View CH7006C-V_2405422.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

201-0000-026 rev 2.1, 8/2/99 *intel i740 is a trademark of intel corp. 1 ch7006c chrontel digital pc to tv encoder features features ? function compatible with ch7004 ? universal digital interface accepts ycrcb (ccir601 or 656) or rgb (15, 16 or 24-bit) video data in both non-interlaced and interlaced formats ? truescale tm rendering engine supports underscan operations for various graphic resolutions ? ? enhanced text sharpness and adaptive flicker removal with up to 5-lines of filtering ? ? enhanced dot crawl control and area reduction ? fully programmable through i 2 c port ? supports ntsc, ntsc-eia (japan), and pal (b, d, g, h, i, m and n) tv formats ? provides composite, s-video and scart outputs ? auto-detection of tv presence ? supports vbi pass-through ? programmable power management ? 9-bit video dac outputs ? complete windows and dos driver software ? offered in 44-pin plcc, 44-pin tqfp general description chrontel ?s ch7006 digital pc to tv encoder is a stand- alone integrated circuit which provides a pc 99 compliant solution for tv output. suggested application use with the intel i740.* it provides a universal digital input port to accept a pixel data stream from a compatible vga controller (or equivalent) and converts this directly into ntsc or pal tv format. this circuit integrates a digital ntsc/pal encoder with 9-bit dac interface, and new adaptive flicker filter, and high accuracy low-jitter phase locked loop to create outstanding quality video. through its truescale tm scaling and deflickering engine, the ch7006 supports full vertical and horizontal underscan capability and operates in 5 different resolutions including 640x480 and 800x600. a new universal digital interface along with full programmability make the ch7006 ideal for system-level pc solutions. all features are software programmable through a standard i 2 c port, to enable a complete pc solution using a tv as the primary display. ? patent number 5,781,241 patent number 5,914,753 figure 1: functional block diagram triple dac pll rgb-yuv converter system clock y/r cvbs/b c/g yuv-rgb converter digital input interface i 2 c register & control block line memory true scale scaling & deflickering engine timing & sync generator ntsc/pal encoder & filters d[15:0] pixel data xclk h v xi xo/fin p-out sc sd reset* ds/bco rset csync
chrontel ch7006c 2 201-0000-026 rev 2.1, 8/2/99 figure 2: 44-pin plcc xo/fin xi dvdd reset* d[3] d[4] d[5] d[8] d[6] dvdd d[7] dgnd sc sd dgnd] d[9] d[10] d[11] avdd vdd rset gnd d [ 2 ] d [ 1 ] v h x c l k d v d d p - o u t d [ 0 ] d g n d d s / b c o a g n d d [ 1 2 ] d [ 1 3 ] d [ 1 4 ] d g n d d [ 1 5 ] d v d d c s y n c g n d c v b s c y 7 8 9 10 13 12 11 39 38 37 36 35 14 15 16 17 34 33 32 31 30 29 6 5 4 3 2 1 4 4 4 3 4 2 4 1 4 0 1 8 1 9 2 0 2 1 2 4 2 3 2 2 2 5 2 6 2 7 2 8 chrontel ch7006
201-0000-026 rev 2.1, 8/2/99 3 chrontel ch7006c figure 3: 44-pin tqfp xo/fin xi dvdd addr d[3] d[4] d[5] d[8] d[6] dvdd d[7] dgnd sc sd dgnd] d[9] d[10] d[11] avdd vdd rset gnd d [ 2 ] d [ 1 ] v x c l k d v d d p - o u t d [ 0 ] d g n d d s / b c o a g n d d [ 1 2 ] d [ 1 3 ] d [ 1 4 ] d g n d d [ 1 5 ] d v d d c s y n c g n d c v b s c y 1 2 3 4 7 6 5 33 32 31 30 29 8 9 10 11 28 27 26 25 24 23 4 4 4 3 4 2 4 1 4 0 3 9 3 8 3 7 3 6 3 5 3 4 1 2 1 3 1 4 1 5 1 8 1 7 1 6 1 9 2 0 2 1 2 2 chrontel ch7006 d[3] d[4] d[5] d[8] d[6] dvdd d[7] dgnd] d[9] d[10] d[11] xo/fin xi dvdd reset* dgnd sc sd avdd vdd rset gnd h
chrontel ch7006c 4 201-0000-026 rev 2.1, 8/2/99 table 1. pin descriptions 44-pin plcc 44pin tqfp type symbol description 4-10, 12-13, 15-21 1,2, 3,4, 6,7,9, 10,11, 12,13, 14,15, 42,43, 44 in d15-d0 digital pixel inputs these pins accept digital pixel data streams with either 8, 12, or 16-bit multiplexed or 16-bit non-multiplexed formats, determined by the input mode setting (see registers and programming section). inputs d0 - d7 are used when operating in 8-bit multiplexed mode. inputs d0 - d11 are used when operating in 12-bit mode. inputs d0 - d15 are used when operating in 16-bit mode. the data structure and timing sequence for each mode is described in the section on digital input port. 43 37 out p-out pixel clock output the ch7006, operating in master mode, provides a pixel data clocking signal to the vga controller. this clock will only be provided in master clock modes and will be tri-stated otherwise. this pin provides the pixel clock output signal (adjustable as 1x,2x or 3x) to the vga controller (see the section on digital video interface, registers and programming for more details). the capacitive loading on this pin should be kept to a minimum. 1 39 in xclk pixel clock input to operate in a pure master mode, the p-out signal should be connected to the xclk input pin. to operate in a pseudo-master mode, the p-out clock is used as a reference frequency, and a signal locked to this output (at 1x, 1/2x, or 1/3x the p-out frequency) is input to the xclk pin. to operate in slave mode, the ch7006 accepts an external pixel clock input at this pin. the capacitive loading on this pin should be kept to a minimum. 3 41 in/out v vertical sync input/output this pin accepts the vertical sync signal from the vga controller, or outputs a vertical sync to the vga controller. the capacitive loading on this pin should kept to a minimum. 2 40 in/out h horizontal sync input/output this pin accepts the horizontal sync from the vga controller, or outputs a horizontal sync to the vga controller. the capacitive loading on this pin should be kept to a minimum. 41 35 in/out ds/bco data/start (input) / buffered clock (output) when configured as an input, the rising edge of this signal identifies the first active pixel of data for each active line. when configured as an output this pin provides a buffered clock output. the output clock can be selected using the bco register (17h) (see registers and programming). 38 32 in xi crystal input a parallel resonance 14.31818 mhz ( 50 ppm) crystal should be attached between xi and xo/fin. however, if an external cmos clock is attached to xo/fin, xi should be connected to ground. 39 33 in xo/fin crystal output or external fref a 14.31818 mhz ( 50 ppm) crystal may be attached between xo/fin and xi. an external cmos compatible clock can be connected to xo/fin as an alternative.
201-0000-026 rev 2.1, 8/2/99 5 chrontel ch7006c 44-pin plcc 44pin tqfp type symbol description 30 24 in rset reference resistor a 360 w resistor with short and wide traces should be attached between rset and ground. no other connections should be made to this pin. 28 22 out y/r luminance output a 75 w termination resistor with short traces should be attached between y and ground for optimum performance. in normal operating modes other than scart and rgb bypass, this pin outputs the composite video signal. in scart and rgb bypass modes, this pin outputs the red signal. 27 21 out c/g chrominance output a 75 w termination resistor with short traces should be attached between c and ground for optimum performance. in normal operating modes other than scart and rgb bypass, this pin outputs the composite video signal. in scart and rgb bypass modes, this pin outputs the green signal. 26 20 out cvbs/b composite video outpu t a 75 w termination resistor with short traces should be attached between cvbs and ground for optimum performance. in normal operating modes other than scart and rgb bypass, this pin outputs the composite video signal. in scart and rgb bypass modes, this pin outputs the blue signal. 23 17 out csync composite sync output a 75 w termination resistor with short traces should be attached between csync and ground for optimum performance. in scart mode, this pin outputs the composite sync signal. 32 26 in/out sd serial data (external pull-up required) this pin functions as the serial data pin of the i 2 c interface port (see the i 2 c port operation section for details). this pin uses the dvdd supply and is not 5v tolerant. 33 27 in sc serial clock (internal pull-up) this pin functions as the serial clock pin of the i 2 c interface port (see the i 2 c port operation section for details). this pin uses the dvdd supply and is not 5v tolerant. 35 29 in reset* reset input when this pin is low, the ch7006 is held in the power-on reset condition. when this pin is high, the device operates normally and reset is controlled through the i 2 c register. 40 34 power agnd analog ground this pin provides the ground reference for the analog section of the ch7006, and must be connected to the system ground, to prevent latchup. refer to the application information section for information on proper supply decoupling. 37 31 power avdd analog supply voltage this pins supplies the 5v power to the analog section of the ch7006. 31 25 power vdd dac power supply this pins supplies the 5v power to ch7006 ?s internal dac ?s. table 1. pin descriptions
chrontel ch7006c 6 201-0000-026 rev 2.1, 8/2/99 digital video interface the ch7006 digital video interface provides a flexible digital interface between a computer graphics controller and the tv encoder ic, forming the ideal quality/cost configuration for performing the tv-output function. this digital interface consists of up to 16 data signals and 4 control signals, all of which are subject to programmable control through the ch7006 register set. this interface can be configured as 8, 12 or 16-bit inputs operating in either multiplexed mode or 16-bit input operation in demultiplexed mode. it will also accept either ycrcb or rgb (15, 16 or 24-bit) data formats and will accept both non-interlaced and interlaced data formats. a summary of the data format modes is as follows: 44-pin plcc 44pin tqfp type symbol description 25, 29 19,23 power gnd dac ground these pins provide the ground reference for ch7006 ?s internal dacs. for information on proper supply decoupling, please refer to the application information section. 11, 22, 36, 44 5,16, 30,38 power dvdd digital supply voltage these pins supply the 3.3v power to the digital section of ch7006. 14, 24, 34, 42 8,18, 28,36 power dgnd digital ground these pins provide the ground reference for the digital section of ch7006, and must be connected to the system ground to prevent latchup. n/a n/a out r r (red) component output this pin provides the analog red component of the digital rgb input in the rgb pass-through mode. n/a n/a out g g (green) component output this pin provides the analog green component of the digital rgb input in the rgb pass-through mode. n/a n/a out b b (blue) component output this pin provides the analog blue component of the digital rgb input in the rgb pass-through mode. table 2. input data formats bus width transfer mode color space and depth format reference 16-bit non-multiplexed rgb 16-bit 5-6-5 each word 15-bit non-multiplexed rgb 15-bit 5-5-5 each word 16-bit non-multiplexed ycrcb (24-bit) cby0,cry1...(ccir656 style) 8-bit 2x-multiplexed rgb 15-bit 5-5-5 over two bytes 8-bit 2x-multiplexed rgb 16-bit 5-6-5 over two bytes 8-bit 3x-multiplexed rgb 24-bit 8-8-8 over three bytes 8-bit 2x-multiplexed ycrcb (24-bit) cb,y0,cr,y1,(ccir656 style) 12-bit 2x-multiplexed rgb 24 8-8-8 over two words - ?c ? version 12-bit 2x-multiplexed rgb 24 8-8-8 over two words - ?i ? version 16-bit 2x-multiplexed rgb 24 (32) 8-8,8x over two words table 1. pin descriptions
201-0000-026 rev 2.1, 8/2/99 7 chrontel ch7006c the clock and timing signals used to latch and process the incoming pixel data is dependent upon the clock mode. the ch7006 can operate in either master (the ch7004 generates a pixel frequency which is either returned as a phase-aligned pixel clock or used directly to latch data), or slave mode (the graphics chip generates the pixel clock). the pixel clock frequency will change depending upon the active image size (e.g., 640x480 or 800x600), the desired output format (ntsc or pal), and the amount of scaling desired. the pixel clock may be requested to be 1x, 2x, or 3x the pixel data rate (subject to a 100mhz frequency limitation). in the case of a 1x pixel clock the ch7006 will automatically use both clock edges, if a multiplexed data format is selected. sync signals: horizontal and vertical sync signals will normally be supplied by the vga controller, but may be selected to be generated by the ch7006. in the case of ccir656 style input (idf = 1 or 9), embedded sync may also be used. (in each case, the period of the horizontal sync should be equal to the duration of the pixel clock, time the first value of the ( total pixels/line x total lines/frame) column of table 16 on page 31(display mode register 00h description). the leading edge of the horizontal sync is used to determine the start of each line. the vertical sync signal must be able to be set to the second value in the ( total pixels/line x total lines/frame) column of table 16 on page 31.) master clock mode: the ch7006 generates a clock signal (output at the p-out pin) which will be used by the vga controller as a frequency reference. the vga controller will then generate a clock signal which will be input via the xclk input. this incoming signal will be used to latch (and de-multiplex, if required) incoming data. the xclk input clock rate must match the input data rate, and the p-out clock can be requested to be 1x, 2x or 3x the pixel data rate. as an alternative, the p-out clock signal can also be used as the input clock signal (connected directly to the xclk input) to latch the incoming data. if this mode is used, the incoming data must meet setup and hold times with respect to the xclk input (with the only internal adjustment being xclk polarity). slave clock mode: the vga controller will generate a clock which will be input to the xclk pin (no clock signal will be output on the p-out pin). this signal must match the input data rate, must occur at 1x, 2x or 3x the pixel data rate, and will be used to latch (and de-multiplex if required) incoming data. also, the graphics ic transmits back to the tv encoder the horizontal and vertical timing signals, and pixel data, each of which must meet the specified setup and hold times with respect to the pixel clock. pixel data: active pixel data will be expected after a programmable number pixels times the multiplex rate after the leading edge of horizontal sync. in other words, specifying the horizontal back porch value (as a pixel count), plus horizontal sync width, will determine when the chip will begin to sample pixels. non-multiplexed mode in the 15/16-bit mode shown in figure4 , the pixel data bus represents a 15/16-bit non-multiplexed data stream, which contains either rgb or ycrcb formatted data. when operating in rgb mode, each 15/16-bit pn value will contain a complete pixel encoded in either 5-6-5 or 5-5-5 format. when operating in ycrcb mode, each 16-bit pn word will contain an 8-bit y (luminance) value on the upper 8 bits, and an 8-bit c (color difference) value on the lower 8 bits. the color difference will be transmitted at half the data rate of the luminance data, with the sequence- being set as cb followed by cr. the cb and cr data will be cosited with the y value transmitted with the cb value, with the data sequence described in table3 . the first active pixel is sav pixels after the trailing edge of horizontal sync, where sav is a bus-controlled register.
chrontel ch7006c 8 201-0000-026 rev 2.1, 8/2/99 figure 4: non-multiplexed data transfers when idf = 1, (ycrcb 16-bit mode), h and v sync signals can be embedded into the data stream. in this mode, the embedded sync will be similar to the ccir656 convention (not identical, since that convention is for 8-bit data streams), and the first byte of the ?video timing reference code ? will be assumed to occur when a cb sample would occur ? if the video stream was continuous. this is delineated in table4 below. in this mode, the s[7-0] byte contains the following data: s[6] = f = 1 during field 2, 0 during field 1 s[5] = v = 1 during field blanking, 0 elsewhere s[4] = h = 1 during eav (the synchronization reference at the end of active video) 0 during sav (the synchronization reference at the start of active video) bits s[7] and s[3-0] are ignored. table 3. ycrcb non-multiplexed mode with embedded syncs idf# format 1 ycrcb 16-bit pixel# p0 p1 p2 p3 p4 p5 p6 p7 bus data d[15] 0 s[7] y0[7] y1[7] y2[7] y3[7] y4[7] y5[7] d[14] 0 s[6] y0[6] y1[6] y2[6] y3[6] y4[6] y5[6] d[13] 0 s[5] y0[5] y1[5] y2[5] y3[5] y4[5] y5[5] d[12] 0 s[4] y0[4] y1[4] y2[4] y3[4] y4[4] y5[4] d[11] 0 s[3] y0[3] y1[3] y2[3] y3[3] y4[3] y5[3] d[10] 0 s[2] y0[2] y1[2] y2[2] y3[2] y4[2] y5[2] d[9] 0 s[1] y0[1] y1[1] y2[1] y3[1] y4[1] y5[1] d[8] 0 s[0] y0[0] y1[0] y2[0] y3[0] y4[0] y5[0] d[7] 1 00 cb0[7] cr0[7] cb2[7] cr2[7] cb4[7] cr4[7] d[6] 1 0 cb0[6] cr0[6] cb2[6] cr2[6] cb4[6] cr4[6] d[5] 1 0 cb0[5] cr0[5] cb2[5] cr2[5] cb4[5] cr4[5] d[4] 1 0 cb0[4] cr0[4] cb2[4] cr2[4] cb4[4] cr4[4] d[3] 1 0 cb0[3] cr0[3] cb2[3] cr2[3] cb4[3] cr4[3] d[2] 1 0 cb0[2] cr0[2] cb2[2] cr2[2] cb4[2] cr4[2] d[1] 1 0 cb0[1] cr0[1] cb2[1] cr2[1] cb4[1] cr4[1] d[0] 1 0 cb0[0] cr0[0] cb2[0] cr2[0] cb4[0] cr4[0] sav p0 p1 p2 p3 p4 p5 t p 1 t ph 1 t hp1 t sp1 t hsw pixel data pout/ xclk hsync t hd p0a p0b p1a p1b p2a p2b t ph t p t hp t sp avr
201-0000-026 rev 2.1, 8/2/99 9 chrontel ch7006c multiplexed mode each rising edge (or each rising and falling edge) of the xclk signal will latch data from the graphics chip. the multiplexed input data formats are shown in figure5 and 6 . the pixel data bus represents an 8, 12, or 16-bit multiplexed data stream, which contains either rgb or ycrcb formatted data. in idf settings of 2, 4, 5, 7, 8 and 9, the input data rate is 2x pclk, and each pair of pn values (e.g., p0a and p0b) will contain a complete pixel, encoded as shown in the tables below. when idf = 6, the input data rate is 3x pclk, and each triplet of pn values (e.g., p0a, p0b and p0c) will contain a complete pixel, encoded as shown in the tables below. when the input is ycrcb, the color-difference data will be transmitted at half the data rate of the luminance data, with the sequence being set as cb, y, cr, y where cb0,y0,cr0 refers to co-sited luminance and color-difference samples ? and the following y1 byte refers to the next luminance sample, per ccir656 standards. however, the clock frequency is dependent upon the current mode, (not 27mhz, as specified in ccir656). figure 5: multiplexed pixel data transfer mode table 4. rgb 8-bit multiplexed mode idf# format 7 rgb 5-6-5 8 rgb 5-5-5 pixel# p0a p0b p1a p1b p0a p0b p1a p1b bus data d[7] g0[2] r0[4] g1[2] r1[4] g0[2] x g1[2] x d[6] g0[1] r0[3] g1[1] r1[3] g0[1] r0[4] g1[1] r1[4] d[5] g0[0] r0[2] g1[0] r1[2] g0[0] r0[3] g1[0] r1[3] d[4] b0[4] r0[1] b1[4] r1[1] b0[4] r0[2] b1[4] r1[2] d[3] b0[3] r0[0] b1[3] r1[0] b0[3] r0[1] b1[3] r1[1] d[2] b0[2] g0[5] b1[2] g1[5] b0[2] r0[0] b1[2] r1[0] d[1] b0[1] g0[4] b1[1] g1[4] b0[1] g0[4] b1[1] g1[4] d[0] b0[0] g0[3] b1[0] g1[3] b0[0] g0[3] b1[0] g1[3] t hsw hs t hd t ph2 t p2 t hp2 t sp2 d[15:0] p0a p0b p1a p1b p2a p2b xclk dec = 0 xclk dec = 1 t sp2 t sp2 t hp2 t hp2
chrontel ch7006c 10 201-0000-026 rev 2.1, 8/2/99 note: the ax[7:0] data is ignored. table 5. rgb 12-bit multiplexed mode idf# format 4 12-bit rgb (12-12) 5 12-bit rgb (12-12) pixel# p0a p0b p1a p1b p0a p0b p1a p1b bus data d[11] g0[3] r0[7] g1[3] r1[7] g0[4] r0[7] g1[4] r1[7] d[10] g0[2] r0[6] g1[2] r1[6] g0[3] r0[6] g1[3] r1[6] d[9] g0[1] r0[5] g1[1] r1[5] g0[2] r0[5] g1[2] r1[5] d[8] g0[0] r0[4] g1[0] r1[4] b0[7] r0[4] b1[7] r1[4] d[7] b0[7] r0[3] b1[7] r1[3] b0[6] r0[3] b1[6] r1[3] d[6] b0[6] r0[2] b1[6] r1[2] b0[5] g0[7] b1[7] g1[7] d[5] b0[5] r0[1] b1[5] r1[1] b0[4] g0[6] b1[4] g1[6] d[4] b0[4] r0[0] b1[4] r1[0] b0[3] g0[5] b1[3] g1[5] d[3] b0[3] g0[7] b1[3] g1[7] g0[0] r0[2] g1[0] r1[2] d[2] b0[2] g0[6] b1[2] g1[6] b0[2] r0[1] b1[2] r1[1] d[1] b0[1] g0[5] b1[1] g1[5] b0[1] r0[0] b1[1] r1[0] d[0] b0[0] g0[4] b1[0] g1[4] b0[0] g0[1] b1[0] g1[1] table 6. rgb 16-bit muliplexed mode idf# format 2 16-bit rgb (16-8) pixel# p0a p0b p1a p1b bus data d[15] g0[7] a0[7] g1[7] r1[7] d[14] g0[6] a0[6] g1[6] r1[6] d[13] g0[5] a0[5] g1[5] r1[5] d[12] g0[4] a0[4] g1[4] r1[4] d[11] g0[3] a0[3] g1[3] r1[3] d[10] g0[2] a0[2] g1[2] r1[2] d[9] g0[1] a0[1] g1[1] r1[1] d[8] g0[0] a0[0] g1[0] r1[0] d[7] b0[7] r0[7] b1[7] a1[7] d[6] b0[6] r0[6] b1[6] a1[6] d[5] b0[5] r0[5] b1[5] a1[5] d[4] b0[4] r0[4] b1[4] a1[4] d[3] b0[3] r0[3] b1[3] a1[3] d[2] b0[2] r0[2] b1[2] a1[2] d[1] b0[1] r0[1] b0[1] a1[1] d[0] b0[0] r0[0] b0[0] a1[0] table 7. ycrcb multiplexed mode idf# format 9 ycrcb 8-bit pixel# p0a p0b p1a p1b p2a p2b p3a p3b bus data d[7] cb0[7] y0[7] cr0[7] y1[7] cb2[7] y2[7] cr2[7] y3[7] d[6] cb0[6] y0[6] cr0[6] y1[6] cb2[6] y2[6] cr2[6] y3[6] d[5] cb0[5] y0[5] cr0[5] y1[5] cb2[5] y2[5] cr2[5] y3[5] d[4] cb0[4] y0[4] cr0[4] y1[4] cb2[4] y2[4] cr2[4] y3[4] d[3] cb0[3] y0[3] cr0[3] y1[3] cb2[3] y2[3] cr2[3] y3[3] d[2] cb0[2] y0[2] cr0[2] y1[2] cb2[2] y2[2] cr2[2] y3[2] d[1] cb0[1] y0[1] cr0[1] y1[1] cb2[1] y2[1] cr2[1] y3[1] d[0] cb0[0] y0[0] cr0[0] y1[0] cb2[0] y2[0] cr2[0] y3[0]
201-0000-026 rev 2.1, 8/2/99 11 chrontel ch7006c when idf = 9 (ycrcb 8-bit mode), h and v sync signals can be embedded into the data stream. in this mode, the embedded sync will follow the ccir656 convention, and the first byte of the ?video timing reference code ? will be assumed to occur when a cb sample would occur if the video stream was continuous. this is delineated in table8 shown below. . in this mode the s[7.0} contains the following data: s[6] = f = 1 during field 2, 0 during field 1 s[5] = v = 1 during field blanking, 0 elsewhere s[4] = h = 1 during eav (the synchronization reference at the end of active video) 0 during sav (the synchronization reference at the start of active video) bits s[7] and s[3-0] are ignored. figure 6: multiplexed pixel data transfer mode (idf = 6) table 8. ycrcb multiplexed mode with embedded syncs idf# format 9 ycrcb 8-bit pixel# p0a p0b p1a p1b p2a p2b p3a p3b bus data d[7] ff 00 00 s[7] cb2[7] y2[7] cr2[7] y3[7] d[6] ff 00 00 s[6] cb2[6] y2[6] cr2[6] y3[6] d[5] ff 00 00 s[5] cb2[5] y2[5] cr2[5] y3[5] d[4] ff 00 00 s[4] cb2[4] y2[4] cr2[4] y3[4] d[3] ff 00 00 s[3] cb2[3] y2[3] cr2[3] y3[3] d[2] ff 00 00 s[2] cb2[2] y2[2] cr2[2] y3[2] d[1] ff 00 00 s[1] cb2[1] y2[1] cr2[1] y3[1] d[0] ff 00 00 s[0] cb2[0] y2[0] cr2[0] y3[0] table 9. rgb 8-bit multiplexed mode (24-bit color) idf# format 6 rgb 8-bit pixel# p0a p0b p0c p1a p1b p1c p2a p2b p2c bus data d[7] b0[7] g0[7] r0[7] b1[7] g1[7] r1[7] b2[7] g2[7] r2(7) d[6] b0[6] g0[6] r0[6] b1[6] g1[6] r1[6] b2[6] g2[6] r2(6) d[5] b0[5] g0[5] r0[5] b1[5] g1[5] r1[5] b2[5] g2[5] r2(5) d[4] b0[4] g0[4] r0[4] b1[4] g1[4] r1[4] b2[4] g2[4] r2(4) d[3] b0[3] g0[3] r0[3] b1[3] g1[3] r1[3] b2[3] g2[3] r2(3) d[2] b0[2] g0[2] r0[2] b1[2] g1[2] r1[2] b2[2] g2[2] r2(2) d[1] b0[1] g0[1] r0[1] b1[1] g1[1] r1[1] b2[1] g2[1] r2(1) d[0] b0[0] g0[0] r0[0] b1[0] g1[0] r1[0] b2[0] g2[0] r2(0) t hsw pixel data pout/ xclk hsync t hd p0a p0b p0c p1a p1b p1c t ph3 t p3 t hp3 t sp3 d[7:0]
chrontel ch7006c 12 *patent number 5,874,846 201-0000-026 rev 2.1, 8/2/99 functional description the ch7006 is a tv-output companion chip to graphics controllers providing digital output in either yuv or rgb format. this solution involves both hardware and software elements which work together to produce an optimum tv screen image based on the original computer generated pixel data. all essential circuitry for this conversion are integrated onchip. onchip circuitry includes memory, memory control, scaling, pll, dac, filters, and ntsc/pal encoder. all internal signal processing, including ntsc/pal encoding, is performed using digital techniques to ensure that the high-quality video signals are not affected by drift issues associated with analog components. no additional adjustment is required during manufacturing. ch7006 is ideal for pc motherboards, web browsers, or vga add-in boards where a minimum of discrete support components (passive components, parallel resonance 14.31818 mhz crystal) are required for full operation. architectural overview the ch7006 is a complete tv output subsystem which uses both hardware and software elements to produce an image on tv which is virtually identical to the image that would be displayed on a monitor. simply creating a compatible tv output from a vga input involves a relatively straightforward process. this process includes a standard conversion from rgb to yuv color space, converting from a non-interlaced to an interlaced frame sequence, and encoding the pixel stream into ntsc or pal compliant format. however, creating an optimum computer-generated image on a tv screen involves a highly sophisticated process of scaling, deflickering, and filtering. this results in a compatible tv output that displays a sharp and subtle image, of the right size, with minimal artifacts from the conversion process. as a key part of the overall system solution, the ch7006 software establishes the correct framework for the vga input signal to enable this process. once the display is set to a supported resolution (either 640x480 or 800x600), the ch7006 software may be invoked to establish the appropriate tv output display. the software then programs the various timing parameters of the vga controller to create an output signal that will be compatible with the chosen resolution, operating mode, and tv format. adjustments performed in software include pixel clock rates, total pixels per line, and total lines per frame. by performing these adjustments in software, the ch7006 can render a superior tv image without the added cost of a full frame buffer memory ? normally used to implement features such as scaling and full synchronization. the ch7005 hardware accepts digital rgb or ycrcb inputs, which are latched in synchronization with the pixel clock. these inputs are then color-space converted into yuv in 4-2-2 format, and stored in a line buffer memory. the stored pixels are fed into a block where scan-rate conversion, underscan scaling and 2-line, 3-line, 4-line and 5- line vertical flicker filtering are performed. the scan-rate converter transforms the vga horizontal scan-rate to either ntsc or pal scan rates; the vertical flicker filter eliminates flicker at the output while the underscan scaling reduces the size of the displayed image to fit onto a tv screen. the resulting yuv signals are filtered through digital filters to minimize aliasing problems. the digital encoder receives the filtered signals and transforms them to composite and s-video outputs, which are converted by the three 9-bit dacs into analog outputs. color burst generation* the ch7006 allows the subcarrier frequency to be accurately generated from a 14.31818 mhz crystal oscillator, leaving the subcarrier frequency independent of the sampling rate. as a result, the ch7006 may be used with any vga chip (with an appropriate digital interface) since the ch7006 subcarrier frequency can be generated without being dependent on the precise pixel rates of vga controllers. this feature is a significant benefit, since even a 0.01% subcarrier frequency variation may be enough to cause some television monitors to lose color lock. in addition, the ch7006 has the capability to genlock the color burst signal to the vga horizontal sync frequency, which enables a fully synchronous system between the graphics controller and the television. when genlocked, the ch7006 can also stop ?dot crawl ? motion (for composite mode operation in ntsc modes) to eliminate the annoyance of moving borders. both of these features are under programmable control through the register set. display modes the ch7006 display mode is controlled by three independent factors: input resolution, tv format, and scale factor, which are programmed via the display mode register. it is designed to accept input resolutions of 640x480, 800x600, 640x400 (including 320x200 scan-doubled output), 720x400, and 512x384. it is designed to support
201-0000-026 rev 2.1, 8/2/99 13 chrontel ch7006c display modes (continued) t is disigned to support output to either ntsc or pal television formats. the ch7005 provides interpolated scaling with selectable factors of 5:4, 1:1, 7:8, 5:6, 3:4 and 7:10 in order to support adjustable overscan or underscan operation when displayed on a tv. this combination of factors results in a matrix of useful operating modes which are listed in detail in table10 . (1) note: percent underscan is a calculated value based on average viewable lines on each tv format, assuming an average tv over- scan of 10%. (negative values) indicate modes which are operating in underscan. for ntsc: 480 active lines - 10% (overscan) = 432 viewable lines (average) for pal: 576 active lines - 10% (overscan) = 518 viewable lines (average) the inclusion of multiple levels of scaling for each resolution have been created to enable optimal use of the ch7006 for different application needs. in general, underscan (modes where percent overscan is negative provides an image that is viewable in its entirety on screen; it should be used as the default for most applications (e.g., viewing text screens, operating games, running productivity applications and working within windows). overscanning provides an image that extends past the edges of the tv screen, exactly like normal television programs and movies appear on tv, and is only recommended for viewing movies or video clips coming from the computer. in addition to the above mode table, the ch7006 also support interlaced input modes, both in ccir 656 and proprietary formats (see display mode register section.) flicker filter and text enhancement the ch7006 integrates an advanced 2-line, 3-line, 4-line and 5-line (depending on mode) vertical deflickering filter circuit to help eliminate the flicker associated with interlaced displays. this flicker circuit provides an adaptive filter algorithm for implementing flicker reduction with selections of high, medium or low flicker content for both luma and chroma channels (see register descriptions). in addition, a special text enhancement circuit incorporates table 10. ch7006 display modes tv format standard input (active) resolution scale factor active tv lines percent (1) overscan pixel clock horizontal total vertical total ntsc 640x480 1:1 480 10% 24.671 784 525 ntsc 640x480 7:8 420 (3%) 28.196 784 600 ntsc 640x480 5:6 400 (8%) 30.210 800 630 ntsc 800x600 5:6 500 16% 39.273 1040 630 ntsc 800x600 3:4 450 4% 43.636 1040 700 ntsc 800x600 7:10 420 (3%) 47.832 1064 750 ntsc 640x400 5:4 500 16% 21.147 840 420 ntsc 640x400 1:1 400 (8%) 26.434 840 525 ntsc 640x400 7:8 350 (19%) 30.210 840 600 ntsc 720x400 5:4 500 16% 23.790 945 420 ntsc 720x400 1:1 400 (8%) 29.455 936 525 ntsc 512x384 5:4 480 10% 20.140 800 420 ntsc 512x384 1:1 384 (11%) 24.671 784 525 pal 640x480 5:4 600 14% 21.000 840 500 pal 640x480 1:1 480 (8%) 26.250 840 625 pal 640x480 5:6 400 (29%) 31.500 840 750 pal 800x600 1:1 600 14% 29.500 944 625 pal 800x600 5:6 500 (4%) 36.000 960 750 pal 800x600 3:4 450 (15%) 39.000 936 836 pal 640x400 5:4 500 (4%) 25.000 1000 500 pal 640x400 1:1 400 (29%) 31.500 1008 625 pal 720x400 5:4 500 (4%) 28.125 1125 500 pal 720x400 1:1 400 (29%) 34.875 1116 625 pal 512x384 5:4 480 (8%) 21.000 840 500 pal 512x384 1:1 384 (35%) 26.250 840 625
chrontel ch7006c 14 201-0000-026 rev 2.1, 8/2/99 display modes (continued) additional filtering for enhancing the readability of text. these modes are fully programmable via i 2 c under the flicker filter register. internal voltage reference an onchip bandgap circuit is used in the dac to generate a reference voltage which, in conjunction with a reference resistor at pin rset, and register controlled divider, sets the output ranges of the dacs. the ch7006 bandgap reference voltage is 1.235 volts nominal for ntsc or pal-m, or 1.317 volts nominal for pal or ntsc-j, which is determined by idf register bit 6 (dacg bit). the recommended value for the reference resistor iset is 360 ohms (though this may be adjusted in order to achieve a different output level). the gain setting for dac output is 1/48 th . therefore, for each dac, the current output per lsb step is determined by the following equation: i lsb = v(rset)/rset reference resistor * 1/gain for dacg=0, this is: i lsb = 1.235/360 * 1/48 = 71.4 m a (nominal) for dacg=1, this is: i lsb = 1.317/360 * 1/48 = 76.2 m a (nominal) power management the ch7006 supports five operating states including normal [on], power down, full power down, s-video off, and composite off to provide optimal power consumption for the application involved. using the programmable power down modes accessed over the i 2 c port, the ch7006 may be placed in either normal state, or any of the four power managed states, as listed below (see ? power management register ? under the register descriptions section for programming information). to support power management, a tv sensing function (see ?connection detect register ? under the register descriptions section) is provided, which identifies whether a tv is connected to either s-video or composite. this sensing function can then be used to enter into the appropriate operating state (e.g., if tv is sensed only on composite, the s-video off mode could be set by software). luminance and chrominance filter options the ch7006 contains a set of luminance filters to provide a controllable bandwidth output on both cvbs and s-video outputs. all values are completely programmable via the video bandwidth register. for all graphs shown, the horizontal axis is frequency in mhz, and the vertical axis is attenuation in dbs. the composite luminance and chrominance video bandwidth output is shown in table12 . table 11. power management operating state functional description normal (on): in the normal operating state, all functions and pins are active power down: in the power-down state, most pins and circuitry are disabled.the ds/bco pin will continue to provide either the vco divided by k3, or 14.318 mhz out when selected as an output, and the p-out pin will continue to output a clock reference when in master clock mode. s-video off: power is shut off to the unused dacs associated with s-video outputs. composite off: in composite-off state, power is shut off to the unused dac associated with cvbs output. full power down: in this power-down state, all but the i 2 c circuits are disabled. this places the ch7006 in its lowest power consumption mode.
201-0000-026 rev 2.1, 8/2/99 15 chrontel ch7006c vbi pass-through support the ch7006 provides the ability to pass-through data with minimal filtering, on vertical blanking lines 10-21 for intercast or close captioned applications (see register descriptions). the composite luminance and chrominance frequency response is depicted in figure7 through 9 . table 12. video bandwidth mode chrominance luminance bandwidth with sin(x) /x (mhz) cvbs s-video s-video cbw[1:0] ycv ysv[1:0], ypeak = 0 ysv[1:0], ypeak = 1 00 01 10 11 0 1 00 01 1x 00 01 1x 0 0.62 0.68 0.80 0.95 2.26 3.37 2.26 3.37 5.23 2.57 4.44 5.23 1 0.78 0.85 1.00 1.18 2.82 4.21 2.82 4.21 6.53 3.21 5.56 6.53 2 0.53 0.58 0.68 0.81 1.93 2.87 1.93 2.87 4.46 2.19 3.79 4.46 3 0.65 0.71 0.83 0.99 2.36 3.52 2.36 3.52 5.46 2.68 4.64 5.46 4 0.83 0.91 1.07 1.27 3.03 4.51 3.03 4.51 7.00 3.44 5.95 7.00 5 1.03 1.13 1.32 1.57 3.75 5.59 3.75 5.59 8.68 4.27 7.38 8.68 6 0.70 0.77 0.90 1.07 2.56 3.81 2.56 3.81 5.92 2.91 5.04 5.92 7 0.87 0.95 1.12 1.33 3.17 4.72 3.17 4.72 7.33 3.60 6.23 7.33 8 0.74 0.81 0.95 1.13 2.69 4.01 2.69 4.01 6.22 3.06 5.29 6.22 9 0.93 1.02 1.20 1.42 3.39 5.05 3.39 5.05 7.84 3.85 6.67 7.84 10 0.63 0.68 0.80 0.95 2.28 3.39 2.28 3.39 5.26 2.59 4.48 5.26 11 0.78 0.86 1.00 1.19 2.84 4.24 2.84 4.24 6.58 3.23 5.59 6.58 12 0.89 0.98 1.15 1.36 3.25 4.84 3.25 4.84 7.52 3.70 6.39 7.52 13 0.62 0.68 0.80 0.95 2.26 3.37 2.26 3.37 5.23 2.57 4.44 5.23 14 0.78 0.85 1.00 1.18 2.82 4.21 2.82 4.21 6.53 3.21 5.56 6.53 15 0.93 1.02 1.20 1.42 3.39 5.05 3.39 5.05 7.84 3.85 6.67 7.84 16 0.64 0.71 0.83 0.98 2.35 3.50 2.35 3.50 5.43 2.67 4.62 5.43 17 0.74 0.81 0.95 1.13 2.70 4.02 2.70 4.02 6.24 3.07 5.30 6.24 18 0.79 0.87 1.02 1.21 2.89 4.31 2.89 4.31 6.68 3.29 5.68 6.68 19 0.77 0.85 1.00 1.18 2.82 4.20 2.82 4.20 6.53 3.21 5.55 6.53 20 0.95 1.03 1.22 1.44 3.44 5.13 3.44 5.13 7.97 3.92 6.77 7.97 21 1.02 1.12 1.32 1.56 3.73 5.56 3.73 5.56 8.63 4.24 7.34 8.63 22 0.77 0.85 0.99 1.18 2.82 4.20 2.82 4.20 6.52 3.20 5.54 6.52 23 0.86 0.94 1.11 1.31 3.13 4.66 3.13 4.66 7.24 3.56 6.16 7.24 24 0.94 1.03 1.21 1.44 3.43 5.11 3.43 5.11 7.94 3.90 6.75 7.94 25 0.71 0.78 0.91 1.08 2.58 3.85 2.58 3.85 5.97 2.94 5.08 5.97 26 0.71 0.78 0.91 1.08 2.58 3.85 2.58 3.85 5.97 2.94 5.08 5.97 27 0.47 0.51 0.60 0.71 1.70 2.53 1.70 2.53 3.92 1.93 3.34 3.92 28 0.38 0.41 0.48 0.57 1.37 2.04 1.37 2.04 3.17 1.56 2.69 3.17
chrontel ch7006c 16 201-0000-026 rev 2.1, 8/2/99 luminance and chrominance filter options (continued) figure 7: composite luminance frequency response (ycv = 0) figure 8: s-video luminance frequency response (ysv = 1x, ypeak = 0) 0 1 2 3 4 5 6 7 8 9 10 11 12 42 36 30 24 18 12 6 0 < > ycvdb i n f , n i 10 6 0 -6 -12 -18 -24 - 30 -36 -42 1 2 3 4 7 8 9 10 11 12 (ycvdb ) n f n,i 10 6 6 0 5 < > ysvdb i f , n i 10 6 0 -6 -12 -18 -24 -30 -36 -42 0 1 2 3 4 5 6 7 8 9 10 11 (ysvdb ) n 12
201-0000-026 rev 2.1, 8/2/99 17 chrontel ch7006c luminance and chrominance filter options (continued) figure 9: chrominance frequency response 0 1 2 3 4 5 6 7 8 9 10 11 12 42 36 30 24 18 12 6 0 < > uvfirdb i n f , n i 10 6 0 -6 -12 -18 -24 -30 -36 -42 0 1 2 3 4 5 6 7 8 9 10 11 12 (uvfirdb ) n f n,i 10 6
chrontel ch7006c 18 201-0000-026 rev 2.1, 8/2/99 ntsc and pal operation composite and s-video outputs are supported in either ntsc or pal format. the general parameters used to characterize these outputs are listed in table13 and shown in figure10 . (see figure13 through 18 for illustrations of composite and s-video output waveforms.) ccir624-3 compliance the ch7006 is predominantly compliant with the recommendations called out in ccir624-3. the following are the only exceptions to this compliance: ? the frequencies of fsc, fh, and fv can only be guaranteed in master or pseudo-master modes, not in slave mode when the graphics device generates these frequencies. ? it is assumed that gamma correction, if required, is performed in the graphics device which establishes the color reference signals. ? all modes provide the exact number of lines called out for ntsc and pal modes respectively, except mode 21, which outputs 800x600 resolution, scaled by 3:4, to pal format with a total of 627 lines (vs. 625). ? chroma signal frequency response will fall within 10% of the exact recommended value. ? pulse widths and rise/fall times for sync pulses, front/back porches, and equalizing pulses are designed to approximate ccir624-3 requirements, but will fall into a range of values due to the variety of clock frequencies used to support multiple operating modes for this table and all subsequent figures, key values are: note: 1. iset = 360 ohms; v(iset) = 1.235v; 75 ohms doubly terminated load. 2. durations vary slightly in different modes due to the different clock frequencies used. 3. active video and black (f, g, h) times vary greatly due to different scaling ratios used in different modes. 4. black times (f and h) vary with position controls. table 13. ntsc/pal composite output timing parameters (in m s) symbol description level (mv) duration (us) ntsc pal ntsc pal a front porch 287 300 1.49 - 1.51 1.48 - 1.51 b horizontal sync 0 0 4.69 - 4.72 4.69 - 4.71 c breezeway 287 300 0.59 - 0.61 0.88 - 0.92 d color burst 287 300 2.50 - 2.53 2.24 - 2.26 e back porch 287 300 1.55 - 1.61 2.62 - 2.71 f black 340 300 0.00 - 7.50 0.00 - 8.67 g active video 340 300 37.66 - 52.67 34.68 - 52.01 h black 340 300 0.00 - 7.50 0.00 - 8.67
201-0000-026 rev 2.1, 8/2/99 19 chrontel ch7006c figure 10: ntsc / pal composite output figure 11: interlaced ntsc video timing a b c d e f g h 520 521 522 523 524 525 1 2 3 4 5 6 7 258 259 260 261 262 263 264 265 266 267 268 269 272 start of vsync analog field 1 analog field 2 8 9 270 271 520 521 522 523 524 525 1 2 3 4 5 6 7 258 259 260 261 262 263 264 265 266 267 268 269 272 start of vsync analog field 1 analog field 2 8 9 270 271 520 521 522 523 524 525 1 2 3 4 5 6 7 258 259 260 261 262 263 264 265 266 267 268 269 272 start of vsync analog field 1 analog field 2 8 9 270 271 pre-equalizing pulse interval reference sub-carrier phase color field 1 line vertical interval vertical sync pulse interva l post-equalizing pulse interval start of field 1 start of field 2 reference sub-carrier phase color field 2 reference sub-carrier phase color field 3 reference sub-carrier phase color field 4 start of field 3 start of field 4 t 1 +v t 2 +v t 3 +v 523 524 525 1 2 3 4 5 6 7 8 9 10 11 12 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 523 524 525 1 2 3 4 5 6 7 8 9 10 11 12 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
chrontel ch7006c 20 201-0000-026 rev 2.1, 8/2/99 figure 12: interlaced pal video timing 621 622 623 624 625 1 2 3 4 5 6 7 620 621 622 623 624 625 1 2 3 4 5 6 7 620 309 310 311 312 313 314 315 316 317 318 319 320 323 308 322 309 310 311 312 313 314 315 316 317 318 319 320 308 start of vsync analog field 1 analog field 2 analog field 3 analog field 4 burst blanking pal switch = 0, +v component burst phase = reference phase = 135 relative to u pal switch = 1, - v component burst phase = reference phase + 90 = 225 relative to u 8 9 10 321 8 9 10 323 322 321 621 622 623 624 625 1 2 3 4 5 6 7 620 621 622 623 624 625 1 2 3 4 5 6 7 620 309 310 311 312 313 314 315 316 317 318 319 320 323 308 322 309 310 311 312 313 314 315 316 317 318 319 320 308 start of vsync analog field 1 analog field 2 analog field 3 analog field 4 burst blanking pal switch = 0, +v component burst phase = reference phase = 135 relative to u pal switch = 1, - v component burst phase = reference phase + 90 = 225 relative to u 8 9 10 321 8 9 10 323 322 321 i n t e r v a l s 4 3 2 1
201-0000-026 rev 2.1, 8/2/99 21 chrontel ch7006c figure 13: ntsc y (luminance) output waveform (dacg = 0) figure 14: pal y (luminance) video output waveform (dacg = 1) color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k color/level ma v white 26.66 1.000 yellow 24.66 0.925 cyan 21.37 0.801 green 19.37 0.726 magenta 16.22 0.608 red 14.22 0.533 blue 11.08 0.415 black 9.08 0.340 blank 7.65 0.287 sync 0.00 0.000 color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k color/level ma v white 26.75 1.003 yellow 24.62 0.923 cyan 21.11 0.792 green 18.98 0.712 magenta 15.62 0.586 red 13.49 0.506 blue 10.14 0.380 blank/ black 8.00 0.300 sync 0.00 0.000
chrontel ch7006c 22 201-0000-026 rev 2.1, 8/2/99 figure 15: ntsc c (chrominance) video output waveform (dacg = 0) figure 16: pal c (chrominance) video output waveform (dacg = 1) color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k color/level ma v cyan/red 25.80 0.968 green/magenta 25.01 0.938 yellow/blue 22.44 0.842 peak burst 18.08 0.678 blank 14.29 0.536 yellow/blue 6.15 0.230 green/magenta 3.57 0.134 cyan/red 2.79 0.105 peak burst 10.51 0.394 3.579545 mhz color burst (9 cycles) color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k color/level ma v cyan/red 27.51 1.032 green/magenta 26.68 1.000 yellow/blue 23.93 0.897 peak burst 19.21 0.720 blank 15.24 0.572 yellow/blue 6.56 0.246 green/magenta 3.81 0.143 cyan/red 2.97 0.111 peak burst 11.28 0.423 4.433619 mhz color burst (10 cycles)
201-0000-026 rev 2.1, 8/2/99 23 chrontel ch7006c figure 17: composite ntsc video output waveform (dacg = 0) figure 18: composite pal video output waveform (dacg = 1) color/level ma v peak chrome 32.88 1.233 white 26.66 1.000 sync 0.00 0.000 peak burst 11.44 0.429 black 9.08 0.340 blank 7.65 0.281 peak burst 4.45 0.145 color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k 3.579545 mhz color burst (9 cycles) color/level ma v peak chrome 33.31 1.249 white 26.75 1.003 sync 0.00 0.000 peak burst 11.97 0.449 blank/black 8.00 0.300 peak burst 4.04 0.151 color bars: w h i t e y e l l o w c y a n m a g e n t a g r e e n r e d b l u e b l a c k 4.433619 mhz color burst (10 cycles)
chrontel ch7006c 24 201-0000-026 rev 2.1, 8/2/99 i 2 c port operation the ch7006 contains a standard i 2 c control port, through which the control registers can be written and read. this port is comprised of a two-wire serial interface, pins sd (bidirectional) and sc, which can be connected directly to the sdb and scb buses as shown in figure19 . the serial clock line (sc) is input only and is driven by the output buffer of the master device (also shown in figure19 ). the ch7006 acts as a slave, and generation of clock signals on the bus is always the responsibility of the master device. when the bus is free, both lines are high. the output stages of devices connected to the bus must have an open-drain or open-collector to perform the wired-and function. data on the bus can be transferred up to 400 kbit/s. figure 19: connectioin of devices to bus electrical characteristics for bus devices the electrical specifications of the bus devices ? inputs and outputs and the characteristics of the bus lines connected to them are shown in figure19 . a pull-up resistor (r p ) must be connected to a 3.3v 10% supply. the ch7006 is a device with input levels related to dvdd. maximum and minimum values of pull-up resistor (r p ) the value of r p depends on the following parameters: ? supply voltage ? bus capacitance ? number of devices connected (input current + leakage current = i input ) the supply voltage limits the minimum value of resistor r p due to the specified minimum sink current of 2ma at vol max = 0.4 v for the output stages: r p >= (v dd ? 0.4) / 2 (r p in k w) the bus capacitance is the total capacitance of wire, connections and pins. this capacitance limits the maximum value of r p due to the specified rise time. the equation for rp is shown below: r p <= 10 3 /c (where: r p is in k w and c, the total capacitance, is in pf) the maximum high level input current of each input/output connection has a specified maximum value of 10 m a. due to the desired noise margin of 0.2v dd for the high level, this input current limits the maximum value of r p . the r p limit depends on v dd and is shown below: r p <= (100 x v dd )/ i input (where: r p is in k w and i input is in m a) sclk in2 data in2 datan2 out scb (serial clock bus) sdb (serial data bus) +dvdd r p slave sclk in1 data in1 datan2 out slave sclk out from master data in master datan2 out master bus master sc sd
201-0000-026 rev 2.1, 8/2/99 25 chrontel ch7006c transfer protocol both read and write cycles can be executed in ?alternating ? and ?auto-increment ? modes. alternating mode expects a register address prior to each read or write from that location (i.e., transfers alternate between address and data). auto-increment mode allows you to establish the initial register location, then automatically increments the register address after each subsequent data access (i.e., transfers will be address, data data data...). a basic serial port transfer protocol is shown in figure20 and described below. figure 20: serial port transfer protocol 1. the transfer sequence is initiated when a high-to-low transition of sd occurs while sc is high; this is the ?start ? condition. transitions of address and data bits can only occur while sc is low. 2. the transfer sequence is terminated when a low-to-high transition of sd occurs while sc is high; this is the ?stop ? condition. 3. upon receiving the first start condition, the ch7006 expects a device address byte (dab) from the master device. the value of the device address is shown in the dab data format below. 4. after the dab is received, the ch7006 expects a register address byte (rab) from the master. the format of the rab is shown in the rab data format below (note that b7 is not used). device address byte (dab) r/w read/write indicator ?0 ?: master device will write to the ch7006 at the register location specified by the address ar[5:0] ?1 ?: master device will read from the ch7006 at the register location specified by the address ar[5:0]. register address byte (rab) b7 b6 b5 b4 b3 b2 b1 b0 1 1 1 0 1 0 1 r/w b7 b6 b5 b4 b3 b2 b1 b0 1 autoinc ar[5] ar[4] ar[3] ar[2] ar[1] ar[0] sd sc 1 - 8 9 data 1 ack condition start condition stop ch7006 1 - 8 data n 9 ack ch7006 ch7006 ch7 device id 8 r/w* 9 ack i 2 c acknowledge acknowledge acknowledge
chrontel ch7006c 26 201-0000-026 rev 2.1, 8/2/99 transfer protocols (continued) autoinc register address auto-increment - to facilitate sequential r/w of registers. ?1 ?: auto-increment enabled (auto-increment mode). write: after writing data into a register, the address register will automatically be incremented by one. read: before loading data from a register to the on-chip temporary register (getting ready to be serially read), the address register will automatically be incremented by one. however, for the first read after an rab, the address register will not be changed. ?0 ?: auto-increment disabled (alternating mode). write: after writing data into a register, the address register will remain unchanged until a new rab is written. read: before loading data from a register to the on-chip temporary register (getting ready to be serially read), the address register will remain unchanged. ar[5:0] specifies the address of the register to be accessed. this register address is loaded into the address register of the ch7006. the r/w access, which follows, is directed to the register specified by the content stored in the address register. the following two sections describe the operation of the serial interface for the four combinations of r/w = 0,1 and autoinc = 0,1. ch7006 write cycle protocols (r/w = 0) data transfer with acknowledge is required. the acknowledge-related clock pulse is generated by the master- transmitter. the master-transmitter releases the sd line (high) during the acknowledge clock pulse. the slave- receiver must pull down the sd line, during the acknowledge clock pulse, so that it remains stable low during the high period of the clock pulse. the ch7006 always acknowledges for writes (see figure21 ). note that the resultant state on sd is the wired-and of data outputs from the transmitter and receiver. figure 21: acknowledge on the bus figure22 shows two consecutive alternating write cycles for autoinc = 0 and r/w = 0. the byte of information, following the register address byte (rab), is the data to be written into the register specified by ar[5:0]. if autoinc = 0, then another rab is expected from the master device, followed by another data byte, and so on. sc from master sd data output by the ch7006 start condition 2 sd data output by master-transmitter 1 8 9 not acknowledge acknowledge clock pulse for acknowledgment
201-0000-026 rev 2.1, 8/2/99 27 chrontel ch7006c note: the acknowledge is from the ch7006 (slave). figure 22: alternating write cycles if autoinc = 1, then the register address pointer will be incremented automatically and subsequent data bytes will be written into successive registers without providing an rab between each data byte. an auto-increment write cycle is shown in figure23. . note: the acknowledge is from the ch7006 (slave). figure 23: auto-increment write cycle when the auto-increment mode is enabled (autoinc is set to 1), the register address pointer continues to increment for each write cycle until ar[5:0] = 3f (3f is the address of the address register). the next byte of information represents a new auto-sequencing ?starting address, ? which is the address of the register to receive the next byte. the auto-sequencing then resumes based on this new ?starting address. ? the auto-increment sequence can be terminated any time by either a ?stop ? or ?restart ? condition. the write operation can be terminated with a ?stop ? condition. ch7006 read cycle protocols (r/w = 1) if a master-receiver is involved in a transfer, it must signal the end of data to the slave-transmitter by not generating an acknowledge on the last byte that was clocked out of the slave. the slave-transmitter ch7006 releases the data line to allow the master to generate the stop condition or the restart condition. to read the content of the registers, the master device starts by issuing a ?start ? condition (or a ?restart ? condition). the first byte of data, after the start condition, is a dab with r/w = 0. the second byte is the rab with ar[5:0], containing the address of the register that the master device intends to read from in ar[5:0]. the master device should then issue a ?restart ? condition ( ?restart ? = ?start, ? without a previous ?stop ? condition). the first byte of data, after this restart condition, is another dab with r/w=1, indicating the master ?s intention to read data hereafter. the master then reads the next byte of data (the content of the register specified in the rab). if autoinc = 0, then another restart condition, followed by another dab with r/w = 0 and rab, is expected from the master device. the master device then issues another restart, followed by another dab. after that, the master may read another data byte, and so on. in summary, a restart condition, followed by a dab, must be produced by the master before each of the rab, and before each of the data read events. two consecutive alternating read cycles are shown in figure24 . sd sc 1 - 8 rab 9 ack condition start condition stop 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge ch7006 acknowledge 1 - 8 rab 9 ack ch7006 acknowledge 1 - 8 data 9 ack i 2 c 1 - 8 data 9 ack ch7006 acknowledge ch7006 acknowledge sd sc 1 - 8 9 rab n ack start stop ch7006 acknowledge 1 - 8 data n 9 1 - 8 9 ack data n+1 ack ch7006 acknowledge ch7006 acknowledge ch7006 acknowledge 1 - 7 device id 8 r/w* 9 ack i 2 c condition condition
chrontel ch7006c 28 201-0000-026 rev 2.1, 8/2/99 transfer protocols (continued) . figure 24: alternating read cycle if autoinc = 1, then the address register will be incremented automatically and subsequent data bytes can be read from successive registers, without providing a second rab figure 25: auto-increment read cycle when the auto-increment mode is enabled (autoinc is set to 1), the address register will continue incrementing for each read cycle. when the content of the address register reaches 2a, it will wrap around and start from 00h again. the auto increment sequence can be terminated by either a ?stop ? or ?restart ? condition. the read operation can be terminated with a ?stop ? condition. figure25 shows an auto-increment read cycle terminated by a stop or restart condition. ch7006 acknowledge ch7006 acknowledge sd sc 1 - 8 rab 1 9 10 ack restart condition start condition stop condition master does not acknowledge 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge ch7006 acknowledge 1 - 8 data 1 9 ack 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge i 2 c i 2 c 10 restart condition 1 - 8 rab 2 9 10 ack restart condition 1 - 7 device id 8 r/w* 9 ack 1 - 8 data 2 9 ack 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge i 2 c i 2 c master does not acknowledge master acknowledge sd sc 1 - 8 rab n 9 10 ack restart condition start condition stop condition master does not acknowledge just before stop condition 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge ch7006 acknowledge 1 - 8 data n 9 ack 1 - 7 device id 8 r/w* 9 ack ch7006 acknowledge 1 - 8 data n+1 9 ack i 2 c
201-0000-026 rev 2.1, 8/2/99 29 chrontel ch7006c registers and programming the ch7006 is a fully programmable device, providing for full functional control through a set of registers accessed from the i 2 c port. the ch7006 contains a total of 37 registers, which are listed in figure14 and described in detail under register descriptions . detailed descriptions of operating modes and their effects are contained in the previous section, functional description. an addition (+) sign in the bits column below signifies that the parameter contains more than 8 bits, and the remaining bits are located in another register. table 14. register map register symbol address bits functional summary display mode dmr 00h 8 display mode selection flicker filter ffr 01h 6 flicker filter mode selection video bandwidth vbw 03h 7 luma and chroma filter bandwidth selection input data format idf 04h 7 data format and bit-width selections clock mode cm 06h 8 sets the clock mode to be used start active video sav 07h 8+ active video delay setting position overflow po 08h 3 msb bits of position values black level blr 09h 8 black level adjustment input latch clock edge select horizontal position hpr 0ah 8+ enables horizontal movement of displayed image on tv vertical position vpr 0bh 8+ enables vertical movement of displayed image on tv sync polarity spr 0dh 4 determines the horizontal and vertical sync polarity power management pmr 0eh 5 enables power saving modes connection detect cdr 10h 4 detection of tv presence contrast enhancement ce 11h 3 contrast enhancement setting pll m and n extra bits mne 13h 5 contains the msb bits for the m and n pll values pll-m value pllm 14h 8+ sets the pll m value - bits (7:0) pll-n value plln 15h 8+ sets the pll n value - bits (7:0) buffered clock bco 17h 6 determines the clock output at pin 41 subcarrier frequency adjust fsci 18h -1fh 4 or 8 each determines the subcarrier frequency pll and memory control pllc 20h 6 controls for the pll and memory sections civ control civc 21h 5 control of civ value calculated fsc increment value civ 22h - 24h 8 each readable register containing the calculated subcarrier increment value version id vid 25h 8 device version number test tr 26h - 29h 30 reserved for test (details not included herein) address ar 3fh 6 current register being addressed
chrontel ch7006c 30 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) table 15. i 2 c alternate register map ( note: controls available only by special arrangement) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 00h ir2 ir1 iro vos1 vos0 sr2 sr1 sr0 01h fc1 fc0 fy1 fy0 ft1 ft0 02h 03h flff cvbw cbw1 cbw0 ypeak ysv1 ysv0 ycv 04h dacg rgbbp idf3 idf2 idf1 idf0 05h 06h cfrb m/s* reserved mcp xcm1 xcm0 pcm1 pcm0 07h sav7 sav6 sav5 sav4 sav3 sav2 sav1 sav0 08h sav8 hp8 vp8 09h bl7 bl6 bl5 bl4 bl3 bl2 bl1 bl0 0ah hp7 hp6 hp5 hp4 hp3 hp2 hp1 hp0 0bh vp7 vp6 vp5 vp4 vp3 vp2 vp1 vp0 0ch 0dh des syo vsp hsp 0eh scart resetb pd2 pd1 pd0 0fh 10h yt ct cvbst sense 11h ce2 ce1 ce0 12h 13h sne spe n9 n8 m8 14h m7 m6 m5 m4 m3 m2 m1 m0 15h n7 n6 n5 n4 n3 n2 n1 n0 16h 17h shf2 shf1 shf0 sco2 sco1 sco0 18h fsci31 fsci30 fsci29 fsci28 19h fsci27 fsci26 fsci25 fsci24 1ah fsci23 fsci22 fsci21 fsci20 1bh p-outp fsci19 fscl18 fscl17 fscl16 1ch dsen fsci15 fscl14 fscl13 fsci12 1dh fsci11 fscl10 fscl9 fsci8 1eh fsci7 fsci6 fsci5 fsci4 1fh fsci3 fsci2 fsci1 fsci0 20h pllcpl pllcap plls pll5vd pll5va mem5v 21h civ25 civ24 clvh1 clvh0 aclv 22h civ23 civ22 civ21 civ20 civ19 civ18 civ17 civ16 23h civ15 civ14 civ13 civ12 civ11 civ10 civ9 civ8 24h civ7 civ6 civ5 civ4 civ3 civ2 civ1 civo 25h vid7 vid6 vid5 vid4 vid3 vid2 vid1 vid0 26h ts3 ts2 ts1 ts0 rsa bst nst te 27h ms2 ms1 mso mtd ylm8 clm8 28h ylm7 ylm6 ylm5 ylm4 ylm3 ylm2 ylm1 ylm0 29h clm7 clm6 clm5 clm4 clm3 clm2 clm1 clm0 3fh reserved reserved ar5 ar4 ar3 ar2 ar1 ar0
201-0000-026 rev 2.1, 8/2/99 31 chrontel ch7006c register descriptions (continued) display mode register a ddress: 00h bits: 8 this register provides programmable control of the ch7006 display mode, including input resolution (ir[2:0]), output tv standard (vos[1:0]), and scaling ratio (sr[2:0]). the mode of operation is determined according to the table below (default is 640x480 input, ntsc output, 7/8 ?s scaling). bit: 7 6 5 4 3 2 1 0 symbol: ir2 ir1 ir0 vos1 vos0 sr2 sr1 sr0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 1 1 0 1 0 1 0 table 16. display modes mode ir[2:0] vos [1:0] sr [2:0] input data format (active video) total pixels/line x total lines/frame output format scaling pixel clock (mhz) 0 000 00 000 512x384 840x500 pal 5/4 21.000000 1 000 00 001 512x384 840x625 pal 1/1 26.250000 2 000 01 000 512x384 800x420 ntsc 5/4 20.139860 3 000 01 001 512x384 784x525 ntsc 1/1 24.671329 4 001 00 000 720x400 1125x500 pal 5/4 28.125000 5 001 00 001 720x400 1116x625 pal 1/1 34.875000 6 001 01 000 720x400 945x420 ntsc 5/4 23.790210 7 001 01 001 720x400 936x525 ntsc 1/1 29.454545 8 010 00 010 640x400 1000x500 pal 5/4 25.000000 9 010 00 001 640x400 1008x625 pal 1/1 31.5000000 10 010 01 000 640x400 840x420 ntsc 5/4 21.146853 11 010 01 001 640x400 840x525 ntsc 1/1 26.433566 12 010 01 010 640x400 840x600 ntsc 7/8 30.209790 13 011 00 000 640x480 840x500 pal 5/4 21.000000 14 011 00 001 640x480 840x625 pal 1/1 26.250000 15 011 00 011 640x480 840x750 pal 5/6 31.5000000 16 011 01 001 640x480 784x525 ntsc 1/1 24.671329 17 011 01 010 640x480 784x600 ntsc 7/8 28.195804 18 011 01 011 640x480 800x630 ntsc 5/6 30.209790 19 100 00 001 800x600 944x625 pal 1/1 29.500000 20 100 00 011 800x600 960x750 pal 5/6 36.0000000 21 100 00 100 800x600 936x836 pal 3/4 39.000000 22 100 01 011 800x600 1040x630 ntsc 5/6 39.272727 23 100 01 100 800x600 1040x700 ntsc 3/4 43.636364 24 100 01 101 800x600 1064x750 ntsc 7/10 47.832168 25* 101 00 001 720x576 864x625 pal 1/1 13.500000 26* 101 01 001 720x480 858x525 ntsc 1/1 13.500000 27* 110 00 001 800x500 1135x625 pal 1/1 17.734375 28* 110 01 001 640x400 910x525 ntsc 1/1 14.318182 * interlaced modes of operation. (for those modes, some functions will be bypassed. for details, please contact the application department.)
chrontel ch7006c 32 201-0000-026 rev 2.1, 8/2/99 flicker filter register symbol: ffr address: 01h bits: 6 the flicker filter register provides for adjusting the operation of the various filters used in rendering the on-screen image. adjusting settings between minimal and maximal values enables optimization between sharpness and flicker content. the fc[1:0] bits determine the settings for the chroma channel. the ft[1:0] bits determine the settings for the text enhancement circuit. the fy[1:0] bits determine the settings for the luma channel. in addition, the chroma channel filtering includes a setting to enable the chroma dot crawl reduction circuit. note: when writing to register o1h, fy[1.0] is bits 3:2. ft[1:0] is bits 1:0. when reading from the register o1h, fy [1:0] is bits 1:0 and ft[1:0] is bits 3:2. register descriptions (continued) vos[1:0] 00 01 10 11 output format pal ntsc pal-m ntsc-j bit: 7 6 5 4 3 2 1 0 symbol: fc1 fc0 fy1 fy0 ft1 ft0 type: r/w r/w r/w r/w r/w r/w default: 1 1 0 0 1 0 table 17. flicker filter settings fy[1:0] settings for luma channel 00 minimal flicker filtering 01 slight flicker filtering 10 maximum flicker filtering 11 invalid ft[1:0] settings for text enhancement circuit 00 maximum text enhancement 01 slight text enhancement 10 minimum text enhancement 11 invalid fc[1:0] settings for chroma channel 00 minimal flicker filtering 01 slight flicker filtering 10 maximum flicker filtering 11 enable chroma dotcrawl reduction
201-0000-026 rev 2.1, 8/2/99 33 chrontel ch7006c register descriptions (continued) video bandwidth register symbol: vbw address: 03h bits: 7 this register enables the selection of alternative filters for use in the luma and chroma channels. there are currently four filter options defined for the chroma channel, 4 filter options in the s-video luma channel and two filter options in the composite luma channel. the table18 and table19 below show the various settings. bit 6 (cvbw) outputs the s-video luma signal on both the s-video luma output and the cvbs output. a "1" in this location enables the output of a black and white image on composite, thereby eliminating the degrading effects of the color signal (such as dot crawl or false colors), which is useful for viewing text with high accuracy. bit 7 (flff) controls the flicker filter used in the 7/10 ?s scaling modes. in these scaling modes, setting flff to 1 causes a five line flicker filter to be used. the default setting of 0 uses a four line flicker filter. bit: 7 6 5 4 3 2 1 0 symbol: flff cvbw cbw1 cbw0 ypeak ysv1 ysv0 ycv type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0 0 0 table 18. luma filter bandwidth ycv luma composite video filter adjust 0 low bandwidth 1 high bandwidth ysv[1:0] luma s-video filter adjust 00 low bandwidth 01 medium bandwidth 10 high bandwidth 11 reserved (decode this and handle the same as 10) ypeak disables the y-peaking circuit 0 disables the peaking filter in luma s-video channel 1 enables the peaking filter in luma s-video channel table 19. chroma filter bandwidth cbw[1:0] luma composite video filter adjust 0 0 low bandwidth 0 1 medium bandwidth 1 0 med-high bandwidth 1 1 high bandwidth
chrontel ch7006c 34 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) input data format register symbol: idf address: 04h bits: 7 this register sets the variables required to define the incoming pixel data stream. rgbbp (bit 5): setting this bit enables the rgb pass-through mode. setting this bit to a 1 causes the input rgb signal to be directly output at the dacs (subject to a pipeline delay). if rgbbp=0, the bypass mode is disabled. dacg (bit 6): this bit controls the gain of the d/a converters. when dacg=0, the nominal dac current is 71 m a, which provides the correct levels for ntsc and pal-m. when dacg=1, the nominal dac current is 76 m a, which provides the correct levels for pal and ntsc-j. clock mode register symbol: cm address: 06h bits: 8 the setting of the clock mode bits determines the clocking mechanism used in the ch7006. the clock modes are shown in the table below. pcm controls the frequency of the pixel clock, and xcm identifies the frequency of the xclk input clock. bit: 7 6 5 4 3 2 1 0 symbol: dacg rgbbp idf3 idf2 idf1 idf0 type: r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0 table 20. input data format idf[3:0] description 0000 16-bit non-multiplexed rgb (16-bit color, 565) input 0001 16-bit non-multiplexed ycrcb (24-bit color) input (y non-multiplexed, crcb multiplexed 0010 16-bit multiplexed rgb (24-bit color) input 0011 15-bit non-multiplexed rgb (15-bit color, 555) input 0100 12-bit multiplexed rgb (24-bit color) input ( ?c ? multiplex scheme) 0101 12-bit multiplexed rgb2 (24-bit color) input ( ?i ? multiplex scheme) 0110 8-bit multiplexed rgb (24-bit color, 888) input 0111 8-bit multiplexed rgb (16-bit color, 565) input 1000 8-bit multiplexed rgb (15-bit color, 555) input 1001-1111 8-bit multiplexed ycrcb (24-bit color) input (y, cr and cb are multiplexed) bit: 7 6 5 4 3 2 1 0 symbol: cfrb m/s* reserved mcp xcm1 xcm0 pcm1 pcm0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 0 0 1 0 0 0 0
201-0000-026 rev 2.1, 8/2/99 35 chrontel ch7006c register descriptions (continued) note: for what was formerly defined as the master mode, the user must now externally connect the p-out clock to the xclk input pin. although it is possible to set the xcm [1:0] and pcm[1:0] values independent of the input data format, there are only certain combinations of input data format, xcm and pcm, that will result in valid data being demultiplexed at the input of the device. refer to the ?input data format register ? for these combinations. note: display modes 25 and 26 must use a 2x multiplexed input data format and a 2x xclk. display modes 27 and 28 must use a 1x xclk input data format. the clock mode register also contains the following bits: ? mcp (bit 4) determines which edge of the pixel clock output will be used to latch input data. zero selects the negative edge, one selects the positive edge. ? m/s* (bit 6) determines whether the device operates in master or slave clock mode. in master mode (1), the 14.31818mhz clock is used as a frequency reference to the pll. in slave mode (0) the xclk input is used as a reference to the pll, and is divided by the value specified by xcm[1:0]. the divide by n and m are forced to one. ? cfrb (bit 7) sets whether the chroma subcarrier free-runs, or is locked to the video signal. one causes the subcarrier to lock to the tv vertical rate, and should be used when the aciv bit is set to zero. zero causes the subcarrier to free-run, and should be used when the aciv bit is set to one. start active video register symbol: sav address: 07h bits: 8 this register sets the delay in pixel increments from leading edge of horizontal sync, or the rising edge of data start, to the start of active video. the entire bit field sav[8:0] is comprised of this register sav[7:0], plus the msb value contained in the position overflow register, bit sav8. this is decoded as a whole number of pixels, which can be set anywhere between 0 and 511 pixels. therefore, in any 2x clock mode, the number of 2x clocks from the leading edge of sync to the first active data must be a multiple of two clocks. in any 3x clock mode, the number of 3x clocks from the leading edge of sync to the first active data must be a multiple of three clocks. when using the ds/bco pin as a data start input, this register should be set to decimal value 11. table 21. input data format register xcm[1:0] pcm[1:0] xclk p-out input data modes supported 00 00 1x 1x 0, 1, 2, 3, 4, 5, 7, 8, 9 00 01 1x 2x 0, 1, 2, 3, 4, 5, 7, 8, 9 00 1x 1x 3x 0, 1, 2, 3, 4, 5, 7, 8, 9 01 00 2x 1x 2, 4, 5, 7, 8, 9 01 01 2x 2x 2, 4, 5, 7, 8, 9 01 1x 2x 3x 2, 4, 5, 7, 8, 9 1x 00 3x 1x 6 1x 01 3x 2x 6 1x 1x 3x 3x 6 bit: 7 6 5 4 3 2 1 0 symbol: sav7 sav6 sav5 sav4 sav3 sav2 sav1 sav0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0 0 0
chrontel ch7006c 36 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) position overflow register symbol: po address: 08h bits: 3 this position overflow register contains the msb values for the sav, hp, and vp values, as follows: ? vp8 (bit 0) is the msb of the vertical position value (see explanation under ?vertical position register ?). ? hp8 (bit 1) is the msb of the horizontal position value (see explanation under ?horizontal position register ?). ? sav8 (bit 2) is the msb of the start of active video value (see explanation under ?start active video register ?). black level register symbol: blr address: 09h bits: 8 this register sets the black level. the luminance data is added to this black level, which must be set between 90 and 208, with the default value being 127. recommended values for ntsc and pal-m are 127, 105 for pal and 100 for ntsc-j. horizontal position register symbol: h p r address: 0ah bits: 8 the horizontal position register is used to shift the displayed tv image in a horizontal direction (left or right) to achieve a horizontally centered image on screen. the entire bit field, hp[8:0] is comprised of this register hp[7:0] plus the msb value contained in the position overflow register, bit hp8. increasing this value moves the displayed image position right; decreasing this value moves the displayed image position left. each increment moves the image position by 4 input pixels. bit: 7 6 5 4 3 2 1 0 symbol: sav8 hp8 vp8 type: r/w r/w r/w default: 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: bl7 bl6 bl5 bl4 bl3 bl2 bl1 bl0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 1 1 1 1 1 1 1 bit: 7 6 5 4 3 2 1 0 symbol: hp7 hp6 hp5 hp4 hp3 hp2 hp1 hp0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0 0 0
201-0000-026 rev 2.1, 8/2/99 37 chrontel ch7006c register descriptions (continued) vertical position register symbol: vpr address: 0bh bits: 8 this register is used to shift the displayed tv image in a vertical direction (up or down) to achieve a vertically cen- tered image on screen. this bit field, vp[8:0] represents the tv line number (relative to the vga vertical sync) used to initiate the generation and insertion of the tv vertical interval (i.e., the first sequence of equalizing pulses). increasing values delay the output of the tv vertical sync, causing the image position to move up on the tv screen. decreasing values, therefore, move the image position down. each increment moves the image position by one tv lines (approximately 4 input lines). the maximum value that should be programmed into the vp[8:0] value is the number of tv lines minus one, divided by two (262, 312 or 313). when panning the image up, the number should be increased until (tvlpf-1) /2 is reached; the next step should be to reset the register to zero. when pan- ning the image down the screen, the vp[8:0] value should be decremented until the value zero is reached. the next step should set the register to (tvlpf-1) /2, and then decrementing can continue. if this value is programmed to a number greater than (tv lines per frame-1) /2, a tv vertical sync will not be generated. sync polarity register symbol: spr address: 0dh bits: 4 this register provides selection of the synchronization signal input to, or output from, the ch7006. ? hsp (bit 0) is horizontal sync polarity - an hsp value of zero means the horizontal sync is active low and a value of one means the horizontal sync is active high. ? vsp (bit 1) is vertical sync polarity - a vsp value of zero means the vertical sync is active low and a value of one means the vertical sync is active high. ? syo (bit 2) is sync direction - a syo value of zero means that h and v sync are input to the ch7006. a value of one means that h and v sync are output from the ch7006. ? des (bit 3) is detect embedded sync - a des value of zero means that h and v sync will be obtained from the direct pin inputs. a des value of one means that h and v sync will be detected from the embedded codes on the pixel input stream. note that this will only be valid for the ycrcb input modes. note: when sync direction is set to be an output, horizontal sync will use a fixed pulse width of 64 pixels and vertical sync will use a fixed pulse width of 1 line. bit: 7 6 5 4 3 2 1 0 symbol: vp7 vp6 vp5 vp4 vp3 vp2 vp1 vp0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: des syo vsp hsp type: r/w r/w r/w r/w default: 0 0 0 0
chrontel ch7006c 38 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) power management register symbol: pmr address: 0eh bits: 5 this register provides control of the power management functions, a software reset (resetb), and the scart output enable. the ch7006 provides programmable control of its operating states, as described in the table below. reset* (bit 3) is soft reset. setting this bit will reset all circuitry requiring a power on reset, except for this bit itself and the i 2 c state machines. scart (bit 4) is the scart enable. setting scart = 0 means the ch7006 will operate normally, outputting y/c and cvbs from the three dacs. scart=1 enables scart output, which will cause r, g and b to be output from the dacs and composite sync from the csync pin. note: for complete details regarding the operation of these modes, see the power management in functional description sections. connection detect register symbol: cdr address: 10h bits: 4 the connection detect register provides a means to sense the connection of a tv to either s-video or composite video outputs. the status bits, yt, ct, and cvbst correspond to the dac outputs for s-video (y and c outputs) and composite video (cvbs), respectively. however, the values contained in these status bits are not valid until a sensing procedure is performed. use of this register requires a sequence of events to enable the sensing of outputs, then reading out the applicable status bits. the detection sequence works as follows: 1. ensure the power management register bits 2-0 is set to 011 (normal mode). bit: 7 6 5 4 3 2 1 0 symbol: scart reset* pd2 pd1 pd0 type: r/w r/w r/w r/w r/w default: 0 1 0 1 1 table 22. power management pd[2:0] operating state functional description 000 composite off cvbs dac is powered down. 001 power down most pins and circuitry are disabled (except for the buffered clock outputs which are limited to the 14mhz output and vco divided output when the ds/bco pin is selected to be an output). 010 s-video off s-video dacs are powered down. 011 normal (on) all circuits and pins are active. 1xx full power down all circuitry is powered down except i 2 c circuit. bit: 7 6 5 4 3 2 1 0 symbol: yt ct cvbst sense type: r r r w default: 0 0 0 0
201-0000-026 rev 2.1, 8/2/99 39 chrontel ch7006c register descriptions (continued) 2. set the sense bit to a 1. this forces a constant current output onto the y, c, and cvbs outputs. note that during sense = 1, these 3 analog outputs are at steady state and no tv synchronization pulses are asserted. 3. reset the sense bit to 0. this triggers a comparison between the voltage sensed on these analog outputs and the reference value expected (v threshold = 1.235v). if the measured voltage is below this threshold value, it is considered connected, if it is above this voltage it is considered unconnected. during this step, each of the three status bits corresponding to individual analog outputs will be set if they are not connected. 4. read the status bits. the status bits, y, c, and cvbst (corresponding to s-video y and c outputs and composite video) now contain valid information which can be read to determine which outputs are connected to a tv. again, a ?0 ? indicates a valid connection, a ?1 ? indicates an unconnected output. contrast enhancement register symbol: ce address: 11h bits: 3 this register provides control of the contrast enhancement feature of the ch7006, according to the table below. at a setting of 000, the video signal will be pulled towards the maximum black level. as the value of ce[2:0] is increased, the amount that the signal is pulled towards black is decreased until unity gain is reached at a setting of 011. from this point on, the video signal is pulled towards the white direction, with the effect increasing with increasing settings of ce[2:0]. bit: 7 6 5 4 3 2 1 0 symbol: ce2 ce1 ce0 type: r/w r/w r/w default: 0 1 1 table 23. contrast enhancement function ce[2:0] description (all gains limited to 0-255) 000 contrast enhancement gain 3 y out = (5/4)*(y in -102) = enhances black 001 contrast enhancement gain 2 y out = (9/8)*(y in -57) 010 contrast enhancement gain 1 y out = (17/16)*(y in -30) 011 normal mode y out = (1/1)*(yin-0) = normal contrast 100 contrast enhancement gain 1 y out = (17/16)*(y in -0) 101 contrast enhancement gain 2 y out = (9/8)*(y in -0) 110 contrast enhancement gain 3 y out = (5/4)*(y in -0) 111 contrast enhancement gain 4 y out = (3/2)*(y in -0) = enhances white
chrontel ch7006c 40 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) figure 26: luma transfer function at different contrast enhancement settings. pll overflow register symbol: mne address: 13h bits: 5 the pll overflow register contains the msb bits for the ?m ? and ?n ? values, which will be described in the pll- m and pll-n registers, respectively. the reserved bits should not be written to. pll m value register symbol: pllm address: 14h bits: 8 the pll m value register determines the division factor applied to the frequency reference clock before it is input to the pll phase detector when the ch7006 is operating in master or pseudo-master clock mode. in slave mode, an external pixel clock is used instead of the frequency reference, and the division factor is determined by the xcm[3:0] value. this register contains the lower 8 bits of the complete 9-bit m value. bit: 7 6 5 4 3 2 1 0 symbol: reserved reserved n9 n8 m8 type: r/w r/w r/w r/w r/w default: 0 0 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: m7 m6 m5 m4 m3 m2 m1 m0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 0 1 0 0 0 0 0 1 0 32 64 96 128 160 192 224 256 0 32 64 96 128 160 192 224 256
201-0000-026 rev 2.1, 8/2/99 41 chrontel ch7006c register descriptions (continued) pll n value register symbol: plln address: 15h bits: 8 the pll n value register determines the division factor applied to the vco output before being applied to the pll phase detector, when the ch7006 is operating in master or pseudo-master mode. in slave mode, the value of ?n ? is always 1. this register contains the lower 8 bits of the complete 10-bit n value. the pixel clock generated in a master and pseudo-master modes is calculated according to the equation below: fpixel = fref* [(n+2) / (m+2)] when using a 14.318 mhz frequency reference, the required m and n values for each mode are shown in the table below buffered clock output register symbol: bco address: 17h bits: 6 bit: 7 6 5 4 3 2 1 0 symbol: n7 n6 n5 n4 n3 n2 n1 n0 type: r/w r/w r/w r/w r/w r/w r/w r/w default: 1 0 0 0 0 0 0 0 table 24. m and n values for each mode mode vga resolution, tv standard, scaling ratio n 10- bits m 9- bits mode vga resolution, tv standard, scaling ratio n 10- bits m 9- bits 0 512x384, pal, 5:4 20 13 15 640x480, pal, 5:6 9 3 1 512x384, pal, 1:1 9 4 16 640x480, ntsc, 1:1 110 63 2 512x384, ntsc, 5:4 126 89 17 640x480, ntsc, 7:8 126 63 3 512x384, ntsc, 1:1 110 63 18 640x480, ntsc, 5:6 190 89 4 720x400, pal, 5:4 53 26 19 800x600, pal, 1:1 647 313 5 720x400, pal, 1:1 339 138 20 800x600, pal, 5:6 86 33 6 720x400, ntsc, 5:4 106 63 21 800x600, pal, 3:4 284 103 7 720x400, ntsc, 1:1 70 33 22 800x600, ntsc, 5:6 94 33 8 640x400, pal, 5:4 108 61 23 800x600, ntsc, 3:4 62 19 9 640x400, pal, 1:1 9 3 24 800x600, ntsc, 7:10 302 89 10 640x400, ntsc, 5:4 94 63 25 720x576, pal, 1:1 31 33 11 640x400, ntsc, 1:1 22 11 26 720x480, ntsc, 1:1 31 33 12 640x400, ntsc, 7:8 190 89 27 800x500, pal, 1:1 242 197 13 640x480, pal, 5:4 20 13 28 640x400, ntsc, 1:1 2 2 14 640x480, pal, 1:1 9 4 bit: 7 6 5 4 3 2 1 0 symbol: shf2 shf1 shf0 sco2 sco1 sco0 type: r/w r/w r/w r/w r/w r/w default: 0 0 0 0 0 0
chrontel ch7006c 42 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) when this pin is selected to be an output, the buffered clock output register determines which clock is selected to be output at the ds/bco clock output pin and what frequency value is output when a vco derived signal is output.the tables below show the possible outputs. subcarrier value registers symbol: fsci address: 18h - 1fh bits: 4 or 8 each the lower four bits of registers 18h through 1fh contain a 32-bit value which is used as an increment value for the rom address generation circuitry. the bit locations are specified as the following: register contents 18h fsci[31:28] 19h fsci[27:24] 1ah fsci[23:20] 1bh fsci[19:16] 1ch fsci[15:12] 1dh fsci[11:8] 1eh fsci[7:4] 1fh fsci[3:0] table 25. clock output selection sco[2:0] buffered clock output 000 14mhz crystal 001 (for test use only) 010 vco divided by k3 (see table26 ) 011 field id signal 100 sine rom msb (for test use only) 101 (for test use only) 110 (for test use only) 111 tv vertical sync (for test use only) table 26. k3 selection shf[2:0] k3 000 2.5 010 3.5 011 4 100 4.5 101 5 110 6 111 7 bit: 7 6 5 4 3 2 1 0 symbol: fsci# fsci# fsci# fsci# type: r/w r/w r/w r/w default:
201-0000-026 rev 2.1, 8/2/99 43 chrontel ch7006c register descriptions (continued) when the ch7006 is operating in the master clock mode, the tables below should be used to set the fsci registers. when using these values, the aciv bit in register 21h should be set to ?0 ? and the cfrb bit in register 06h should be set to ?1 ? . when the ch7006 is operating in the slave clock mode, the aciv bit in register 21h should be set to ?1 ?and the cfrb bit in register 06h should be set to ?0 ?. *note: for reduced cross-color and cross-luminance artifacts, a value of 488,265,597 can be used with cfrb = "0" & aciv = "0". table 27. fsci values (ntsc modes) mode ntsc ?normal dot crawl ? ntsc ?no dot crawl ? pal-m ?normal dot crawl 2 763,363,328 763,366,524 762,524,467 3 623,153,737 623,156,346 622,468,953 6 574,429,782 574,432,187 573,798,541 7 463,962,517 463,964,459 463,452,668 10 646,233,505 646,236,211 645,523,358 11 516,986,804 5165,988,968 516,418,687 12 452,363,454 452,365,347 451,866,351 16 623,153,737 623,156,346 622,468,953 17 545,259,520 545,261,803 544,660,334 18 508,908,885 508,911,016 508,349,645 22 521,957,831 521,960,016 521,384,251 23 469,762,048 469,764,015 469,245,826 24 428,554,851 438,556,645 428,083,911 26 569,408,543 569,410,927 568,782,819 28 1,073,741,824 1,073,746,319 1,072,561,888 table 28. fsci values (pal modes) mode pal ?normal dot crawl ? pal-n ?normal dot crawl ? 0 806,021,060 651,209,077 1 644,816,848 520,967,262 4 601,829,058 486,236,111 5 485,346,014 392,125,896 8 677,057,690 547,015,625 9 537,347,373 434,139,385 13 806,021,060 651,209,077 14 644,816,848 520,967,262 15 537,347,373 434,139,385 19 645,499,916 521,519,134 20 528,951,320 427,355,957 21 488,262,757* 394,482,422 25 705,268,427 569,807,942 27 1,073,747,879 867,513,766
chrontel ch7006c 44 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) symbol: address: 1bh bits: 8 note: p out p (bit 4) is used to invert the p-out signal. symbol: address: 1ch bits: 6 note: dsen (bit4) controls the bco / data start i/o pin. when this bit is low, the pin continues to operate as the bco pin described in the bco register description. when this bit is high, the pin becomes an input for the data start signal. pll control register symbol: pllc address: 20h bits: 6 the following pll and memory controls are available through the pll control register: mem5v mem5v is set to 1 when the memory supply is 5 volts. the default value of 0 is used when the memory supply is 3.3 volts. pll5va pll5va is set to 1 when the phase-locked loop analog supply is 5 volts (default). a value of 0 is used when the phase-locked loop analog supply is 3.3 volts. pll5vd pll5vd is set to 1 when the phase-locked loop digital supply is 5 volts. a value of 0 is used when the phase-locked loop digital supply is 3.3 volts (default). plls plls controls the number of stages used in the pll. when the pll5va is 1 (5v analog pll supply) plls should be 1, and seven stages are used. when pll5va is 0 (3.3v analog pll supply) plls should be 0, and five stages are used. pllcap pllcap controls the loop filter capacitor of the pll. a recommended listing of pllcap vs. mode is shown below pllcpi pllchi controls the charge pump current of the pll. the default value should be used. bit: 7 6 5 4 3 2 1 0 symbol: p-outp fsci19 fsci18 fsci17 fsci16 type: r/w r/w r/w r/w r/w default: 0 0 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: dsen fsci15 fsci14 fsci13 fsci12 type: r/w r/w r/w r/w r/w default: 1 0 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: pllcpi pllcap plls pll5vd pll5va mem5v type: r/w r/w r/w r/w r/w r/w default: 0 0 1 0 1 0
201-0000-026 rev 2.1, 8/2/99 45 chrontel ch7006c register descriptions (continued) table 29. pll capacitor setting mode pllcap value 0 1 1 1 2 1 3 0 4 1 5 0 6 1 7 1 8 0 9 1 10 1 11 1 12 0 13 1 14 1 15 1 16 0 17 0 18 0 19 0 20 1 21 0 22 1 23 1 24 0 25 1 26 1 27 0 28 1
chrontel ch7006c 46 201-0000-026 rev 2.1, 8/2/99 register descriptions (continued) civ control register symbol: civc address: 21h bits: 5 the following controls are available through the civ control register: aciv when the automatic calculated increment value is 1, the number calculated and present at the civ registers will automatically be used as the increment value for subcarrier generation, removing the need for the user to read the civ value and write in a new fsci value. whenever this bit is set to 1, the subcarrier generation must be forced to free-run mode. civh[1:0] these bits control the hysteresis circuit which is used to calculate the civ value. civ[25:24] see descriptions in the next section. calculated increment value register symbol: civ address: 22h - 24h bits: 8 the civ registers 22h through 24h contain a 26-bit value, which is the calculated increment value that should be used as the upper 26 bits of fsci. this value is determined by a comparison of the pixel clock and the 14mhz clock. the bit locations and calculation of civ are specified as the following: register contents 21h civ[25:24] 22h civ[23:16] 23h civ[15:8] 24h civ[7:0] version id register symbol: vid address: 25h bits: 8 this read-only register contains a 8-bit value indicating the identification number assigned to this version of the ch7006. the default value shown is pre-programmed into this chip and is useful for checking for the correct version of this chip, before proceeding with its programming. bit: 7 6 5 4 3 2 1 0 symbol: civ25 civ24 civh1 civh0 aciv type: r r r/w r/w r/w default: 0 0 0 0 1 bit: 7 6 5 4 3 2 1 0 symbol: civ# civ# civ# civ# civ# civ# civ# civ# type: r r r r r r r r default: 0 0 0 0 0 0 0 0 bit: 7 6 5 4 3 2 1 0 symbol: vid7 vid6 vid5 vid4 vid3 vid2 vid1 vid0 type: r r r r r r r r default: 0 0 1 0 1 0 1 0
201-0000-026 rev 2.1, 8/2/99 47 chrontel ch7006c register descriptions (continued) address register symbol: ar address: 3fh bits: 6 the address register points to the register currently being accessed. electrical specifications notes: 1. stresses greater than those listed under absolute maximum ratings may cause permanent damage to the device. these are stress ratings only. functional operation of the device at these or any other conditions above those indicated under the normal operating condition of this specification is not recommended. exposure to absolute maximum rating conditions for extended periods my affect reliability. 2. the device is fabricated using high-performance cmos technology. it should be handled as an esdsensitive device. voltage on any signal pin that exceeds the power supply voltage by more than +0.5v can induce destructive latch. bit: 7 6 5 4 3 2 1 0 symbol: ar5 ar4 ar3 ar2 ar1 ar0 type: r/w r/w r/w r/w r/w r/w default: x x x x x x table 30. absolute maximum ratings symbol description min typ max units v dd relative to gnd - 0.5 7.0 v input voltage of all digital pins 1 gnd - 0.5 vdd + 0.5 v t sc analog output short circuit duration indefinite sec t amb ambient operating temperature - 55 125 c t s t o r storage temperature - 65 150 c t j junction temperature 150 c t v p s vapor phase soldering (one minute) 220 c table 31. recommended operating conditions symbol description min typ max units vdd dac power supply voltage 4.75 5.00 5.25 v avdd analog supply voltage 5.00 5.25 dvdd digital supply voltage 3.3 3.6 r l output load to dac outputs 37.5 w
chrontel ch7006c 48 201-0000-026 rev 2.1, 8/2/99 note: rset = 360 w w , vref = 1.235v, and ntsc ccir601 operation. note: 1. v iic -refers to i2c pins sd and sc. 2. v data - refers to all digital pixel and clock inputs. 3. v sd - refers to i2c pin sd as an output. 4. v p-out - refers to pixel data output time - graphics. table 32. electrical characteristics (operating conditions: t a = 0 o c - 70 o c, v dd = 5v 5% ) symbol description min typ max unit video d/a resolution 9 9 9 bits full scale output current 33.89 ma video level error 10 % vdd & avdd (5.v) current (simultaneous s-video & composite outputs) 105 ma dvdd current (3.3v) 40 ma table 33. timing - tv encoder symbol description min typ max unit t p1 pixel clock period 20 50 ns t ph1 pixel clock high time 8 25 ns tdc1 pixel clock duty cycle (t ph1 /t p1 ) 40 50 60 % t p2 pixel clock period 10 25 ns t ph2 pixel clock high time ns tdc2 pixel clock duty cycle (t ph2 /t p2 ) 40 50 60 % t p3 pixel clock period 10 17 ns t ph3 pixel clock high time ns tdc3 pixel clock duty cycle (t ph3 /t p3 ) 40 50 60 % table 34. digital inputs / outputs symbol description test condition min typ max unit v sdol sd output low voltage iol = 2.0 ma 0.4 v v iicih sd input high voltage 2.7 vdd + 0.5 v v iicil sd input low voltage gnd-0.5 1.4 v v dataih d[0-15] input high voltage 2.5 dvdd+0.5 v v datail d[0-15] input low voltage gnd-0.5 0.8 v v p-outoh p-out output high voltage iol = - 400 m a 2.8 v v p-outol p-out output low voltage iol = 3.2 ma 0.2 v
chrontel 2210 o ?toole avenue san jose, ca 95131-1326 tel: (408) 383-9328 fax: (408) 383-9338 www.chrontel.com e-mail: sales@chrontel.com 1998 chrontel, inc. all rights reserved. chrontel products are not authorized for and should not be used within life support systems or nuclear facility applications wit hout the specific written consent of chrontel. life support systems are those intended to support or sustain life and whose failure to p erform when used as directed can reasonably expect to result in personal injury or death. chrontel reserves the right to make changes at any time without notice to improve and supply the best possible product and is not responsible and does not assume any liability for misapplication or use outside the limits specified in this document. we provi de no warranty for the use of our products and assume no liability for errors contained in this document. printed in the u.s.a. ordering information part number package type number of pins voltage supply CH7006C-V plcc 44 3v/5v ch7006c-t tqfp 44 3v/5v 201-0000-026 rev 2.1, 8/2/99 49 chrontel ch7006c table 35. timing graphics symbol description min typ max unit t hsw horizontal sync pulse width 1 t p t hd pixel clock to horizontal leading edge delay 2 17 ns t sp1, t sp2, t sp3 setup time from pixel data to pixel clock 2 ns t ph1, t hp2, t ph3 hold time from pixel clock to pixel data 2 ns

▲Up To Search▲

Price & Availability of CH7006C-V

	To Download CH7006C-V Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .