Mar 9 2006, 05:52 AM
Post
#1
|
|
![]() Group: Members Posts: 1,565 Joined: 7-April 05 From: Sydney, Australia Member No.: 6,806 |
Ok here are the details of my research to date, at the mement i am 1 pin away from being technically able to pull this hack off (GPIO18 for those who are intrested aka RDY, only used for VLIO)
hope this comes in handy, after this the next problem is sourcing chips CODE Note this is all 2700G7 specific, i plan to use the 2700G5 as it has suport for external Mem (up to 32MB) and 704KB of SRAM for frame buffers, the 2700G7 has only 16MB of built in MEM with no support for more as well as the FRAME Buffer, depending on which is cheaper + size constraints + avalibility + mem requirements i may change.
2700G PXA270 Description Avalible SYS_nCAS nSDCAS System Bus Address Strobe Yes (off SDRAM) SYS_MA[25:2] MA[25:2] System Bus Address Yes (off SDRAM and flash) SYS_MD[31:0] MD[31:0] System Bus Data Yes (off SDRAM and flash) SYS_nOE nOE System Bus Output Enable Yes (off flash) SYS_RDnWR RDnWR System Bus Read/Write Yes (off CF card) SYS_RDY RDY System Bus Ready SYS_nCS[1:0] nCS[1:0] System Bus Chip Select Yes (off flash, mabey, it looks like we could steal it from that second flash slot, there are 6, we need 2 and they cant be used for anything else) SYS_nWE nWE System Bus SRAM Write Enable Yes (off SDRAM and flash) SYS_nPWE nPWE System Bus VLIO Write Enable Yes (off CF card) SYS_DQM[3:0] DQM[3:0] System Bus Data Masks Yes (off SDRAM) System Bus Signals = 68 pins For LCD's you get a pixel clock that incicates pixels, an Line Clock which indicates End of line (or HSYNC) and a Frame Clock (VSYNC) so you can see if you get rid of the pixel clock and add timing before clocking and after clocking the pixels out before/after the line/frame clock lines are triggered it would be compatible with a CRT, not sure but it might work with the new DVI connector with out insertint the "blanking", need to do some more research if this (the 2700G) dosen't pay off. Local Mem Signals = 57 pins (all on PCB) LCD_IN_DD[17:0] LCD Data Inputs LCD_IN_PCLK Pixel Clock Input LCD_IN_LCLK Line Clock Input LCD_IN_FCLK Frame Clock Input LCD_IN_DEN LCD Data Enable Input LCD Input Pin Signals = 22 pins (header going to LCD screen on zaurus) LCD1_DD[23:0] LCD Data LCD1_PCLK LCD Pixel Clock LCD1_LCLK LCD Line Clock LCD1_FCLK LCD Frame Clock LCD1_DEN LCD Data Enable LCD1_PWM Allows changing of backlight brightness LCD2_DD[23:0] LCD Data LCD2_PCLK LCD Pixel Clock LCD2_LCLK LCD Line Clock LCD2_FCLK LCD Frame Clock LCD2_DEN LCD Data Enable LCD2_PWM Allows changing of backlight brightness LCD Output Pin Signals = 29 pins * 2 (primary and secondary) (One semi on board: Internal LCD connector is flat cable) XTAL_IN CLOCK Input XTAL_OUT CLOCK Input (or outptu depending on how you like it) CLKIN Clock input if XTAL_* not used GPIO General pourpos IO (Yeah!!) POLL_FLAG Connect to GPIO (PXA) for flow controll nRESET_IN Reset on LOW nINT Interupt to main processor (can i even acsess this line :( ) RSVD 100K pull down to ground resistor on these lines MISC Pins = 13 Pins (not all connected) VCC_CORE Core Power Supply (1.2V 28 pins) VCC_SYS System Bus Supply (1.8V, 2.5V 9 pins) VCC_LM Local Mem Bus Supply (1.8V 10 pins) VCC_SDRAM SDRAM Device Supply (1.8V 13 pins) VCC_LCD_IN LCD Input Power (1.8V, 2.5V 2 pins) VCC_LCD1 Primary LCD interface Power (1.8V, 2.5V, 3.3V 4 pins) VCC_LCD2 Secondary LCD interface Power (1.8V, 2.5V, 3.3V 4 pins) VCC_IO Misc IO Power (3.3V 15 pins) GND GND (Come on you should know this 50 pins) VCCA_CORE_PLL Core PLL Analog Power (2.5V 1 pin) VSSA_CORE_PLL Core PLL Analog Gnd (Gnd 1 pin) VCCA_DISP_PLL Display PLL Analog Power (2.5V 1 pin) VSSA_DISP_PLL Display PLL Analog Gnd (Gnd 1 pin) VAA_XTAL Crystal Osscilator Analog Power (2.5V 1 pin) VSSA_XTAL Crystal Osscilator Analog Power (Gnd 1 pin) Power Pins = 141 (all on PCB) |
|
|
|
![]() |
Jun 9 2006, 07:36 PM
Post
#2
|
|
![]() Group: Members Posts: 1,565 Joined: 7-April 05 From: Sydney, Australia Member No.: 6,806 |
Pitiful is 16MBs and cuts into avalible sdram bandwidth and microdrive acsess so any of those will affect each other
if you do the 16M / yres by xres * (colurdepth/4) you get the refresh rate for any res so for moive playback you need to drop the res a bit, however i can do a bit of 2d accelaration and moive playback stuff to help bump up the speed a bit as well as some register hacks i have been looking into battery wise im not sure, it definattly wont chew as much as a wifi card |
|
|
|
Da_Blitz 2700g Video Card Mar 9 2006, 05:52 AM
speculatrix I really hope you can pull this off. Great work... Mar 12 2006, 01:24 PM
Da_Blitz Just need to find that last IO and it should be ea... Mar 12 2006, 04:07 PM
Da_Blitz Ok so gpio18 is avalible thanks to albetr os now i... Jun 3 2006, 10:55 PM
speculatrix http://www.reghardware.co.uk/2006/06/07/nv...bilem... Jun 7 2006, 03:59 AM
speculatrix er, I think I meant this page, sorry:
http://www.d... Jun 7 2006, 03:59 AM
Da_Blitz the Three top players that i can see are also the ... Jun 9 2006, 12:26 AM
Da_Blitz WOW, long post. Sorry bout that Jun 9 2006, 12:27 AM
Reaper Video output by itself is very nice (especially if... Jun 9 2006, 12:40 AM
adf QUOTE(Reaper @ Jun 9 2006, 08:40 AM)Video out... Jun 9 2006, 01:57 AM
the_oak QUOTE*CF graphics card with pitiful transfer rate ... Jun 9 2006, 04:48 AM![]() ![]() |
|
Lo-Fi Version | Time is now: 19th June 2013 - 06:32 AM |