
are used. In a field picture, it should be ‘0’ .
If progressive_frame is ‘1’ , Frame_pred_frame_dct should be
set to ‘1’ . This flag affects the syntax of the bitstream.
8
Page 8
Pin Description (continued)
227,226,225,
224,221,220,
191,190,187,
NAMEPINTYPEDESCRIPTION
MBFI
PDATA[31:0]
169O
184,183,182
GDC21D401B
Macroblock Field IDCT.
This signal has the meaning when a decoded picture is a frame
picture. If this is set to ‘0’ , the output of a decoded
macroblock has the form of frame IDCT.
If this is set to ‘1’ , the output of a decoded macroblock has the
form of field IDCT.
O
Picture data.
This is a bundle of four adjacent pixel data.
A decoded macroblock consists of 96 consequent PDATA.
The order of PDATA in a decoded macroblock depends on the
MBFI signal.
9
Page 9
4. Block Diagram
GDC21D401B
MHz System Clock
27
Display Sync
Video
Bitstream
I C I/F
2
Video
Bitstream
Predecoder
FIFO
STC
DTS check
Decoding
Controller
Sequence
Parcer
FIFO
IDCT
Coefficient
Decoder
Macroblock
Parameter
Decoder
Motion
Vector
Decoder
64
Internal Data Bus
IDCT
coefficients
MB
parameters
MB
Motion
Vector
Quantization
Matrix
IQ
IQ
&
&
Buffer
Buffer
MB
Decoding
Controller
MV
Processor
SDRAM
Controller
Address
Data
Max. M sample/sec
200
High-Speed IDCT
32
IDCT
IDCT
FIFO
Half-pel
Predictor
Predictor
FIFO
Display
Processor
Control
Decoded
MB Data
32
Data
Window
10
Figure 2. MPEG-2 MP@HL Video Decoder Block Diagram
Page 10
5. Functional Description
5.1 Initialization and Decoding Start
GDC21D401B
5.2 Picture Decoding
¥¡) Power on reset: \RESET = ‘0’ during several
clocks. In this state, bitstream buffers are flushed,
all decoding controllers are reset, and all internal
registers have default values.
¥¢) Decoding start: \RESET = ‘1’ . In this state,
the bitstream buffers begin to be filled with input
video bitstreams. First, the bitstream is discarded
until the 1st sequence header code appears, and
then the sequence headers are decoded. Next, the
picture data bitstream is decoded, and it continues
to be decoded unless the buffers are underflow.
WAIT
SYNC
Picture decoding control state diagram is shown in
Fig 3. Each picture decoding is synchronized with
the external ‘pic_dis_sync‘ signal. If
‘pic_dis_sync‘ signal is changed, command is
executed, then some decoding conditions are
checked (sync parity: top or bottom, underflow,
DTS check and skip check). If the condition is
“repeat”, decoding is suspended until this
condition is ended. If the condition is “skip”, one
picture frame data is skipped. Or else decoding
sync is generated, and controller is suspended for
given ‘pic_dis_sync‘ duration which is
determined by picture format information. If
picture decoding is completed within the given
‘pic_dis_sync’ duration, the state is changed to
wait sync state, and next decoding cycle is
executed. If picture decoding is not completed
within the given duration, the state is changed to
task overrun state. In the task overrun state, as
soon as picture decoding is completed, next
decoding cycle is executed without waiting
‘pic_dis_sync’ transition.
PIC_DIS_SYNC =
or
DEC. COMPLETE
WAIT
PIC_DIS_SYNC =or
Figure 3. Picture Decoding Control State Diagram
DEC. COMPLETE
TASK
OVERRUN
DEC.
SYNC
GEN.
COMMAND
DEC.
CHECK
SKIP
SKIP
WAIT
REPEAT
11
Page 11
5.3 STC (System Time Clock)
Generation
Internal STC counter is supported. STC value
should be loaded after the reset by host
microprocessor or transport decoder, and be
counted at 90 KHz clock derived from external 27
MHz system clock. Also it should be loaded
whenever the system time base is changed. 27
MHz system clock is supplied by transport
decoder through external 27 MHz clock input.
5.4 DTS (Decoding Time Stamp)
Synchronization
The DTS values are determined when the
associated pictures are decoded referring to the
STC. The transport decoder must extract PTS &
DTS values from MPEG-2 system layer, and
transfer them to the VD through host interface or
transport data interface. DTS values are extracted
and stored in DTS-FIFO with 8-bit associated
picture numbers(they are defined by the VD
internally).
Before each picture decoding, picture decoding
controller checks if there is a DTS associated with
each picture. And it compares the DTS value from
DTS-FIFO with STC value.
The DTS value is considered to match the current
STC value if the DTS satisfies this equation:
other words, the time for current picture
decoding has already passed), picture
decoding controller discards B picture in
VBV delay mode and any picture in low
delay mode without decoding it.
If the DTS is greater than STC + jitter_f (the
time for current picture decoding has not
come yet), picture decoding controller pauses
decoding process until the DTS value falls
within the allowed tolerance of STC.
GDC21D401B
When system time base is changed, the time base
of the current DTS can differ from that of the new
STC temporarily because of video decoding delay.
In this case, for about 0.5 sec, DTS
synchronization mode must be disabled by using
decoding mode register.
5.5 Error Concealment
When the VLD detects a bitstream syntax error or
an MPEG-2 error code (0x000001B4), it performs
appropriate error handling and error concealment
to continue the decoding and to minimize the
effect of the error on decoded video. If an error
occurs at picture data layer, the slice-based error
concealment is performed. If an error occurs at
header layer, the associated sequence is skipped.
5.6 User Data Read
User data in MPEG Video Sequence can be read
through host interface(I2C) as follows. If user data
interrupt is enabled, user data is stored in internal
user data FIFO, and once ‘PIC_DEC_SYNC’ is
set to ‘high’ , a interrupt signal is generated which
informs host that user data is stored in user data
FIFO. If user data interrupt is disabled, user data
is discarded.
When user data interrupt signal is generated, host
reads the ‘U_D_COUNT’ register for bytes of
user data FIFO. And then it reads the
‘USER_DATA’ register(user data FIFO output)
repeatedly as many as the number of
‘U_D_COUNT’ register. In this case, the host
had better use I2C burst read cycle for reading
speed efficiency.
If internal user data FIFO becomes full during user
data interrupts, ‘\UBUFF_FULL’ signal is
generated. And also the host reads user data by
using the same method as that of user data
interrupt.
12
Page 12
GDC21D401B
5.7 Bitstream Buffer Over/Underflow
The bitsream buffer should not be overflowed or
underflowed in proper operation. When bitstream
buffer is overflowed, video bitstream input is
disabled during overflow. At the beginning of each
picture, buffer underflow is checked. If the
bitstream buffer doesn’t include 1 picture
bitstream, decoding is paused until this buffer is
filled with 1 picture bitstream.
5.8 VLD (Variable Length Decoder)
The VLD executes the variable length decoding of
MPEG-2 MP@HL. It is composed of sequence
syntax parsing and code decoding. Headers, IDCT
coefficients, and motion vectors are generated from
input video bitstream by the VLD.
5.9 Inverse Quantization
For decoding a macroblock, various parameters
are transmitted to IQ block and MC block in
sequence of decoding syntax from the VLD.
Coefficient data transmitted to IQ block is
multiplied by quantization matrix and quantization
scale. Quantized coefficients are stored in two
coefficient buffers, and if a request from MC block
occurs, the buffer outputs 4 coefficients in parallel.
Inverse scan processing is performed in the
coefficient buffer by varying read/write address
according to scan type.
5.10 IDCT (Inverse Discrete Cosine
Transform)
Coefficient data from IQ block is transmitted to
IDCT block. IDCT block transforms these
coefficients to pixel data or differential pixel data
which have the values in the range of -256 to 255.
For one 8x8 DCT block, the elapse time through
IDCT block is a 30-VDCLK period. In the IDCT
block, 8 adjacent coefficients are processed in
parallel. The maximum transform speed of the
IDCT block is 1 VDCLK x 4 Sample/sec. This
speed is sufficient for decoding 1920x1088 30Hz
frame data.
5.11 MC(Motion Compensation)
Motion Compensation Block restores macroblocks
by using macroblock difference data from IDCT
block and reference data which are read from
external SDRAM. Address Generation Block in
the MC generates SDRAM address by
transforming macroblock address and motion
vector to row/column address of SDRAM. The
MC block supports all motion types of frames and
field pictures. Motion vector range decoded in the
MC has the range of f_code=9 for horizontal
motion vector and f_code=8 for vertical motion
vector. Reconstructed macroblock in the MC block
is stored in the external frame memory for I or P
picture, but in the case of B picture, decoded
macroblock is transmitted directly to the VDP
without being stored in the frame memory.
Because reordering occurs in I or P picture
decoding, macroblock data to be displayed is read
from reference frame memory and is transmitted to
the VDP in the form of macroblock.
13
Page 13
5.12 Transport Interface
GDC21D401B
8-bit parallel data(vid_data[7:0]) interface is
supported for transport data interface. Three kinds
of data (Video bitstream, STC, and DTS) can be
transferred through transport data interface.
If \vid_en is low, video bitstream is transferred. If
vstcw is high, STC is transferred. If tsw is high,
PTS & DTS are transferred. Or else no data is
VID_STRB
TSW
VSTCW
\VID_EN
VID_DATA
VID_STRB
TSW
32
23:1631:24
PTS[32:0]
15:8
7:0
32
all ‘ 0’
all ‘ 1’
23:1631:24
DTS[32:0]
all ‘ 0’
transferred.
Vid_data[7:0] is latched on the rising edge of
vid_strb. When \VID_REQ is high, transport
data interface is disabled, and when \VID_REQ is
low, transport data interface is enabled. See the
following Fig. 4.
15:8
7:0
32
23:1631:24
PTS[32:0]
15:8
7:0
14
VSTCW
\VID_EN
VID_DATA
32
23:1631:24
STC[32:0]
15:8
7:0
. . . . .
Video Bitstream
Figure 4. Transport Data Interface Timing Diagram
Page 14
5.13 Host Interface
GDC21D401B
I2C bus interface
I2C bus interface is used for host data interface. It
operates only as a Slave. The Chip-ID(dev
address) of this IC is “0001111”b. Data on the
I2C-bus can be transferred at the rate up to 100
START
CHIP ID
0001111
S0
START
CHIP ID
0001111
S0
WRITE
ACK
0 REG. ADDR.(7:0) 0 S00011110 REG. DATA(7:0)00 P
ACKACK
Figure 5. Write Cycle Diagram
WRITE
ACK
0 REG. ADDR.(7:0) 0 S00011110 REG. DATA(7:0)11 P
ACKACK
START
START
kbit/s in the standard mode, or up to 400 kbit/s in
the fast mode. I2C bus data read/write formats of
this IC are shown in the Fig 5, 6, and 7. Burst
Read Cycle can be used for the user data reading.
For more information, see the I2C bus interface
standard.
PAUSE
PAUSE
CHIP ID
CHIP ID
WRITE
ACK
READ
ACK
START
CHIP ID
0001111
S0
Figure 6. Read Cycle Diagram
WRITE
ACK
0REG. ADDR.(7:0) 0S00011110REG. DATA(7:0)10
START
ACKACK
CHIP ID
ACKACK
READ
ACK
REG. DATA(7:0) 1PREG. DATA(7:0) 0. . . . . .
Figure 7. Burst Read Cycle Diagram
PAUSE
15
Page 15
GDC21D401B
Interrupt Mechanism
The host enables specific interrupt events to occur
by setting the mask value of internal interrupt.
When an event occurs and corresponding interrupt
enable is set, an interrupt bit is set in an interrupt
register. Everytime each picture decodes sync, host
checks the contents of that register, and if any bit
Table 2. Bit Definitions for Interrupt Register & Mask
BIT #MNEMONICEVENTEVENT DEFINITION
0PIC_SPicture decodingNew picture is decoded.1GOP_HGOP header decodingGOP headers are decoded.2SEQ_HSeq. header decodingSequence headers are decoded.3SEQ_ESeq. end decodingSequence end code is decoded.4USRUser data ready
User data has been extracted and stored in the buffer. When
the interrupt is disabled, user data is discarded.
5OVFBuffer overflowBitstream buffer is full.6UND/PTSBuffer underflow/
PTS received
Dec_mode(4) = ‘0’ :Bitstream buffer doesn’t have 1 picture bitstream.Dec_mode(4) = ‘1’PTS value is received through transport interface.
7ERRBitstream errorError code or syntax error is detected in the bitstream
(header layer).
is set, it generates an interrupt. Exceptionally, in
the case of OVF, UND and ERR interrupt signals
are generated as soon as the error is detected.
When the host serves the interrupt, interrupt
register and interrupt signal are reset.
16
Page 16
GDC21D401B
Register Description
Table 3. Internal Register Description
NONAMESIZEDESCRIPTIONDEFAULTR/W
0Command3High level decoding command.play()R/W1Interrupt8Interrupt register.all ‘0’R2DTS08DTS[7:0]. When it is written by host
microprocessor, DTS[19:0] values are stored in the
DTS-FIFO.
When DTS value is received through transport
interface, it can be read via the DTS registers
3DTS18DTS[15:8].all ‘0’R/W4DTS24DTS[19:16].all ‘0’R/W5Jitter_f08Forward tolerance of DTS synchronization,
Jitter_f[7:0]. When it is written by host
microprocessor, Jitter_f[19:16] values are applied.
6Jitter_f18Jitter_f[15:8].h”0E”R/W7Jitter_f24Jitter_f[19:16].h“0”R/W8Jitter_b08Backward tolerance of DTS synchronization,
Jitter_b[7:0]. When it is written by host
microprocessor, Jitter_b[19:16] values are applied.
9Jitter_b18Jitter_b[15:8].h“FA”R/W10Jitter_b24Jitter_b[19:16].h“F”R/W11Slm_num6Slow motion command of repeated numbers.h’8’R/W12Header8Header data register. It has header value indicated
by H_addr.
13H_addr8Header data address.all ‘0’R/W14User_data8User data register. It is an output of the User data
FIFO.
15U_d_count7User data count in the User data FIFO.all ‘0’R16Int_mask8Interrupt mask.all ‘0’R/W17Dec_mode8Decoding mode.all ‘0’R/W18STC08System time clock, STC[7:0]. When it is written by
host microprocessor, STC[19:0] values are loaded
to internal STC counter.
19STC18STC[15:8].all ‘0’R/W20STC24STC[19:16].all ‘0’R/W21W_ptr8Write address pointer of the bitstream buffer. Unit
value is 65,536 bits.
22R_ptr8Read address pointer of the bitstream buffer. Unit
value is 65,536 bits.
23PTS08PTS[7:0]. When PTS value is received through
tranport interface, it is stored in the PTS registers.
0 Reset() Bitstream buffer is flushed and decoding controllers are reset.
1 Play() Start normal bitstream decoding.
2 Slow_motion()
3 Scan() Skip to I picture and continue decoding.
4 Single_step() Decode 1 picture and pause.
5 Pause() Decoding pause.
6 Fast_forward() Decode only I or P picture.
7 NoCOM() No command is loaded in the register. In this state bitstream buffer
Decode pictures and pause during slm_num duration, and repeat this
Video decoder transmits decoded macroblock data
to the VDP via 32-bit bus. This bus consists of 4
adjacent pixel components. For a macroblock data,
the number of 32-bit data is 96. The order of
output data is the same of decoded block in the
MPEG-2 Video bitstream syntax. For this reason,
output sequence depends on the DCT_TYPE of
decoded macroblock. Except for the macroblock
whose parameter is Frame picture and Field DCT,
all macroblock outputs have the sequence of below
figure(MBFI=0) in Fig 8. The top field data of the
macroblock whose DCT_TYPE is field
type(MBFI=1) is transmitted before the
macroblock bottom field data for the luminance
data. But the chrominance data has always the
same sequence.
MBFI == 0
MBFI == 1
Figure 8. The Sequence of Macroblock Output
20
Page 20
5.15 Video Data Output Timing
GDC21D401B
During decoding the picture which is field picture
or frame picture, PDWIN signal is high. Between
each picture decoding time, there is low level
period of PDWIN signal, and it is longer than 128
VDCLKs. Picture parameter such as PSTR[1:0] or
\FFPN is determined 2 clocks before PDWIN
PDWIN
PSTR[1:0]
\FFPN
SCLK
2 VDCLKsMinimum
SCLK
MBCLK
MBFI
1 VDCLK
rising edge. Its value is not changed until PDWIN
falling edge. SCLK shows that new slice decoding
is started on its rising edge. Width of MBCLK
high pulse is always 96 clocks, and low value
period is longer than 2 clocks.
Minimum
128 VDCLKs
PDATA[31:0]
VDCLK
01239594
Figure 9. Timing Diagram
21
Page 21
5.16 SDRAM Interface
GDC21D401B
For storing reference frame data and bitstream,
very large external memory is required. For this
reason, 64-Mbit external memory
(GM72V161621AT or same type of memory) is
needed. The memory configuration is shown in the
following Fig 10. Four 16-Mbit memories are
directly attached to the VD. Data bus for
Revision Note for SDRAM Clock Application Note(1):
Recommended
CSN,WEN,RASN,CASN, BA0
SDRAM_ADDR[10:0]
SDRAM_DATA[63:0]
MCLK
# 94
4 ~ 10.5ns
5
11
64
SDRAMs is configured to 64-bit bus. Memory
interface clock should be supplied to the VD and
four SDRAMs with same clock phase. Higher
frequency for MCLK and higher decoding speed
can be achieved. The maximum frequency of
MCLK is 81 MHz. This clock doesn’t need to be
locked with any other input clock.
16
SDRAM
CLK
SDRAM
16
CLK
22
CLK
Driver
MCLK_IN
MCLK_OUT
VSS
N.C.
(No Connect)
Figure 10. External Memory Interface 1
16
16
SDRAM
CLK
SDRAM
CLK
Page 22
Revision Note for SDRAM Clock Application Note(2)
GDC21D401B
CSN,WEN,RASN,CASN, BA0
SDRAM_ADDR[10:0]
SDRAM_DATA[63:0]
MCLK
MCLK_IN
MCLK_OUT
CLK
Driver
# 94
# 234
# 237
5
11
64
# 234# 237
Clock delay equivalent circuit
16
16
16
16
2 ~ 5 ns
SDRAM
CLK
SDRAM
CLK
SDRAM
CLK
SDRAM
CLK
Figure 11. External Memory Interface 2
23
Page 23
GDC21D401B
6. Electrical Specification
6.1 Absolute Maximum Rating
SYMBOLPARAMETERSVALUESUNIT
V
DD
V
I
Vo
T
stg
P
d
*Note :Absolute Maximum Ratings mean that the safety of the device cannot be guaranteed beyond these values,
and this doesn’t imply that the device should be operated within these limits.
6.2 Recommended Operating Range
SYMBOLPARAMETERSVALUESUNIT
V
DD
T
opr
6.3 DC Characteristics (VDD = 3.3 V¡¾10%, TA = 0 ~ 70¡É)
Power Supply Voltage-0.33 to 5.5V
Digital Input Voltage-0.33 to VDD + 0.5V
Digital Output Voltage-0.33 to VDD + 0.5V
Storage Temperature-55 to 125
°C
Power Dissipation4.6W
Power Supply Voltage
Operating Temperature0 to 70
3.3 ± 10%
°C
V
SYMBOLPARAMETERMINMAXUNIT
V
V
V
V
I
I
F
IH
IL
OH
OL
DD
DDQ
OPR
Input High Voltage0.7X V
Input Low Voltage-0.330.2X V
Output High Voltage2.4V
Output Low Voltage00.4V
Operating Current-900mA
Standby Current-10uA
Operating Frequency-54MHz
DD
V
+0.33V
DD
DD
DD
V
V
24
Page 24
6.4 AC Characteristics (VDD = 3.3 V¡¾10%, TA = 0 ~ 70¡É)
All Outputs
All Intputs
CHclk27m/tCLvdclk
CHvid strb/tCLvdclk
t
CP
MCLK
VDCLK
CLK27M
VID_STRB
t
CH
t
CL
GDC21D401B
except for
SDRAM I/F
signals
except for
SRAM I/F
signals
t
S
t
OD
t
H
Figure 12. Input/Output Timing
SYMBOLPARAMETERSMINMAXTYPEUNIT
t
CPmclk
t
CPvdclk
t
CPclk27m
t
CPvid strb
t
CHmclk/tCLmclk
t
CHvdclk/tCLvdclk
t
t
t
H
ts
t
OD
Clock Period--13ns
Clock Period--20ns
Clock Period--36ns
Clock Period--36ns
Clock High Time/Clock Low Time5.57.5-ns
Clock High Time/Clock Low Time812-ns
Clock High Time/Clock Low Time1620-ns
Clock High Time/Clock Low Time1620-ns
Input Hold Time7--ns
Input Setup Time7--ns
Output Delay Time-17-ns
*Note : Low voltage input signal rising and falling edge switching time = 1.0 ns