• 1
  • 8
  • 9
  • 10(current)
  • 11
  • 12
CrystalHD & current SVN builds of XBMC?
archie Wrote:Did your card have the EEPROM installed?

LOL, that's a silly statement. Unless you know exactly how the hardware works and what's in the on-board EEPROM and what function is performs, it's meaning less.

P.S I have Philmatic's card and it's truly dead, could have been a coincident. It's very difficult to determine from the postmortem
Reply
davilla Wrote:LOL, that's a silly statement.

Actually, it's a question.

Quote:Unless you know exactly how the hardware works and what's in the on-board EEPROM and what function is performs, it's meaning less.

If the EEPROM is populated, the contents may have gotten corrupted. Like this: http://linux.slashdot.org/linux/08/09/23/133258.shtml

If the EEPROM is not populated, we can rule out that possibility and something else must have gone wrong.

I never claimed I knew how to fix it. I don't even own that model, which is why I asked if it has the onboard EEPROM.
Reply
archie Wrote:Actually, it's a question.



If the EEPROM is populated, the contents may have gotten corrupted. Like this: http://linux.slashdot.org/linux/08/09/23/133258.shtml

If the EEPROM is not populated, we can rule out that possibility and something else must have gone wrong.

I never claimed I knew how to fix it. I don't even own that model, which is why I asked if it has the onboard EEPROM.

Nope, a firmware download can't trash the EEPROM. The downloader code can't see that memory and you need to follow a very specific sequence to actually re-program the EEPROM.
Reply
davilla Wrote:Nope, a firmware download can't trash the EEPROM. The downloader code can't see that memory and you need to follow a very specific sequence to actually re-program the EEPROM.

IF his card really was bricked through software (big if), a corrupted EEPROM would be my first guess. On many other chipsets, the EEPROM can be used to override PCI vendor/product IDs and other important settings.

Since we know the EEPROM is not strictly required for this application, it might be interesting to lift or short the SDA/SCL pins (simulating an unpopulated EEPROM) and see if the board comes back to life.

Second guess would be overclocking/overheating the DRAM. Seems like a long shot, and the voltage is not under software control AFAICT.
Reply
archie Wrote:IF his card really was bricked through software (big if), a corrupted EEPROM would be my first guess. On many other chipsets, the EEPROM can be used to override PCI vendor/product IDs and other important settings.

Since we know the EEPROM is not strictly required for this application, it might be interesting to lift or short the SDA/SCL pins (simulating an unpopulated EEPROM) and see if the board comes back to life.

Second guess would be overclocking/overheating the DRAM. Seems like a long shot, and the voltage is not under software control AFAICT.

Actually the voltage is under software control Smile

One very big issue with old layout cards seems to be defective DRAM. I've got another old layout card that has a stick bit. Now that I'm home again, I can start poking at them again.
Reply
pkolkkal Wrote:Ok sorry guys I got tired of my own ranting and reimaged my drive with pin87a's latest image. Put the latest r144 driver with svn31718 and I finally get working video on my old layout CrystalHD card. Seeking is not as fast as with r1.0.3 & svn28256 but it works Big Grin

Also updated to latest Nvidia stable 256.44 drivers because I'm seeing faster 60Hz-24Hz mode change with them.

So it must have been something I modified on Sam's image. Probably at some point I did something stupid like apt-get update on my kernel <- FAIL

Hi pkolkkal,

pin87a's image is also running on my internal drive after some time (as a linux noob).

Could someone tell me which latest Version of XBMC and which ChrystalHD Driver is recommable? Should i also update the Nvidia Driver (faster 60Hz-24Hz sounds good..)?

Could someone tell me how to update and in which order?

thanks in advance!

MasteROe
Reply
davilla pointed out r137 should be stable for BCM970012 again. I am running r144 myself but can't see much of a performane difference between both revisions.

As for the 'faster' refresh rate switching, I'm not noticing any improvement at all. I went from 195.x to 256.x, on an AppleTV.
* MikroTik RB5009UG+S+IN :: ZyXEL GS1900-8HP v1 :: EAP615-Wall v1 :: Netgear GS108T v3 running OpenWrt 23.05
* LibreELEC 11:  HTPC Gigabyte Brix GB-BXA8-5545 with CEC adapter, Sony XR-64A84K :: Desktop AMD Ryzen 7 5800X / Sapphire Nitro+ Radeon 6700XT  / 27" Dell U2717D QHD
* Debian Bookworm x86_64: Celeron G1610, NFS/MariaDB/ZFS server
* Blog
Reply
.:B:. Wrote:davilla pointed out r137 should be stable for BCM970012 again. I am running r144 myself but can't see much of a performane difference between both revisions.

As for the 'faster' refresh rate switching, I'm not noticing any improvement at all. I went from 195.x to 256.x, on an AppleTV.

r144 brought osx/linux driver/lib in sync with windows driver/lib and fixed a major opps with bmc70015 handling as well as some other minor issues.

As for refresh rate switching, you might try extracting the EDID and using it statically instead of dynamically. The driver might be doing something silly like excessively probing the HDTV capabilities on refresh rate switch.
Reply
I assume checking Xorg's logs (or using some tool like read-edid), adding that data to Xorg.conf and playing with the UseEDID option for the nvidia driver is sufficient for that?
* MikroTik RB5009UG+S+IN :: ZyXEL GS1900-8HP v1 :: EAP615-Wall v1 :: Netgear GS108T v3 running OpenWrt 23.05
* LibreELEC 11:  HTPC Gigabyte Brix GB-BXA8-5545 with CEC adapter, Sony XR-64A84K :: Desktop AMD Ryzen 7 5800X / Sapphire Nitro+ Radeon 6700XT  / 27" Dell U2717D QHD
* Debian Bookworm x86_64: Celeron G1610, NFS/MariaDB/ZFS server
* Blog
Reply
.:B:. Wrote:I assume checking Xorg's logs (or using some tool like read-edid), adding that data to Xorg.conf and playing with the UseEDID option for the nvidia driver is sufficient for that?

Yes, there were some posts here (somewhere) that detailed that operation.
Reply
Thanks, I'll start looking around then.
* MikroTik RB5009UG+S+IN :: ZyXEL GS1900-8HP v1 :: EAP615-Wall v1 :: Netgear GS108T v3 running OpenWrt 23.05
* LibreELEC 11:  HTPC Gigabyte Brix GB-BXA8-5545 with CEC adapter, Sony XR-64A84K :: Desktop AMD Ryzen 7 5800X / Sapphire Nitro+ Radeon 6700XT  / 27" Dell U2717D QHD
* Debian Bookworm x86_64: Celeron G1610, NFS/MariaDB/ZFS server
* Blog
Reply
Is there any way to find out which layout my card has without opening the apple TV?

Sorry for this noob question (and the other questions today). I just bought an Apple TV with the 700012 preinstalled.

Could anyone describe me what disadvantage i have if i have the old layout broadcom chip? Will i never get stutterfree 1080p with this version?

Thank you!
Reply
noreason4nothin Wrote:Is there any way to find out which layout my card has without opening the apple TV?

Sorry for this noob question (and the other questions today). I just bought an Apple TV with the 700012 preinstalled.

Could anyone describe me what disadvantage i have if i have the old layout broadcom chip? Will i never get stutterfree 1080p with this version?

Thank you!

Nope, you might get lucky if the PCI subsystem was set to 2616, some old layout card were programmed that way.

Old layout vs new layout has nothing to do with stutterfree 1080p. 1080p seems to be working fine for me with AppleTVs running XBMC under Linux.
Reply
Okay cool. Will this old Chip support all of your updates until sometime 1080p will rune fine on atvos?

Thx for everything!
Reply
noreason4nothin Wrote:Okay cool. Will this old Chip support all of your updates until sometime 1080p will rune fine on atvos?

Thx for everything!

Yes, I've already provided patches to crystalhd upstream to continue driver/lib support under old layout cards and I have an old layout card now for testing. However, if an old layout card has defective DRAM, there's nothing I can do about that.
Reply
  • 1
  • 8
  • 9
  • 10(current)
  • 11
  • 12

Logout Mark Read Team Forum Stats Members Help
CrystalHD & current SVN builds of XBMC?0