From: Holly G. <hgates@EINK.com> - 2005-11-10 14:56:30
|
I'm sure I will be less useful to you since I have not worked with that particular display. But in general what I do for an active display is: - get the fb set up to the right size, as described on the wiki page. Probably this will be just the resolution of your screen. If you don't know it, start out with something bigger than you think the screen is. - install new kernel, boot up - make sure pixel clock is running using scope - set LCCR values to make timing like datasheet values, use scope to check - if no data sheet is available (I have not done this; I would buy an LCD only if a data sheet was available from somewhere), start out by setting PCD so that FCLK is about 60 Hz, as seen on a scope. Next, set HSW, BLW, and ELW to about 1uS. Set VSW, EFW, an BFW to 0. Now if nothing is working, try changing PCP, VSP, HSP, and OEP one at a time, then in various combinations. If you know the resolution of the panel and have a connector pinout for it, you have a non-trivial chance of getting it working. If you don't even know the panel pinout, I think it would take a huge amount of time and probably some burned out panels before you got anywhere. Hope that helps, -Holly Mark Wiesbock wrote: >Hello, > >We have order and now have, the ALPS LCD display listed on the Display >page on gumstix.org. It looked like a good LCD for the price, and the >pictures from http://hubbard.engr.scu.edu/embedded/lcd/lfh8p4032b/lfh8p4032b.html >looked like a pretty good sign as well, but so far we haven't had much >luck getting the screen working.. or finding the proper timing >parameters in general. > >Does anyone have any methods or ideas on how to go about getting this >screen up and running? Is manually just changing the timing values the >only option, or is there some way we could use an oscilloscope or >other equipment in order to get some results? > >As well, does anyone have any file-system images we would be able to >throw on our gumstix that would be able to get a picture loaded into >the frame buffer easily, so just in case we do have to go about the >manual procedure with the timing, at least we would know there is >something on the display to see? > >Thanks for all your help, >-Mark > > >------------------------------------------------------- >SF.Net email is sponsored by: >Tame your development challenges with Apache's Geronimo App Server. Download >it for free - -and be entered to win a 42" plasma tv or your very own >Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php >_______________________________________________ >gumstix-users mailing list >gum...@li... >https://lists.sourceforge.net/lists/listinfo/gumstix-users > > |
From: ohm u. <oun...@ya...> - 2005-11-14 17:30:40
|
Hello all, I am working alongside Mark on this ALPS project. I just needed to ask a few newb questions. 1) How do we know the size of fb0 and if it has been set up correctly? I get this with ls -l in /dev/: crw-r----- 1 root tty 29, 0 Nov 7 2005 fb0 It gives me the major and minor numbers, which is correct, i think, since it is a special file. How can I find how big the device file is? Further, Mr. Jim Walters suggested using dd if=/dev/zero of=/dev/fb0 to get a read on how much memory is allocated to fb0: my output: # dd if=/dev/zero of=/dev/fb0 dd: /dev/fb0: No space left on device 2) I used an oscilloscope to find initial frequencies. my output: L_PCLK ~4Mhz L_FCLK nothing L_LCLKC nothing can this be connected to this dmesg output? -- pxa2xx-fb pxa2xx-fb: overriding resolution: 240x320 -- pxa2xx-fb pxa2xx-fb: overriding bit depth: 8 -- PXA: invalid hsync_len 0 -- PXA: invalid left_margin 0 -- PXA: invalid right_margin 0 -- PXA: invalid vsync_len 0 (edit before sending: i set the h and vsync_len and margin values to values other than zero, and still no output) What would cause L_FCLK to not be outputting a waveform? What can I do to put out a frequency? 3) Holly Gates mentioned this in a response to Mark: ... Next, set HSW, BLW, and ELW to about 1uS. ... if i were to do this using pxaregs, would the usage be: root/pxaregs LCCR1_HSW 1 ? if i were to do it in the bootargs, would the usage be: ... video=pxafb:mode:240x160-8,passive,hsynclen:1 ... ? Specifically, what are the units when using pxaregs and when using the command line? Thank you very much, we'll all get this ALPS display working, yet! Ohm __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com |
From: <hg...@ei...> - 2005-11-15 05:02:03
|
The units for HSW, ELW, and BLW are in pixel clocks, so the actual length in time depends on your pixel clock speed; e.g. if you had a pixel clock speed of 1MHz (I think this would be a LCCR3_PCD of 100?), you would set them to be 1 for a 1uS length because your pixel clock period is 1uS. -Holly > Hello all, > I am working alongside Mark on this ALPS project. I > just needed to ask a few newb questions. > > 1) How do we know the size of fb0 and if it has been > set up correctly? I get this with ls -l in /dev/: > crw-r----- 1 root tty 29, 0 Nov 7 2005 fb0 > > It gives me the major and minor numbers, which is > correct, i think, since it is a special file. How can > I find how big the device file is? > > Further, Mr. Jim Walters suggested using dd > if=3D/dev/zero of=3D/dev/fb0 to get a read on how much > memory is allocated to fb0: > > my output: > # dd if=3D/dev/zero of=3D/dev/fb0 > dd: /dev/fb0: No space left on device > > > 2) I used an oscilloscope to find initial frequencies. > > > my output: > L_PCLK ~4Mhz > L_FCLK nothing > L_LCLKC nothing > can this be connected to this dmesg output? > > -- pxa2xx-fb pxa2xx-fb: overriding resolution: 240x320 > -- pxa2xx-fb pxa2xx-fb: overriding bit depth: 8 > -- PXA: invalid hsync_len 0 > -- PXA: invalid left_margin 0 > -- PXA: invalid right_margin 0 > -- PXA: invalid vsync_len 0 > > (edit before sending: i set the h and vsync_len and > margin values to values other than zero, and still no > output) > > What would cause L_FCLK to not be outputting a > waveform? > > What can I do to put out a frequency? > > 3) Holly Gates mentioned this in a response to Mark: > > ... > > Next, set > HSW, BLW, and ELW to about 1uS. > > ... > > > if i were to do this using pxaregs, would the usage > be: > root/pxaregs LCCR1_HSW 1 ? > > if i were to do it in the bootargs, would the usage > be: > ... video=3Dpxafb:mode:240x160-8,passive,hsynclen:1 ... > ? > > Specifically, what are the units when using pxaregs > and when using the command line? > > Thank you very much, we'll all get this ALPS display > working, yet! > > Ohm > > > > > __________________________________ > Yahoo! FareChase: Search multiple travel sites in one click. > http://farechase.yahoo.com > > > ------------------------------------------------------- > SF.Net email is sponsored by: > Tame your development challenges with Apache's Geronimo App Server. > Download > it for free - -and be entered to win a 42" plasma tv or your very own > Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > |
From: Michael P. <mk...@in...> - 2005-11-15 04:45:06
|
Hi Ohm - I'm a newb here too but been messing with gumstix and LCDs for a few weeks. I'm having success with the Sony LCD thanks to the wiki docs. A few points - try dd if=/dev/fb0 of=/dev/null :) if being the input file and of the output file. the command as you typed it will send zeros to the fb until its full which gave you the no space left error (and would blank the screen). At least the fb exists and accepts input. This way round it should dump all of /dev/fb0 to null without error and report a count of bytes (or kb or some indication of filesize). Nothing on your LCLK and FCLK and the dmesg output means the kernel isn't setting those up - you need to give some values in the bootargs (and I suggest using them rather than pxaregs - more human readable) Have a look at the file pxafb.txt in the kernel source Documentation/fb/ directory for some more info on using the bootargs. You set the PIXCLOCK in picoseconds then I think the other values are derived from that ie the HSYNCLEN will be x number of pixclock periods as will left and right. It's pretty tough figuring these values even with a good datasheet, pity the ALPS datasheet doesn't have timing data :( Also in an earlier post Mark was asking for easy ways to get a picture on the screen so you know there's something there. I suggest that anyone playing with LCD setups when compiling your kernel for framebuffer support make sure to also: enable Virtual terminal in Character Devices + support for console on virtual terminal then in Graphics Support --> Console display driver support enable Framebuffer Console support and Select compiled-in fonts eg VGA 8x16 font then add 'console=tty0' to your bootargs in front of 'console=ttyS0,115200n8' so for example my bootargs look like this for the Sony LCD: setenv bootargs console=tty0 console=ttyS0,115200n8 root=1f02 rootfstype=jffs2 reboot=cold,hard video=pxafb:mode:240x160-16,active,hsynclen:7,left:12,right:16, upper:19,lower:19,vsynclen:2,hsync:1,vsync:1,pixclockpol:0,pixclock:216363 this way you get the nice little penguin and kernel messages on screen as soon as the kernel initializes and even if things look wrong you know you're on track. So good luck with the ALPS screen I'm sure you'll get there, I hope this info helps a little - Michael >Hello all, >I am working alongside Mark on this ALPS project. I >just needed to ask a few newb questions. > >1) How do we know the size of fb0 and if it has been >set up correctly? I get this with ls -l in /dev/: >crw-r----- 1 root tty 29, 0 Nov 7 2005 fb0 > >It gives me the major and minor numbers, which is >correct, i think, since it is a special file. How can >I find how big the device file is? > >Further, Mr. Jim Walters suggested using dd >if=/dev/zero of=/dev/fb0 to get a read on how much >memory is allocated to fb0: > >my output: ># dd if=/dev/zero of=/dev/fb0 >dd: /dev/fb0: No space left on device > > >2) I used an oscilloscope to find initial frequencies. > > >my output: >L_PCLK ~4Mhz >L_FCLK nothing >L_LCLKC nothing >can this be connected to this dmesg output? > >-- pxa2xx-fb pxa2xx-fb: overriding resolution: 240x320 >-- pxa2xx-fb pxa2xx-fb: overriding bit depth: 8 >-- PXA: invalid hsync_len 0 >-- PXA: invalid left_margin 0 >-- PXA: invalid right_margin 0 >-- PXA: invalid vsync_len 0 > >(edit before sending: i set the h and vsync_len and >margin values to values other than zero, and still no >output) > >What would cause L_FCLK to not be outputting a >waveform? > >What can I do to put out a frequency? > >3) Holly Gates mentioned this in a response to Mark: > >... > >Next, set >HSW, BLW, and ELW to about 1uS. > >... > > >if i were to do this using pxaregs, would the usage >be: >root/pxaregs LCCR1_HSW 1 ? > >if i were to do it in the bootargs, would the usage >be: >... video=pxafb:mode:240x160-8,passive,hsynclen:1 ... >? > >Specifically, what are the units when using pxaregs >and when using the command line? > >Thank you very much, we'll all get this ALPS display >working, yet! > >Ohm |
From: Michael P. <mk...@in...> - 2005-11-15 07:01:28
|
Actually what I said about pixclock and hsync etc probably applies more to active screens than passive ones ? like I said noob here good luck > > > Have a look at the file pxafb.txt in the kernel source > Documentation/fb/ directory for some more info on using the bootargs. > You set the PIXCLOCK in picoseconds then I think the other values are > derived from that ie the HSYNCLEN will be x number of pixclock periods > as will left and right. |
From: ohm u. <oun...@ya...> - 2005-11-16 19:57:14
|
Wow, great information! Getting closer (flickering orange screen instead of nothing), but baby-step speed: I tried to enable the virtual console through the make menuconfig. Upon cd ../../ and make, I get these errors: /*snippet 1*/ make[1]: Leaving directory `/home/IQ/Desktop/gumstix/gumstix_builds/gumstix-buildroot/build_arm_nofpu/linux-2.6.11gum' make: *** [/home/IQ/Desktop/gumstix/gumstix_builds/gumstix-buildroot/build_arm_nofpu/linux-2.6.11gum/arch/arm/boot/compressed/vmlinux] Error 2 /*end snippet 1*/ /*snippet 2*/ make[4]: *** [drivers/video/console/vgacon.o] Error 1 make[3]: *** [drivers/video/console] Error 2 make[2]: *** [drivers/video] Error 2 make[2]: *** Waiting for unfinished jobs.... CC net/ipv4/fib_semantics.o make[1]: *** [drivers] Error 2 /*End Snippet 2*/ Is there anything I need to know about enabling the virtual console? Thanks for the help, two newbs working together are better than two single newbs. =p __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com |
From: Michael P. <mk...@in...> - 2005-11-17 03:04:49
|
Sounds like progress :) baby-steps the way to go. I was messing with an 800x600 panel from a dead laptop before getting hold of the sony LCD, was interesting to see how changing individual timing values changed the output tho I never quite got it working right. I found it very useful referring to the Intel PXA255 Developer Manual section on LCD control registers and then comparing that with the panel data sheet and the pxafb.txt kernel docs to figure out what's going on between hardware and software. (It's pretty heavy reading, made my head spin but I learned a lot) With the kernel I can tell you what I did which worked for me - not sure exactly what's causing your compile errors: first up I was just compiling the kernel, all fb stuff in kernel and not as modules, didn't compile full root file system so- in gumstix/gumstix-buildroot/build_arm/linux-2.6.11gum/ type make ARCH=arm CROSS_COMPILE=arm-linux- menuconfig go through and set it up how you like then type make ARCH=arm CROSS_COMPILE=arm-linux- uImage to compile. The freshly made kernel is located at gumstix/gumstix-buildroot/build_arm/linux-2.6.11gum/arch/arm/boot/uImage I put this uImage on my mmc card and booted it with appropriate bootargs from u-boot as described in the wiki before making it permanent, there's other ways to do it tho depending on your setup. As far as I remember this worked fine and since then I have rebuilt the full root filesystem incorporating this kernel so ? Have a look through older postings on the forum too as I think others have had similar compile errors You'll be so excited the first time you see that penguin on your gumstix screen I know I was - keep us updated with your progress.. At the moment I'm thinking about how to interface an input device mouse/keyboard whatever with my setup as I have TinyX working and plan to put matchbox on there (ideally aiming at functionality similar to DamnSmallLinux with apps on mmc card) You'll have the advantage of the touch-screen with the ALPS. cheers - Michael ohm unmongkolthavong wrote: >Wow, great information! > >Getting closer (flickering orange screen instead of >nothing), but baby-step speed: I tried to enable the >virtual console through the make menuconfig. Upon cd >../../ and make, I get these errors: > >/*snippet 1*/ >make[1]: Leaving directory >`/home/IQ/Desktop/gumstix/gumstix_builds/gumstix-buildroot/build_arm_nofpu/linux-2.6.11gum' >make: *** >[/home/IQ/Desktop/gumstix/gumstix_builds/gumstix-buildroot/build_arm_nofpu/linux-2.6.11gum/arch/arm/boot/compressed/vmlinux] >Error 2 >/*end snippet 1*/ > >/*snippet 2*/ >make[4]: *** [drivers/video/console/vgacon.o] Error 1 >make[3]: *** [drivers/video/console] Error 2 >make[2]: *** [drivers/video] Error 2 >make[2]: *** Waiting for unfinished jobs.... > CC net/ipv4/fib_semantics.o >make[1]: *** [drivers] Error 2 >/*End Snippet 2*/ > >Is there anything I need to know about enabling the >virtual console? > >Thanks for the help, two newbs working together are >better than two single newbs. =p > > > >__________________________________ >Yahoo! FareChase: Search multiple travel sites in one click. >http://farechase.yahoo.com > > >------------------------------------------------------- >This SF.Net email is sponsored by the JBoss Inc. Get Certified Today >Register for a JBoss Training Course. Free Certification Exam >for All Training Attendees Through End of 2005. For more info visit: >http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click >_______________________________________________ >gumstix-users mailing list >gum...@li... >https://lists.sourceforge.net/lists/listinfo/gumstix-users > > > |
From: Bas v. K. <bas...@gm...> - 2005-11-19 20:29:29
|
On 11/16/05, ohm unmongkolthavong <oun...@ya...> wrote: > ../../ and make, I get these errors: > > make[4]: *** [drivers/video/console/vgacon.o] Error 1 > Make sure that in "Graphics support" -> "Console display driver support", "VGA text Console" is turned OFF in make menuconfig, you only need "Framebuffer console support" to be turned on (I stumbled over this same error as you did before). BTW, I got a Philips LB064V02-A1 LCD working at first try based on the info in this thread, very useful! The LB064V02 is a 640x480 18bits color active TFT screen. My bootargs for this Philips screen are: setenv bootargs console=3Dtty0 console=3DttyS0,115200n8 root=3D1f02 rootfstype=3Djffs2 reboot=3Dcold,hard video=3Dpxafb:mode:640x480-16,active,hsynclen:64,vsynclen:45,left:24,right:= 40,upper:10,lower:33,hsync:0,vsync:0,outputen:1,pixclockpol:0,pixclock:4000= 0 I found out that this screen is quite forgiving in terms of timing, the onl= y settings that really make a difference between functioning and not functioning are (for this Philips screen): mode (if you fill in the wrong widthxheight your display will be running like a mal-adjusted old fashioned b/w tv) outputen (polarity of the data enable line, if this one is wrongly set the lcd shows nothing) pixclockpol (this one indicates whether it should acquire data on a falling or rising edge of the pixclock signal. If this one is wrongly set, the display shows nothing or shows garbage, as the data won't be valid). if pixclock is (very) far off, it doesn't show much on screen either. The rest of the settings mainly concern minimal calibration and are not sho= w stoppers if maladjusted. TIP: if your LCD is making noise (as in a soft hissing sound) and/or shows noise on screen, check if your ground (GND) connections are well connected. Now I've got another phenomena: after about 5 min. the screen goes blank (like as if a screen saver kicked in or so), does anybody know how to avoid this from happening? Anyway, even if the info didn't concern an ALPS display, but I hope this helps nevertheless, best regards, Bas |
From: Michael P. <mk...@in...> - 2005-11-19 22:53:26
|
Bas van Klinkenberg wrote: > > Now I've got another phenomena: after about 5 min. the screen goes > blank (like as if a screen saver kicked in or so), does anybody know > how to avoid this from happening? > Hi Bas - yeah the screen blanking is a default behaviour and not nice with LCDs, you can send some control codes to the console to disable. Check out 'man console_codes' on your linux machine for a full listing there's heaps you can do. echo -ne "\033[9;0]\033[14;0]" > /dev/tty0 switches off both screen blanking and VESA powerdown (havn't actually confirmed that both come into play but no harm disabling both), fixes the problem you describe. You can put this in a startup script somewhere. echo -ne "\033[13]" > /dev/tty0 will unblank an already blanked screen echo -ne "\033c" > /dev/tty0 is handy to clear and reset the screen you can set text colours and other cool stuff. You can send the text output of any program to the screen this way eg ls /bin > /dev/tty0 echo -e "Hello\nWorld!" > /dev/tty0 cat > /dev/tty0 - Michael |
From: Bas v. K. <bas...@gm...> - 2005-11-28 00:14:11
|
On 11/19/05, Michael Petchkovsky <mk...@in...> wrote: Hi Bas - yeah the screen blanking is a default behaviour and not nice > with LCDs, you can send some control codes to the console to disable. > > Check out 'man console_codes' on your linux machine for a full listing > there's heaps you can do. > > echo -ne "\033[9;0]\033[14;0]" > /dev/tty0 > > switches off both screen blanking and VESA powerdown (havn't actually > confirmed that both come into play but no harm disabling both), fixes > the problem you describe. You can put this in a startup script somewhere. > > echo -ne "\033[13]" > /dev/tty0 will unblank an already blanked screen > echo -ne "\033c" > /dev/tty0 is handy to clear and reset the > screen > > you can set text colours and other cool stuff. > > You can send the text output of any program to the screen this way > > eg ls /bin > /dev/tty0 > echo -e "Hello\nWorld!" > /dev/tty0 > cat > /dev/tty0 > > - Michael > Sorry for the late reaction, but thanks a lot for the tips, they work great= ! Regards, Bas |
From: ohm u. <oun...@ya...> - 2005-11-21 21:07:09
|
Hi, Pardon the long post, I think I've exhausted my options (no pun intended) with this LCD screen. # pxaregs lccr LCD Controller Control Register 0 (7-23) LCCR0 0x00300e78 00000000 00110000 00001110 01111000 LCCR0_ENB 0 LCD controller enable LCCR0_CMS 0 LCD monochrome operation enable LCCR0_SDS 0 LCD dual panel display enable LCCR0_LDM 1 LCD disable done IRQ disable LCCR0_SFM 1 LCD start of frame IRQ disable LCCR0_IUM 1 LCD fifo underrun error IRQ disable LCCR0_EFM 1 LCD end of frame IRQ disable LCCR0_PAS 0 LCD active display enable LCCR0_DPD 1 LCD send 8 pixel on L_DD[7:0] at each clock LCCR0_DIS 1 LCD controller disable LCCR0_QDM 1 LCD quick disable IRQ disable LCCR0_PDD 0 LCD palette DMA request delay LCCR0_BM 1 LCD branch start IRQ disable LCCR0_OUM 1 LCD fifo underrun IRQ disable LCD Controller Control Register 1 (7-26) LCCR1 0x010128ef 00000001 00000001 00101000 11101111 LCCR1_PPL 239 LCD pixels per line (+1) LCCR1_HSW 10 LCD horizontal sync pulse width (+1) LCCR1_ELW 1 LCD end of line pixel clock wait count (+1) LCCR1_BLW 1 LCD beginning of line pixel clock wait count (+1) LCD Controller Control Register 2 (7-28) LCCR2 0x0101293f 00000001 00000001 00101001 00111111 LCCR2_LPP 319 LCD lines per panel (+1) LCCR2_VSW 10 LCD vertical sync pulse width (+1) LCCR2_EFW 1 LCD end of frame line clock wait count (+1) LCCR2_BFW 1 LCD beginning of frame line clock wait count (+1) LCD Controller Control Register 3 (7-31) LCCR3 0x03b0001a 00000011 10110000 00000000 00011010 LCCR3_PCD 26 LCD pixel clock divisor (+1) LCCR3_ACB 0 LCD AC bias pin frequency (+1) LCCR3_API 0 LCD AC bias pin transitions per interrupt LCCR3_VSP 1 LCD L_FCLK vertical sync polarity active low LCCR3_HSP 1 LCD L_LCLK horizontal sync polarity active low LCCR3_PCP 0 LCD data sampled on falling edge of L_PCLK LCCR3_OEP 1 LCD L_BIAS output enable active low LCCR3_BPP 8 LCD bits per pixel LCCR3_DPC 0 LCD double pixel clock rate at L_PCLK Above are my values for pxaregs, most of the values are experimental, except for LCCR3_PCD, which I found gave me 60Hz at F_LCLK. Here are my bootargs, again aribtrary because I don't have the timing parameters for the ALPS display: bootargs=console=tty0 console=ttyS0, 115200n8 root=1f02 rootfstype=jffs2 reboot=cold,hardvideo=pxafb:mode:240x320-8,passive,hsynclen:10,left:1,right:5,vsynclen:10 The LCD enable pin (18 DISP_OFF: (L) Display OFF / (H) Display ON) is connected to GPIO16 when I type: echo "GPIO out set" > /proc/gpio/GPIO16 the entire screen turns orange. Sometimes the orange will slowly fade away at the edges until no orange is left. A dd command gives: # dd if=/dev/null of=/dev/tty0 0+0 records in 0+0 records out (or, just to be sure) # dd if=/dev/null of=/dev/fb0 0+0 records in 0+0 records out and a snippet from dmesg: ... pxa2xx-fb pxa2xx-fb: overriding resolution: 240x320 pxa2xx-fb pxa2xx-fb: overriding bit depth: 8 pxa2xx-fb pxa2xx-fb: override hsynclen: 10 pxa2xx-fb pxa2xx-fb: override left: 1 pxa2xx-fb pxa2xx-fb: override right: 5 pxa2xx-fb pxa2xx-fb: override vsynclen: 10 Console: switching to colour frame buffer device 30x40 fb1: Virtual frame buffer device, using 1024K of video memory ... Steps taken so far: ------------------- 1) Tried contacting the author of http://hubbard.engr.scu.edu/embedded/lcd/lfh8p4032b/lfh8p4032b.html ... no response 2)Tried contacting ALPS, was told the item was discontinued so they cannot support it anymore. (sounds like a cop-out to me... All I was asking for were datasheets) 3) Contacted HCS (distributor). Tried to help, but did not have datasheets. Not too much info was gained. 4) Enabled all recommended settings within make menuconfig as noted on the wiki. 5) Enabled virtual console 6) DISabled the VGA text console in make menuconfig due to compiling errors. 7) Ordered a Sony screen, as mentioned on the wiki =p 8) Waiting for breakout board to become available again. Questions: ---------- 1) Is there any way to get close to the pixel clock value without random guess-checking? 2) Can anyone supply a root_fs_arm_nofpu file that works with their own LCD screen? I would like to know where the problem is. 3) Is this simply a timing problem? What other considerations must I make when trying to find these timing parameters? Thank you everyone, this will be a good project to complete since it seems a lot of people are having trouble with this screen. Regards, Ohm __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com |
From: Bas v. K. <bas...@gm...> - 2005-11-22 17:51:12
|
A dd command gives: > # dd if=3D/dev/null of=3D/dev/tty0 > 0+0 records in > 0+0 records out To test the framebuffer allocation you should do the following: dd if=3D/dev/fb0 of=3D/dev/null, so that you are actually writing from /dev= /fb0 to /dev/null (not from /dev/null to /dev/fb0 as you did) Then the complete framebuffer will be dumped to /dev/null, and the amount o= f records dumped will be shown. The result should be 150+0 records for your application (240 times 320 time= s 1 byte (8bit color) divided by 512 (standard blocksize) =3D 150 records) For the rest, I can't really see anything wrong with the pxaregs. I assume you already tried playing with pixclock polarity and that you checked with the oscilloscope that you actually receive good signals on the correct LCD pins? Regards, Bas |
From: Jim W. <jwa...@qu...> - 2005-11-28 16:44:38
|
Hi, Here comes another long post! When I run the same pxaregs command, I show the LCD as being enabled on my system, where your LCD looks to be not enabled in LCCR0_ENB. # ./pxaregs lccr LCD Controller Control Register 0 (7-23) LCCR0 0x003008f9 00000000 00110000 00001000 11111001 LCCR0_ENB 1 LCD controller enable LCCR0_CMS 0 LCD monochrome operation enable LCCR0_SDS 0 LCD dual panel display enable LCCR0_LDM 1 LCD disable done IRQ disable LCCR0_SFM 1 LCD start of frame IRQ disable LCCR0_IUM 1 LCD fifo underrun error IRQ disable LCCR0_EFM 1 LCD end of frame IRQ disable LCCR0_PAS 1 LCD active display enable LCCR0_DPD 0 LCD send 8 pixel on L_DD[7:0] at each clock LCCR0_DIS 0 LCD controller disable LCCR0_QDM 1 LCD quick disable IRQ disable LCCR0_PDD 0 LCD palette DMA request delay LCCR0_BM 1 LCD branch start IRQ disable LCCR0_OUM 1 LCD fifo underrun IRQ disable LCD Controller Control Register 1 (7-26) LCCR1 0x630ade7f 01100011 00001010 11011110 01111111 LCCR1_PPL 639 LCD pixels per line (+1) LCCR1_HSW 55 LCD horizontal sync pulse width (+1) LCCR1_ELW 10 LCD end of line pixel clock wait count (+1) LCCR1_BLW 99 LCD beginning of line pixel clock wait count (+1) LCD Controller Control Register 2 (7-28) LCCR2 0x220201df 00100010 00000010 00000001 11011111 LCCR2_LPP 479 LCD lines per panel (+1) LCCR2_VSW 0 LCD vertical sync pulse width (+1) LCCR2_EFW 2 LCD end of frame line clock wait count (+1) LCCR2_BFW 34 LCD beginning of frame line clock wait count (+1) LCD Controller Control Register 3 (7-31) LCCR3 0x04400001 00000100 01000000 00000000 00000001 LCCR3_PCD 1 LCD pixel clock divisor (+1) LCCR3_ACB 0 LCD AC bias pin frequency (+1) LCCR3_API 0 LCD AC bias pin transitions per interrupt LCCR3_VSP 0 LCD L_FCLK vertical sync polarity active low LCCR3_HSP 0 LCD L_LCLK horizontal sync polarity active low LCCR3_PCP 1 LCD data sampled on falling edge of L_PCLK LCCR3_OEP 0 LCD L_BIAS output enable active low LCCR3_BPP 16 LCD bits per pixel LCCR3_DPC 0 LCD double pixel clock rate at L_PCLK # I have put up a rootfs file that you can try out, if you like. It is really old! The LCD should be enabled, but I didn't actually test this particular file on the gumstix with the LCD because I didn't want to mess up some of the other settings I've changed on that particular board. You should be able to run the pxaregs program (which may not be included in the buildroot, it's on my MMC card) and get the same (or similar) results as I do. The file is located at http://www.quad-c.com/root_fs_arm and I will leave it up there for a while. Also, it looks like I screwed up the dd command that I gave you! Someone else on the list came to my rescue and posted the correct command, so I didn't react (other than groaning). Thanks, Jim Walters ohm unmongkolthavong wrote: >Hi, >Pardon the long post, I think I've exhausted my >options (no pun intended) with this LCD screen. > ># pxaregs lccr > >LCD Controller Control Register 0 (7-23) >LCCR0 0x00300e78 00000000 00110000 >00001110 01111000 >LCCR0_ENB 0 LCD controller enable >LCCR0_CMS 0 LCD monochrome operation >enable >LCCR0_SDS 0 LCD dual panel display >enable >LCCR0_LDM 1 LCD disable done IRQ >disable >LCCR0_SFM 1 LCD start of frame IRQ >disable >LCCR0_IUM 1 LCD fifo underrun error >IRQ disable >LCCR0_EFM 1 LCD end of frame IRQ >disable >LCCR0_PAS 0 LCD active display enable >LCCR0_DPD 1 LCD send 8 pixel on >L_DD[7:0] at each clock >LCCR0_DIS 1 LCD controller disable >LCCR0_QDM 1 LCD quick disable IRQ >disable >LCCR0_PDD 0 LCD palette DMA request >delay >LCCR0_BM 1 LCD branch start IRQ >disable >LCCR0_OUM 1 LCD fifo underrun IRQ >disable > >LCD Controller Control Register 1 (7-26) >LCCR1 0x010128ef 00000001 00000001 >00101000 11101111 >LCCR1_PPL 239 LCD pixels per line (+1) >LCCR1_HSW 10 LCD horizontal sync pulse >width (+1) >LCCR1_ELW 1 LCD end of line pixel >clock wait count (+1) >LCCR1_BLW 1 LCD beginning of line >pixel clock wait count (+1) > >LCD Controller Control Register 2 (7-28) >LCCR2 0x0101293f 00000001 00000001 >00101001 00111111 >LCCR2_LPP 319 LCD lines per panel (+1) >LCCR2_VSW 10 LCD vertical sync pulse >width (+1) >LCCR2_EFW 1 LCD end of frame line >clock wait count (+1) >LCCR2_BFW 1 LCD beginning of frame >line clock wait count (+1) > >LCD Controller Control Register 3 (7-31) >LCCR3 0x03b0001a 00000011 10110000 >00000000 00011010 >LCCR3_PCD 26 LCD pixel clock divisor >(+1) >LCCR3_ACB 0 LCD AC bias pin frequency >(+1) >LCCR3_API 0 LCD AC bias pin >transitions per interrupt >LCCR3_VSP 1 LCD L_FCLK vertical sync >polarity active low >LCCR3_HSP 1 LCD L_LCLK horizontal sync >polarity active low >LCCR3_PCP 0 LCD data sampled on >falling edge of L_PCLK >LCCR3_OEP 1 LCD L_BIAS output enable >active low >LCCR3_BPP 8 LCD bits per pixel >LCCR3_DPC 0 LCD double pixel clock >rate at L_PCLK > >Above are my values for pxaregs, most of the values >are experimental, except for LCCR3_PCD, which I found >gave me 60Hz at F_LCLK. > >Here are my bootargs, again aribtrary because I don't >have the timing parameters for the ALPS display: > >bootargs=console=tty0 console=ttyS0, 115200n8 >root=1f02 rootfstype=jffs2 >reboot=cold,hardvideo=pxafb:mode:240x320-8,passive,hsynclen:10,left:1,right:5,vsynclen:10 > >The LCD enable pin (18 DISP_OFF: (L) Display OFF / (H) >Display ON) is connected to GPIO16 >when I type: echo "GPIO out set" > /proc/gpio/GPIO16 >the entire screen turns orange. Sometimes the orange >will slowly fade away at the edges until no orange is >left. > >A dd command gives: ># dd if=/dev/null of=/dev/tty0 >0+0 records in >0+0 records out > >(or, just to be sure) > ># dd if=/dev/null of=/dev/fb0 >0+0 records in >0+0 records out > >and a snippet from dmesg: > >... >pxa2xx-fb pxa2xx-fb: overriding resolution: 240x320 >pxa2xx-fb pxa2xx-fb: overriding bit depth: 8 >pxa2xx-fb pxa2xx-fb: override hsynclen: 10 >pxa2xx-fb pxa2xx-fb: override left: 1 >pxa2xx-fb pxa2xx-fb: override right: 5 >pxa2xx-fb pxa2xx-fb: override vsynclen: 10 >Console: switching to colour frame buffer device 30x40 >fb1: Virtual frame buffer device, using 1024K of video >memory >... > >Steps taken so far: >------------------- >1) Tried contacting the author of >http://hubbard.engr.scu.edu/embedded/lcd/lfh8p4032b/lfh8p4032b.html >... no response > >2)Tried contacting ALPS, was told the item was >discontinued so they cannot support it anymore. >(sounds like a cop-out to me... All I was asking for >were datasheets) > >3) Contacted HCS (distributor). Tried to help, but did >not have datasheets. Not too much info was gained. > >4) Enabled all recommended settings within make >menuconfig as noted on the wiki. > >5) Enabled virtual console > >6) DISabled the VGA text console in make menuconfig >due to compiling errors. > >7) Ordered a Sony screen, as mentioned on the wiki =p > >8) Waiting for breakout board to become available >again. > >Questions: >---------- >1) Is there any way to get close to the pixel clock >value without random guess-checking? > >2) Can anyone supply a root_fs_arm_nofpu file that >works with their own LCD screen? I would like to know >where the problem is. > >3) Is this simply a timing problem? What other >considerations must I make when trying to find these >timing parameters? > >Thank you everyone, this will be a good project to >complete since it seems a lot of people are having >trouble with this screen. > >Regards, >Ohm > > > > > > >__________________________________ >Yahoo! Mail - PC Magazine Editors' Choice 2005 >http://mail.yahoo.com > > >------------------------------------------------------- >This SF.Net email is sponsored by the JBoss Inc. Get Certified Today >Register for a JBoss Training Course. Free Certification Exam >for All Training Attendees Through End of 2005. For more info visit: >http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click >_______________________________________________ >gumstix-users mailing list >gum...@li... >https://lists.sourceforge.net/lists/listinfo/gumstix-users > > |
From: ohm u. <oun...@ya...> - 2005-11-28 20:39:43
|
Mr. Walters, I checked out your fs image. It gives me a nice cream colored screen with moving horizontal scan lines. I then checked the pxaregs and everything looked fine (controller was ENABLED, this time =p ) I checked the data lines with an O'scope and found that I have no data signals coming out! Am using a custom daughterboard, so I checked out the connections and they all lead to the correct gumstix pins. So I'm confused yet again. Why are there no data signals coming out of LDD0:7 ? I'll reload my old fs image and see if I have the same problem there. This ALPS screen will not get the best of us!! Regards, Ohm. Date: Mon, 28 Nov 2005 10:44:46 -0600 From: Jim Walters <jwa...@qu...> To: gum...@li... Subject: Re: [Gumstix-users] Re: Find LCD timing for the ALPS on gumstix.org Reply-To: gum...@li... Hi, Here comes another long post! When I run the same pxaregs command, I show the LCD as being enabled on my system, where your LCD looks to be not enabled in LCCR0_ENB. # ./pxaregs lccr LCD Controller Control Register 0 (7-23) LCCR0 0x003008f9 00000000 00110000 00001000 11111001 LCCR0_ENB 1 LCD controller enable LCCR0_CMS 0 LCD monochrome operation enable LCCR0_SDS 0 LCD dual panel display enable LCCR0_LDM 1 LCD disable done IRQ disable LCCR0_SFM 1 LCD start of frame IRQ disable LCCR0_IUM 1 LCD fifo underrun error IRQ disable LCCR0_EFM 1 LCD end of frame IRQ disable LCCR0_PAS 1 LCD active display enable LCCR0_DPD 0 LCD send 8 pixel on L_DD[7:0] at each clock LCCR0_DIS 0 LCD controller disable LCCR0_QDM 1 LCD quick disable IRQ disable LCCR0_PDD 0 LCD palette DMA request delay LCCR0_BM 1 LCD branch start IRQ disable LCCR0_OUM 1 LCD fifo underrun IRQ disable LCD Controller Control Register 1 (7-26) LCCR1 0x630ade7f 01100011 00001010 11011110 01111111 LCCR1_PPL 639 LCD pixels per line (+1) LCCR1_HSW 55 LCD horizontal sync pulse width (+1) LCCR1_ELW 10 LCD end of line pixel clock wait count (+1) LCCR1_BLW 99 LCD beginning of line pixel clock wait count (+1) LCD Controller Control Register 2 (7-28) LCCR2 0x220201df 00100010 00000010 00000001 11011111 LCCR2_LPP 479 LCD lines per panel (+1) LCCR2_VSW 0 LCD vertical sync pulse width (+1) LCCR2_EFW 2 LCD end of frame line clock wait count (+1) LCCR2_BFW 34 LCD beginning of frame line clock wait count (+1) LCD Controller Control Register 3 (7-31) LCCR3 0x04400001 00000100 01000000 00000000 00000001 LCCR3_PCD 1 LCD pixel clock divisor (+1) LCCR3_ACB 0 LCD AC bias pin frequency (+1) LCCR3_API 0 LCD AC bias pin transitions per interrupt LCCR3_VSP 0 LCD L_FCLK vertical sync polarity active low LCCR3_HSP 0 LCD L_LCLK horizontal sync polarity active low LCCR3_PCP 1 LCD data sampled on falling edge of L_PCLK LCCR3_OEP 0 LCD L_BIAS output enable active low LCCR3_BPP 16 LCD bits per pixel LCCR3_DPC 0 LCD double pixel clock rate at L_PCLK # I have put up a rootfs file that you can try out, if you like. It is really old! The LCD should be enabled, but I didn't actually test this particular file on the gumstix with the LCD because I didn't want to mess up some of the other settings I've changed on that particular board. You should be able to run the pxaregs program (which may not be included in the buildroot, it's on my MMC card) and get the same (or similar) results as I do. The file is located at http://www.quad-c.com/root_fs_arm and I will leave it up there for a while. Also, it looks like I screwed up the dd command that I gave you! Someone else on the list came to my rescue and posted the correct command, so I didn't react (other than groaning). Thanks, Jim Walters __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com |
From: N.E.Whiteford <N.E...@so...> - 2005-11-28 21:06:37
|
Sorry, I've not ben keeping track of this thread, but are you using 8bit colour mode in pxafb? IIRC nothing below 16bit is currently working in pxafb, try switching to 16bit mode, you should get data coming out then. nav On Mon, 28 Nov 2005, ohm unmongkolthavong wrote: > Mr. Walters, > > I checked out your fs image. It gives me a nice cream > colored screen with moving horizontal scan lines. > > I then checked the pxaregs and everything looked fine > (controller was ENABLED, this time =p ) > > I checked the data lines with an O'scope and found > that I have no data signals coming out! Am using a > custom daughterboard, so I checked out the connections > and they all lead to the correct gumstix pins. > > So I'm confused yet again. Why are there no data > signals coming out of LDD0:7 ? I'll reload my old fs > image and see if I have the same problem there. > > This ALPS screen will not get the best of us!! > > Regards, > Ohm. > > Date: Mon, 28 Nov 2005 10:44:46 -0600 > From: Jim Walters <jwa...@qu...> > To: gum...@li... > Subject: Re: [Gumstix-users] Re: Find LCD timing for > the ALPS on > gumstix.org > Reply-To: gum...@li... > > Hi, > > Here comes another long post! When I run the same > pxaregs command, I > show the LCD as being enabled on my system, where your > LCD looks to be > not enabled in LCCR0_ENB. > > # ./pxaregs lccr > > LCD Controller Control Register 0 (7-23) > LCCR0 0x003008f9 00000000 00110000 > 00001000 > 11111001 > LCCR0_ENB 1 LCD controller > enable > LCCR0_CMS 0 LCD monochrome > operation enable > LCCR0_SDS 0 LCD dual panel > display enable > LCCR0_LDM 1 LCD disable done > IRQ disable > LCCR0_SFM 1 LCD start of > frame IRQ disable > LCCR0_IUM 1 LCD fifo underrun > error IRQ > disable > LCCR0_EFM 1 LCD end of frame > IRQ disable > LCCR0_PAS 1 LCD active > display enable > LCCR0_DPD 0 LCD send 8 pixel > on L_DD[7:0] at > each clock > LCCR0_DIS 0 LCD controller > disable > LCCR0_QDM 1 LCD quick disable > IRQ disable > LCCR0_PDD 0 LCD palette DMA > request delay > LCCR0_BM 1 LCD branch start > IRQ disable > LCCR0_OUM 1 LCD fifo underrun > IRQ disable > > LCD Controller Control Register 1 (7-26) > LCCR1 0x630ade7f 01100011 00001010 > 11011110 > 01111111 > LCCR1_PPL 639 LCD pixels per > line (+1) > LCCR1_HSW 55 LCD horizontal > sync pulse width > (+1) > LCCR1_ELW 10 LCD end of line > pixel clock wait > count (+1) > LCCR1_BLW 99 LCD beginning of > line pixel clock > wait count (+1) > > LCD Controller Control Register 2 (7-28) > LCCR2 0x220201df 00100010 00000010 > 00000001 > 11011111 > LCCR2_LPP 479 LCD lines per > panel (+1) > LCCR2_VSW 0 LCD vertical sync > pulse width (+1) > LCCR2_EFW 2 LCD end of frame > line clock wait > count (+1) > LCCR2_BFW 34 LCD beginning of > frame line clock > wait count (+1) > > LCD Controller Control Register 3 (7-31) > LCCR3 0x04400001 00000100 01000000 > 00000000 > 00000001 > LCCR3_PCD 1 LCD pixel clock > divisor (+1) > LCCR3_ACB 0 LCD AC bias pin > frequency (+1) > LCCR3_API 0 LCD AC bias pin > transitions per > interrupt > LCCR3_VSP 0 LCD L_FCLK > vertical sync polarity > active low > LCCR3_HSP 0 LCD L_LCLK > horizontal sync > polarity > active low > LCCR3_PCP 1 LCD data sampled > on falling edge > of > L_PCLK > LCCR3_OEP 0 LCD L_BIAS output > enable active > low > LCCR3_BPP 16 LCD bits per > pixel > LCCR3_DPC 0 LCD double pixel > clock rate at > L_PCLK > # > > I have put up a rootfs file that you can try out, if > you like. It is > really old! The LCD should be enabled, but I didn't > actually test this > particular file on the gumstix with the LCD because I > didn't want to > mess up some of the other settings I've changed on > that particular > board. You should be able to run the pxaregs program > (which may not be > included in the buildroot, it's on my MMC card) and > get the same (or > similar) results as I do. The file is located at > http://www.quad-c.com/root_fs_arm and I will leave it > up there for a > while. > > Also, it looks like I screwed up the dd command that I > gave you! > Someone else on the list came to my rescue and posted > the correct > command, so I didn't react (other than groaning). > > Thanks, > > Jim Walters > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > |
From: ohm u. <oun...@ya...> - 2005-11-30 21:29:37
|
Enter our friendly penguin!! So the penguin is now displayed after about 5 weeks of timing issues. However, the only thing i changed right before it worked was 240x320-8 to 240x320-16 in the bootargs ! Though, timing issues still abound. I have a flickering penguin and the text fades out from left to right. I will put up some pics when I get home tonight, but I will try my best to describe it now: alternating bright and dark columns from left to right, and the columns alternate from darker to darkest the closer you get to the right side of the screen. Also, Horizontal scanning lines at a rate of about 3-4 per second moving from top to bottom. Since I don't want to go on another wild goose chase, I will leave it up to the experts. I do'nt know what to change to compensate for the symptoms. Like I said, I will post up some pictures tonight so you get a better undersatnding of the symptoms. Thanks for the help, Jim, Mike, Bas, Nav, Holly and all others for the extra help through this. Ohm. __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ |
From: Jim W. <jwa...@qu...> - 2005-12-01 00:16:59
|
Hi, Now that you have something visible on the screen, you can use pxaregs to change the LCD timing parameters from the command line. It works really well, since you can specify the portions of each control register to change and see the results immediately. Given that you have some bands moving on the screen, I would try changing the pixel clock and pixel clock divider first and see if that makes them go away. Have you ever really chased wild geese? It's actually kind of fun! Thanks, Jim ohm unmongkolthavong wrote: >Enter our friendly penguin!! > >So the penguin is now displayed after about 5 weeks of >timing issues. However, the only thing i changed right >before it worked was 240x320-8 to 240x320-16 in the >bootargs ! > >Though, timing issues still abound. I have a >flickering penguin and the text fades out from left to >right. > >I will put up some pics when I get home tonight, but I >will try my best to describe it now: > >alternating bright and dark columns from left to >right, and the columns alternate from darker to >darkest the closer you get to the right side of the >screen. Also, Horizontal scanning lines at a rate of >about 3-4 per second moving from top to bottom. > >Since I don't want to go on another wild goose chase, >I will leave it up to the experts. I do'nt know what >to change to compensate for the symptoms. > >Like I said, I will post up some pictures tonight so >you get a better undersatnding of the symptoms. > >Thanks for the help, Jim, Mike, Bas, Nav, Holly and >all others for the extra help through this. > >Ohm. > > > >__________________________________ >Yahoo! Music Unlimited >Access over 1 million songs. Try it free. >http://music.yahoo.com/unlimited/ > > >------------------------------------------------------- >This SF.net email is sponsored by: Splunk Inc. Do you grep through log files >for problems? Stop! Download the new AJAX search engine that makes >searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! >http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click >_______________________________________________ >gumstix-users mailing list >gum...@li... >https://lists.sourceforge.net/lists/listinfo/gumstix-users > > |
From: Michael P. <mk...@in...> - 2005-12-01 09:07:15
|
>ohm unmongkolthavong wrote: >Enter our friendly penguin!! Hooray for the penguin! as Jim says - much easier now you've got something to look at on screen. (sorry nothing useful to add but you're almost there now:) re: wild geese, Jim that does sound like fun :) Ohm, you might have to chase that penguin round the screen a while before you nail those values - congrats on the progress! Jim Walters wrote: > Hi, > > Now that you have something visible on the screen, you can use pxaregs > to change the LCD timing parameters from the command line. It works > really well, since you can specify the portions of each control > register to change and see the results immediately. > > Given that you have some bands moving on the screen, I would try > changing the pixel clock and pixel clock divider first and see if that > makes them go away. > > Have you ever really chased wild geese? It's actually kind of fun! > > Thanks, > > Jim > > ohm unmongkolthavong wrote: > >> Enter our friendly penguin!! >> So the penguin is now displayed after about 5 weeks of >> timing issues. However, the only thing i changed right >> before it worked was 240x320-8 to 240x320-16 in the >> bootargs ! >> >> Though, timing issues still abound. I have a >> flickering penguin and the text fades out from left to >> right. >> I will put up some pics when I get home tonight, but I >> will try my best to describe it now: >> alternating bright and dark columns from left to >> right, and the columns alternate from darker to >> darkest the closer you get to the right side of the >> screen. Also, Horizontal scanning lines at a rate of >> about 3-4 per second moving from top to bottom. >> Since I don't want to go on another wild goose chase, >> I will leave it up to the experts. I do'nt know what >> to change to compensate for the symptoms. >> Like I said, I will post up some pictures tonight so >> you get a better undersatnding of the symptoms. >> Thanks for the help, Jim, Mike, Bas, Nav, Holly and >> all others for the extra help through this. >> Ohm. >> >> >> __________________________________ Yahoo! Music Unlimited Access over >> 1 million songs. Try it free. http://music.yahoo.com/unlimited/ >> >> >> ------------------------------------------------------- >> This SF.net email is sponsored by: Splunk Inc. Do you grep through >> log files >> for problems? Stop! Download the new AJAX search engine that makes >> searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! >> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click >> _______________________________________________ >> gumstix-users mailing list >> gum...@li... >> https://lists.sourceforge.net/lists/listinfo/gumstix-users >> >> > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > |
From: ohm u. <oun...@ya...> - 2005-12-02 03:41:56
|
Hi all, I got nipped by a goose as a child, so I don't go near them =p j/k I just don't know what to really change, and what combinations to try. I FINALLY posted up some pics. The camera doesn't show the white horizontal scanning line that goes from top to bottom 2-3 times a second. however, notice the dark columns. pics: http://mason.gmu.edu/~ounmongk/penguin1.jpg http://mason.gmu.edu/~ounmongk/penguin2.jpg I will spend this Friday (tomorrow) playing with the pxaregs. Seriously, I almost fell out of my chair when I saw our little penguin. And as promised, timing specs will be posted once I find the right ones! =) Thanks a bunch, everyone. Credit will be given where it's due! Ohm __________________________________________ Yahoo! DSL Something to write home about. Just $16.99/mo. or less. dsl.yahoo.com |
From: Doug S. <do...@pr...> - 2005-12-02 12:44:57
|
Ohm, How long is the cable from the gumstix to the LCD? I had similar output on an embedded PC because my custom made cable was too long. These LCD connections are high speed and the signal attenuates quickly. Also the quality of the joints matters, are they hand soldered? -- Doug ohm unmongkolthavong wrote: > Hi all, > > I got nipped by a goose as a child, so I don't go near > them =p > > j/k I just don't know what to really change, and what > combinations to try. I FINALLY posted up some pics. > The camera doesn't show the white horizontal scanning > line that goes from top to bottom 2-3 times a second. > however, notice the dark columns. > > pics: > http://mason.gmu.edu/~ounmongk/penguin1.jpg > http://mason.gmu.edu/~ounmongk/penguin2.jpg > > I will spend this Friday (tomorrow) playing with the > pxaregs. Seriously, I almost fell out of my chair when > I saw our little penguin. ab > > And as promised, timing specs will be posted once I > find the right ones! =) > > Thanks a bunch, everyone. Credit will be given where > it's due! > Ohm > > > > __________________________________________ > Yahoo! DSL – Something to write home about. > Just $16.99/mo. or less. > dsl.yahoo.com > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > > > |
From: ohm u. <oun...@ya...> - 2005-12-09 19:35:08
|
Hi Mr. Sutherland, The cable is about 4.5" long. the connectors are the back of the LCD screen and a custom daughtboard with a 24-pin wire connector. I checked the connections and they are all sound and solid. I can get it to a point where i still lose one side of the screen due to bad timing, but the text is now readable. I wrote a little script that changed the LCCR3_PCD from 0 to 255, hoping that somewhere along script, the screen would be readable. However, the loss of the screen (graying of a the screen) goes from the left to the right as the numbers increase. -Can anyone suggest another path to take towards finding the bug-free display? -Is there a way to change the actual pixel clock using pxaregs? I was wondering if there was another register that handled the clock. I'm reading the PXA developer's manual, and as mentioned before, it really DOES make your head spin. Hopefully, though, i'll find some info on the pixel clock. While we're here, I noticed that text that i type does NOT show up on the screen. It only prints out the boot-up messages (the dmesg stuff) and then stops there. I will do a search around the archives to find some info, but if anyone could point me in the right direction, I would totally appreciate it. Thanks for everyone's help! I hope this newb can be like the LCD masters here one day. Ohm __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com |
From: Michael P. <mk...@in...> - 2005-12-10 00:22:17
|
hi Ohm, the photos you posted look quite good in that you have real data going to the screen, colors are good, pixels are ordered and text can be made out and horiz and vert syncing seems ok ie things appear in the right place on screen. I suspect the problem with the 'faded' look could be a polarity issue. Look at the three polarity settings pixclockpol (reg LCCR3_PCP) hsync (LCCR3_HSP) and vsync (LCCR3_VSP). Each of these can be one or zero. I would try all eight combinations from 0,0,0 .... to 1,1,1 see what changes and what looks best. (Of course I have to add the disclaimer - if the magic smoke gets out it's not my fault!) Maybe you've tried this already. If you hit the right combination and the text looks crisp it may throw the syncing right out, then you have to tweak hsynclen (LCCR1_HSW) vsynclen(LCCR1_VSW) left (LCCR1_BLW) and right (LCCR1_ELW) to straighten things up According to the Intel docs upper (LCCR2_BFW) and lower (LCCR2_EFW) should be set to zero for a passive screen and vsynclen(LCCR1_VSW) should be increased to allow wait states at the start of frame but YMMV Also you can send text to the screen like this: echo "Hello World" > /dev/tty0 there's actually a lot you can do with this redirection method, you can send the text output of any program to the LCD by appending ' > /dev/tty0' It would be cool if what was typed at the serial console appeared directly on the text console, any linux console gurus know how to do this ? Is it something trivial or does it require extensive kernel hacking :) ;) in all my searching of console-HOWTOS and docs I havn't come across an answer to this one. Another question to the gurus, should I be using /dev/tty1 /dev/console or /dev/tty0 in the above examples (or just whatever works) ? cheers, Michael ohm unmongkolthavong wrote: >Hi Mr. Sutherland, > >The cable is about 4.5" long. the connectors are the >back of the LCD screen and a custom daughtboard with a >24-pin wire connector. > >I checked the connections and they are all sound and >solid. > >I can get it to a point where i still lose one side of >the screen due to bad timing, but the text is now >readable. I wrote a little script that changed the >LCCR3_PCD from 0 to 255, hoping that somewhere along >script, the screen would be readable. However, the >loss of the screen (graying of a the screen) goes from >the left to the right as the numbers increase. > >-Can anyone suggest another path to take towards >finding the bug-free display? >-Is there a way to change the actual pixel clock using >pxaregs? I was wondering if there was another register >that handled the clock. I'm reading the PXA >developer's manual, and as mentioned before, it really >DOES make your head spin. Hopefully, though, i'll find >some info on the pixel clock. > >While we're here, I noticed that text that i type does >NOT show up on the screen. It only prints out the >boot-up messages (the dmesg stuff) and then stops >there. I will do a search around the archives to find >some info, but if anyone could point me in the right >direction, I would totally appreciate it. > >Thanks for everyone's help! I hope this newb can be >like the LCD masters here one day. >Ohm > >__________________________________________________ >Do You Yahoo!? >Tired of spam? Yahoo! Mail has the best spam protection around >http://mail.yahoo.com > > >------------------------------------------------------- >This SF.net email is sponsored by: Splunk Inc. Do you grep through log files >for problems? Stop! Download the new AJAX search engine that makes >searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! >http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click >_______________________________________________ >gumstix-users mailing list >gum...@li... >https://lists.sourceforge.net/lists/listinfo/gumstix-users > > > |
From: Craig H. <cr...@gu...> - 2005-12-10 00:53:31
|
On Dec 9, 2005, at 3:22 PM, Michael Petchkovsky wrote: > It would be cool if what was typed at the serial console appeared > directly on the text console, any linux console gurus know how to > do this ? Is it something trivial or does it require extensive > kernel hacking :) ;) in all my searching of console-HOWTOS and docs > I havn't come across an answer to this one. You could probably do this pretty easily by having socat just copy / dev/ttyS0 to /dev/tty0; or else you could just modify the bootarg variable in u-boot (after you get the screen working of course) to have console=/dev/tty0 instead of console=/dev/ttyS0 C |
From: Michael P. <mk...@in...> - 2005-12-10 04:19:44
|
Thanks Craig OK I dumped console=/dev/ttyS0 from bootargs, this suppressed kernel messages on that console (of course getty comes up as per inittab so login and bash remain on ttyS0) and left tty0 as the system default console. The last mentioned console=... argument the kernel receives becomes the default console so putting console=/dev/tty0 after console=/dev/ttyS0 has the effect that both receive kernel messages and tty0 is regarded as the default system console. To get ttyS0's STDIO echoed on tty0 it looks like socat may be the answer, I was pleased to see it in the buildroot and compiled no problems. Could take me a while to figure out the commandline options though :) When you hook up an LCD and see text output on it there's this visceral expectation that your keypresses (on a separate machine over a serial line) should be visible immediately, when you think about it it's obvious why they're not. I guess this could be useful to someone with a setup like Machine A (gumstix with attached LCD) + Machine B (no graphics capabilities but a keyboard and serial link to Machine A). I can live without the feature though. BTW I cross-compiled xvkbd (a virtual X keyboard) and have it running on tinyX, could be a useful app for people who get touchscreens going. Michael Craig Hughes wrote: > > You could probably do this pretty easily by having socat just copy / > dev/ttyS0 to /dev/tty0; or else you could just modify the bootarg > variable in u-boot (after you get the screen working of course) to > have console=/dev/tty0 instead of console=/dev/ttyS0 > > C > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > |
From: Michael P. <mk...@in...> - 2005-12-10 08:00:09
|
Doh! - just occurred to me to redirect the shell to /dev/tty0 so from login on /dev/ttyS0 #sh > /dev/tty0 now all your keystrokes go straight up on the lcd screen, you can type commands and see their results all on the gumstix screen. there would be neat ways to handle this through inittab or something to make it permanent if you wanted that behaviour. Michael |