Kid Chaos glitch

[kremiso]

Originally Posted by amilo3438

Hmm, if it's not original, maybe you could make an .adf image of that version that works on a real OCS machine so that it can be checked on the emulator as well !?

EDIT: Also, does the disks from the Tosec also glitch on your real Amiga !?

next time ill do the travel ill check better the game version, but almost sure is not a version matter

(Prestige crack + trainer iirc)

[amilo3438]

Originally Posted by ross

If, as suggested by Toni, it is a signal timing problem there is very little that can be done: the game may or may not glitch on your OCS system.

Unless in fact there is a patched version of the game, but apparently this is not the case.

If it is a signal timing problem or not that should be checked/tested on his real OCS machine with the Tosec verions too! (does it happens the same i.e. glitches)

Is his tested version patched how we could know without checking the same! (sound like guessing)

[ross]

Originally Posted by amilo3438

If it is a signal timing problem or not that should be checked/tested on his real OCS machine with the Tosec verions too! (does it happens the same i.e. glitches)

Oh well, the perfect thing would be to have his machine to do the tests.
Just send it to Toni, then we'll do all the necessary checks

[amilo3438]

Originally Posted by ross

Oh well, the perfect thing would be to have his machine to do the tests.
Just send it to Toni, then we'll do all the necessary checks

Exactly!

[Toni Wilen]

It is timing problem inside Agnus, either Denise gets the sprite data (sprite DMA slot was used) or it does not (sprite DMA slot was not used). But that can be tested without any external hardware.

[ross]

Originally Posted by Toni Wilen

..4-bit DRAMs need 9-bit refresh but 8371 only has 8-bit refresh. ECS Agnus is needed to support 4-bit DRAMs.

Wait! This is not the right place because we are totally off topic and have nothing to do with the problem of the game, but I just checked the specs of the 8371 PLCC and it has 9 bits for RAS refresh! (MA0~8, RAS0=RAS1=low)
I think we need to check a few things, I could be completely wrong

[Toni Wilen]

Originally Posted by ross

Wait! This is not the right place because we are totally off topic and have nothing to do with the problem of the game, but I just checked the specs of the 8371 PLCC and it has 9 bits for RAS refresh! (MA0~8, RAS0=RAS1=low)
I think we need to check a few things, I could be completely wrong

Yes, I noticed the same few minutes ago. I mixed two things: 8371 does not work reliably with rev6 mainboard with 4 chip ram chips but trapdoor expansion can work normally.

[ross]

Originally Posted by Toni Wilen

Yes, I noticed the same few minutes ago. I mixed two things: 8371 does not work reliably with rev6 mainboard with 4 chip ram chips but trapdoor expansion can work normally.

Ok, so the 4 256x4 chips are not 'strange'.

At this point you should check the batch dates of the 8371s and understand if they behave differently.
If by chance they all behave the same on different motherboards, then you could already understand which ones have the problem with back-to-back spr/bpl.
Otherwise it is a problem with the timing of some signal on the motherboard..

[ross]

Attached a super-simple test.

Start the ADF and check result:
- full red screen: real ECS+ Agnus
- black screen with green vertical stripe: OCS Agnus without bug
- black screen with white/striped vertical stripe: OCS Agnus with bug

If you want to try it in WinUAE remember that the OCS Agnus without bugs is not selectable, so the only thing you can do to emulate the situation is to start the system in OCS and then change the chipset in ECS while the test is active.

To report, just in case: everything written on Agnus chip and the motherboard revision.

[kremiso]

thank you ross, ill take with me the next time to complete the test, also with the other machine

[kremiso]

one thing about the glitch itself

now the game contain some levels that work as arcade sequence, with less icons on the left, just score and time counter;

now those sequences works without glitch also on OCS

so, do you think that the glitch is generated from the lower 3 icons on the left (flowers, health and another i dont renember atm), or to the fact that those sequences doesnt have left/right scrolling?

you can see how that sequences behave typing 'ARCADEGAMES' as password

[ross]

Originally Posted by kremiso

.. or to the fact that those sequences doesnt have left/right scrolling?

This.
DDFSTRT is high enough to not "collide" with the sprites DMA.

[Superman]

I just tested the prestige crack on my A500. I have the ECS Agnus and OCS Denise and no glitches for me. The same config on WinUAE is also glitch free so the emulation for that part is correct.

[ross]

Originally Posted by Superman

I just tested the prestige crack on my A500. I have the ECS Agnus and OCS Denise and no glitches for me. The same config on WinUAE is also glitch free so the emulation for that part is correct.

Yep, no ECS Agnus Amiga suffers from this bug.

The problem is instead to understand which OCS Agnus do not suffer from it and above all why.
WinUAE is conservative and acts as if all OCS Agnus are affected, also because there was no evidence to the contrary.

In any case the bug seems 'stable', i.e. if the conditions are met it is repeatable.