CD32 Graphics Fault

Dipski · 03 July 2018, 01:07

I've got a CD32 with a strange graphics fault. It happens with vertical columns on static sprites/graphics. On the loading screen the Amiga CD32 logo and the disc image appear normally as they spin & move, but when they stop and it becomes a static screen there are vertical columns that appear corrupted and have parts of the background there instead.

I can't seem to find any issues on the address lines or video signals on the oscilloscope, so I'm leaning towards a fault with the Lisa chip as it handles sprites (I'm assuming the logos are sprites). Could any other chip or fault cause the issue and is there any way to confirm the fault before finding a replacement Lisa?

solidcore · 03 July 2018, 01:36

Quote:

Originally Posted by Dipski

I've got a CD32 with a strange graphics fault. It happens with vertical columns on static sprites/graphics. On the loading screen the Amiga CD32 logo and the disc image appear normally as they spin & move, but when they stop and it becomes a static screen there are vertical columns that appear corrupted and have parts of the background there instead.

I can't seem to find any issues on the address lines or video signals on the oscilloscope, so I'm leaning towards a fault with the Lisa chip as it handles sprites (I'm assuming the logos are sprites). Could any other chip or fault cause the issue and is there any way to confirm the fault before finding a replacement Lisa?

Failing lisa.. you checked the ram?

Dipski · 03 July 2018, 02:32

Yeah I've checked the RAM, everything looks normal there. With the vertical bands there are 20 in total, every second one with glitched sprite graphics (only when static) and normal background graphics.. I can't help but feel like I'm missing something.

dlfrsilver · 03 July 2018, 14:25

those are jailbars, check the RAMs and ROM of the CD32.

supaduper · 03 July 2018, 20:08

What about the excellent diagarom, you can burn an eprom for the CD32
http://eab.abime.net/showthread.php?t=89183

Toni Wilen · 03 July 2018, 20:31

ROM can't cause this. It can be memory (It isn't if games boot without crashing) or Alice (It most likely isn't if sprites work fine, check if NVRAM save screen "key" sprite is corrupted). It is most likely Lisa if everything else works fine.

Dipski · 04 July 2018, 03:22

Thanks guys, interestingly the NVRAM save screen key is fine - and the games I have load and run fine (Alien Breed / Qwak) with no graphics glitches. This would seem to rule out a RAM issue.

I did test with another ROM in the CD32 but had the same glitch on the loading screen, and the ROM socket looked fine.

I'll check the ROM socket and its connections further. My main concern is if it's specific to a picture mode (for games/programs I don't have to test it with) if I can't find any fault associated with the ROM.

Dipski · 04 July 2018, 06:20

The ROM socket tests fine and continuity for all of the connections from the ROM is there, the signals all look normal. I've checked continuity between Alice and Lisa, checked the signals there which all look normal as well. I've got some EPROM's on their way so I can burn a CD32 DiagROM so if everything passes on the DiagROM that should rule out any memory fault conclusively and point to a Lisa issue. I feel like there's something obvious that I haven't checked on the board but nothing else stands out going over the schematics.

Toni Wilen · 04 July 2018, 08:47

Try some "true" AGA game, not OCS/ECS port, game that uses higher bandwidth AGA modes because problem only seems to affect them (every other 16 pixel block is corrupted) and only bitplanes because "key" is fine (it is made from sprites).

Which points to fault inside Lisa. It can't be Alice or Alice to Lisa data bus because sprites work.

supaduper · 04 July 2018, 12:18

@ Dipski, Just to add if you are burning Diagrom to a 27c800 using the TL866 and the 42pin adapter like I have, then after flashing you will have to bend pin 31 up so its not connected for it to work on the CD32

Dipski · 04 July 2018, 12:24

Quote:

Originally Posted by supaduper

Just to add if you are burning Diagrom to a 27c800 using the TL866 and the 42pin adapter like I have, then after flashing you will have to bend pin 31 up so its not connected for it to work on the CD32

Yes spot on, I've got a TL866 and have 27c800's on the way and did not know about pin 31 needing to be disconnected - thankyou!!

Dipski · 04 July 2018, 12:58

Ok so I just found an AGA game to load up, and it does indeed have the glitched graphic bars. So I think that confirms that Alice is working, RAM and ROM are working and Lisa is the problem. I'll double check continuities on Lisa and if that all checks out then I think it must be the chip itself.

03 July 2018, 01:07	#1
Dipski Registered User Join Date: Jun 2016 Location: Victoria, Australia Posts: 32	CD32 Graphics Fault I've got a CD32 with a strange graphics fault. It happens with vertical columns on static sprites/graphics. On the loading screen the Amiga CD32 logo and the disc image appear normally as they spin & move, but when they stop and it becomes a static screen there are vertical columns that appear corrupted and have parts of the background there instead. I can't seem to find any issues on the address lines or video signals on the oscilloscope, so I'm leaning towards a fault with the Lisa chip as it handles sprites (I'm assuming the logos are sprites). Could any other chip or fault cause the issue and is there any way to confirm the fault before finding a replacement Lisa? Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Picasso-IV Fault	Daedalus	support.Hardware	12	04 September 2018 01:12
CD32 reset boot up fault	trolleymaster	support.Hardware	31	09 May 2018 14:58
CD32 FMV cartridge fault	AmigaFun	support.Hardware	13	14 July 2016 05:32
Amiga CD32/SX-1 Graphics Palette Issues	BanditLOAF	support.Hardware	0	06 June 2014 03:04
A1200 possible CIA fault	edimax	support.Hardware	11	20 May 2012 23:16

03 July 2018, 02:32	#3
Dipski Registered User Join Date: Jun 2016 Location: Victoria, Australia Posts: 32	Yeah I've checked the RAM, everything looks normal there. With the vertical bands there are 20 in total, every second one with glitched sprite graphics (only when static) and normal background graphics.. I can't help but feel like I'm missing something.

03 July 2018, 14:25	#4
dlfrsilver CaptainM68K-SPS France Join Date: Dec 2004 Location: Melun nearby Paris/France Age: 46 Posts: 10,505	those are jailbars, check the RAMs and ROM of the CD32.

03 July 2018, 20:08	#5
supaduper Registered User Join Date: Oct 2012 Location: Krypton Posts: 1,221	What about the excellent diagarom, you can burn an eprom for the CD32 http://eab.abime.net/showthread.php?t=89183

03 July 2018, 20:31	#6
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,553	ROM can't cause this. It can be memory (It isn't if games boot without crashing) or Alice (It most likely isn't if sprites work fine, check if NVRAM save screen "key" sprite is corrupted). It is most likely Lisa if everything else works fine.

04 July 2018, 03:22	#7
Dipski Registered User Join Date: Jun 2016 Location: Victoria, Australia Posts: 32	Thanks guys, interestingly the NVRAM save screen key is fine - and the games I have load and run fine (Alien Breed / Qwak) with no graphics glitches. This would seem to rule out a RAM issue. I did test with another ROM in the CD32 but had the same glitch on the loading screen, and the ROM socket looked fine. I'll check the ROM socket and its connections further. My main concern is if it's specific to a picture mode (for games/programs I don't have to test it with) if I can't find any fault associated with the ROM.

04 July 2018, 06:20	#8
Dipski Registered User Join Date: Jun 2016 Location: Victoria, Australia Posts: 32	The ROM socket tests fine and continuity for all of the connections from the ROM is there, the signals all look normal. I've checked continuity between Alice and Lisa, checked the signals there which all look normal as well. I've got some EPROM's on their way so I can burn a CD32 DiagROM so if everything passes on the DiagROM that should rule out any memory fault conclusively and point to a Lisa issue. I feel like there's something obvious that I haven't checked on the board but nothing else stands out going over the schematics.

04 July 2018, 08:47	#9
Toni Wilen WinUAE developer Join Date: Aug 2001 Location: Hämeenlinna/Finland Age: 49 Posts: 26,553	Try some "true" AGA game, not OCS/ECS port, game that uses higher bandwidth AGA modes because problem only seems to affect them (every other 16 pixel block is corrupted) and only bitplanes because "key" is fine (it is made from sprites). Which points to fault inside Lisa. It can't be Alice or Alice to Lisa data bus because sprites work.

04 July 2018, 12:18	#10
supaduper Registered User Join Date: Oct 2012 Location: Krypton Posts: 1,221	@ Dipski, Just to add if you are burning Diagrom to a 27c800 using the TL866 and the 42pin adapter like I have, then after flashing you will have to bend pin 31 up so its not connected for it to work on the CD32

04 July 2018, 12:58	#12
Dipski Registered User Join Date: Jun 2016 Location: Victoria, Australia Posts: 32	Ok so I just found an AGA game to load up, and it does indeed have the glitched graphic bars. So I think that confirms that Alice is working, RAM and ROM are working and Lisa is the problem. I'll double check continuities on Lisa and if that all checks out then I think it must be the chip itself.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)