As some of you probably already know, there is a graphics problem on certain hardware in Taito's F3 system.
Let's put the situation in context:
Some of the first pcbs of the F3 system, specifically those that came out between 1992 and 1994, were designed as single-pcb and there were two different revisions of them. As @undamned called on his blog:
http://therealundamned.blogspot.com/2009/08/taito-f3-pcb-101.html
However, before transitioning to the motherboard and cartridge model, some of the latest games were released using the first revision (orange components) of the system, reused or converted to run some of the games originally designed to run on pcbs of the second revision (white components). You could say they were factory-conversions. The games that I've been able to identify in which this was done are:
To do this, Taito used an “Expand PCB” mounted onto original revision 1 board using two connectors on each edge of the board. This was done because we suppose that the original mask-roms of the games (arabian magic, Riding Fight ...) were not extracted from the pcb for reasons of saving time or money, and instead they used the "Expand PCB" to place the roms of the new "converted" games, as they didn't fit on the original board.
In addition there are several factory-patches made on the solder-side of all of them to cover some additional addressing that the old board (orange components) did not originally cover, but that the next version of the board (white components) does:
So far so good, but it's well known that specifically these games using "expand pcb" have a problem of "color bleeding", in which it seems that there is a blur effect that goes from left to right, (and from bottom to top in gunlock, vertical game), a problem that doesn't occur in other games like grid seeker, arabian magic ... using exactly the same base hardware (weird?!). (More info here: https://shmups.system11.org/viewtopic.php?f=6&t=50619):
This made me suspect that there must be some differences between the components used in pcb of each game, apart from the "Expand PCB", and that difference must be located in the analog video output stage, since the effect is visible in every single layer of graphics in the game: both sprites, text and backgrounds.
Comparing the pcb with the Grid Seeker one, I found several amplifier transistors had been replaced by jumpers, and some resistors had been removed (or not installed at all), as well as other transistors replaced by a different model.
Super Cup Finals:
Grid Seeker:
Components originally installed in Grid Seeker:
After waiting over a month (ughh) for them to arrive to my door, I finally had them available to try.
Obviously, instead of testing that in Gunlock's pcb, a Super Taito Cup Finals with “Expand PCB” was the first pcb chosen to test whether using the correct components that were originally mounted on the first F3 boards would make the annoying bleeding effect disappear...:
New components installed:
Once the "original" components had been installed (like the ones we saw in the Grid Seeker), let's see if the result was successful:
YAAY!
Seeing that was successful on Super Cup Finals pcb, I decided to do it also in Gunlock pcb and results are obviously the same and lot better than the previous one:
Sharp and colorful image as any other F3 game!
I do not know if this change of components was made for compatibility with European or USA chassis or if there was simply no stock of those components and a cheaper workaround was carried out.
If you have any of the pcbs involved, don't hesitate to fix it!
PD: I also made a YouTube video explaining this (only in Spanish, sorry
)
Let's put the situation in context:
Some of the first pcbs of the F3 system, specifically those that came out between 1992 and 1994, were designed as single-pcb and there were two different revisions of them. As @undamned called on his blog:
http://therealundamned.blogspot.com/2009/08/taito-f3-pcb-101.html
- First revision using orange components (early games): Ring Rage, Arabian Magic, Riding Fight, Grid Seeker, (Taito Cup Finals?)
- Second revision using white components (93-94 games): Taito Cup Finals, Top Ranking Stars, Prime Time Fighter, RayForce, Super Cup Finals (also have additional rom space)
However, before transitioning to the motherboard and cartridge model, some of the latest games were released using the first revision (orange components) of the system, reused or converted to run some of the games originally designed to run on pcbs of the second revision (white components). You could say they were factory-conversions. The games that I've been able to identify in which this was done are:
- Gunlock
- Super Taito Cup Finals
- Light Bringer / Dungeon Magic
To do this, Taito used an “Expand PCB” mounted onto original revision 1 board using two connectors on each edge of the board. This was done because we suppose that the original mask-roms of the games (arabian magic, Riding Fight ...) were not extracted from the pcb for reasons of saving time or money, and instead they used the "Expand PCB" to place the roms of the new "converted" games, as they didn't fit on the original board.
In addition there are several factory-patches made on the solder-side of all of them to cover some additional addressing that the old board (orange components) did not originally cover, but that the next version of the board (white components) does:
So far so good, but it's well known that specifically these games using "expand pcb" have a problem of "color bleeding", in which it seems that there is a blur effect that goes from left to right, (and from bottom to top in gunlock, vertical game), a problem that doesn't occur in other games like grid seeker, arabian magic ... using exactly the same base hardware (weird?!). (More info here: https://shmups.system11.org/viewtopic.php?f=6&t=50619):
This made me suspect that there must be some differences between the components used in pcb of each game, apart from the "Expand PCB", and that difference must be located in the analog video output stage, since the effect is visible in every single layer of graphics in the game: both sprites, text and backgrounds.
Comparing the pcb with the Grid Seeker one, I found several amplifier transistors had been replaced by jumpers, and some resistors had been removed (or not installed at all), as well as other transistors replaced by a different model.
Super Cup Finals:
Grid Seeker:
Components originally installed in Grid Seeker:
- 3 x A1015 (Toshiba transistor) placed at TR1, TR2 and TR3
- 3 x C1815 (Toshiba transistor) placed at TR4, TR5 and TR6
- 3 x 2K Ohm resistors placed at R32, R33 and R34
After waiting over a month (ughh) for them to arrive to my door, I finally had them available to try.
Obviously, instead of testing that in Gunlock's pcb, a Super Taito Cup Finals with “Expand PCB” was the first pcb chosen to test whether using the correct components that were originally mounted on the first F3 boards would make the annoying bleeding effect disappear...:
New components installed:
Once the "original" components had been installed (like the ones we saw in the Grid Seeker), let's see if the result was successful:
YAAY!
Seeing that was successful on Super Cup Finals pcb, I decided to do it also in Gunlock pcb and results are obviously the same and lot better than the previous one:
Sharp and colorful image as any other F3 game!
I do not know if this change of components was made for compatibility with European or USA chassis or if there was simply no stock of those components and a cheaper workaround was carried out.
If you have any of the pcbs involved, don't hesitate to fix it!
PD: I also made a YouTube video explaining this (only in Spanish, sorry
)Attachments
Last edited:
