MiSTer FPGA Forum

The Grobda arcade core isn't working properly - primary button is "shoot" and works, but there is supposed to be a secondary "shield" button which does not activate. The core allows you to map a shield button on control remapping, but it just doesn't respond in-game. It's not a dip switch setting, and the shield appears in the game's attract mode fine, so it's not a strange "shield-less" rom.

Verified not working on multiple MiSTers.

The Grobda rom uses the "The Tower of Druaga Inc" core - (Which includes Dig Dug II, Mappy, Motos, Super Pacman, Pac & Pal, Pac-Man & Chomp Chomp and Grobda)

Can anyone else verify this?

Verified. I reported it on the GitHub page. Maybe you can add a comment on this there:

https://github.com/MiSTer-devel/Arcade- ... /issues/25

Thanks, PistolsAtDawn.

I played this a lot back in the day at the arcades - FYI for those who haven't - it's not a "pickup" or powerup - you get the shields immediately. (on another forum post, some people thought it may be something you get "later" in the game , so possibly not a bug - this isn't the case!).

You have to hold the shield button down to keep it activated - a lot of people in the arcades didn't realise that and thought it didn't do anything much - this was the early 80's where people were only used to mashing buttons - yes I'm that old

Hopefully there is a fix, as it's a really fun game, but unplayable in it's current state. So much so that I'd consider it being removed from update_all until it's playable.

I would say it's very, very challenging without the shields, but not quite unplayable. This game was one of the MiSTer Community Game Club picks here: viewtopic.php?t=5149

I didn't get a reply through Github. If I were the author, I think it's likely that I would ignore my bug report as it seems like user error. If you were to also report it, then maybe it would get fixed.

I updated the SYS files and had a look at this, but there is more happening than just the shield not working.

The self test in GROBDA reports 0 0 3 (rather than the expected 0 0 0) so something beyond the control mapping needs sorting out, need to find out what the '3' is indicating.

it does look (from Mame driver) as though GROBDA does some bank switching on the inputs, and that the second fire button for each player is multiplexed into one of the dip switches. so if this is what it is complaining about then it needs adding into the code.

will let you know how I get on!

macro wrote: ↑Wed Dec 07, 2022 9:24 pm

I updated the SYS files and had a look at this, but there is more happening than just the shield not working.

The self test in GROBDA reports 0 0 3 (rather than the expected 0 0 0) so something beyond the control mapping needs sorting out, need to find out what the '3' is indicating.

it does look (from Mame driver) as though GROBDA does some bank switching on the inputs, and that the second fire button for each player is multiplexed into one of the dip switches. so if this is what it is complaining about then it needs adding into the code.

will let you know how I get on!

If I can be of assistance with testing, please let me know!

I think I have worked out where the problem is.

It seems that each game on the platform uses different IO controller chips and Grobda is the only one that expects button 2 in that particular location.

It also seems that the IO chip is used in a bootup check, so the information being returned then will also be wrong (which may explain the '3' on self test screen)

I will construct a suitable 'ioctrl_2' to be used for Grobda that fixes these two problems and hope that it then works as expected.

macro wrote: ↑Thu Dec 08, 2022 11:33 pm

I think I have worked out where the problem is.

It seems that each game on the platform uses different IO controller chips and Grobda is the only one that expects button 2 in that particular location.

It also seems that the IO chip is used in a bootup check, so the information being returned then will also be wrong (which may explain the '3' on self test screen)

I will construct a suitable 'ioctrl_2' to be used for Grobda that fixes these two problems and hope that it then works as expected.

That's great you found the problem, Grobda is a fun game that doesn't get a lot of attention.

There is one other thing missing from Grobda, and that's the sound sample "Get Ready" that plays before the start of every stage. At least I think it is missing because I don't hear it on my MiSTer build. If I remember it is not really a digital sample, it's actually played directly using the Namco 8channel sound hardware by manipulating the sound registers. At least that's what I remember from the MAME driver notes.

goldenage wrote: ↑Sat Dec 10, 2022 6:08 pm

That's great you found the problem, Grobda is a fun game that doesn't get a lot of attention.

There is one other thing missing from Grobda, and that's the sound sample "Get Ready" that plays before the start of every stage. At least I think it is missing because I don't hear it on my MiSTer build. If I remember it is not really a digital sample, it's actually played directly using the Namco 8channel sound hardware by manipulating the sound registers. At least that's what I remember from the MAME driver notes.

I'm using a modified version of Mame at the moment to dump IO chip info to a log file. Initial tests seem to indicate that it uses the protection from both the 56xx and the 58xx chips, I updated one and the error changed to 02, so hopefully fixing both will give a 00.

There is also some oddity with coining up, but that may be down to the missing DIP settings

Just trying to decide whether to re-structure the code for these chips so it is in one file and is a bit easier to follow, but will continue with getting grobda working properly (with dips!) first.

sound is a sample, played by the CPU through a 4 bit DAC so just need to add an output port for the value and mix it into the sound channels - will add that as well.

I have the shield button working and it now passes the self test.

Just need to tidy up the code and add the DIPS and DAC port.

I've not done the DIPs yet, but self test, controls and speech should now be fine and I have tried some of the other games and they still seem to work.

so whilst I sort out DIP settings for Grobda, can you please do some testing and make sure that all of the other MRAs using this core still work as expected.

Thank you @marco !
I have tested all the games, they start Ok.
I will do more test this evening.

I have differences for sound (louder sound and maybe a slightly faster) between this RBF and the last release for the game "The Tower of Druaga" (I just checked this game for sound)

There is DAC added for Grobda to be able to use speech. Could this change also influence other games ?

wark91 wrote: ↑Wed Dec 14, 2022 10:25 am

I have differences for sound (louder sound and maybe a slightly faster) between this RBF and the last release for the game "The Tower of Druaga" (I just checked this game for sound)

There is DAC added for Grobda to be able to use speech. Could this change also influence other games ?

The overall clock is the same at 49.147727Mhz so speed should be the same. (just checked it to make sure since I had to move the file to another subdirectory)

The DAC is only active for Grobda (it uses the MODEL ID from the MRA file) so grobda will be lower overall volume, since it has to mix two sources, but other games should remain the same as they were. I'll check what 'silence' on the DAC port is in the game, in case it leaves it set to FF or something (rather than 00)

I thought the screen seemed a bit wide on a horizontal monitor, but it uses the same aspect settings as it did before.

n.b. sound may also be affected by the update to latest SYS, that has had a lot of changes since the one in there originally. should not change speed though

Thank your responses.
Yeah sys update could affect it also.
For the "slightly faster", it is just an impression. The sound volume on my side is different so difficult to compare it clearly.

Regarding screen, Mame shows those informations for Grodba : resolution = 288*224 for 60.60hz)

Edit: Also the new volume sound maybe more accurate for an Arcade.

I tested some of the arcades based on this core. There is one noticable regression: playing on a regular (yoko) monitor with the "Vertical" orientation setting, the aspect ratio is ignored. Both "original" and "wide" now fill the screen completely.

Fixed the aspect, added DIPs to mra (except coinage, since code for IO chip does not take any notice anyway)

Added some code to zero the DAC when it plays other sound effects (Grobda leaves it set to $07)

Please test this whilst I sort out what needs to be checked in as a pull request.

Thanks for the update. The aspect ratio is fine now.

Thank you !

For aspect ratio (custom aspect ratio), it misses something
the definition uses the older format ("HFO1,Aspect Ratio,Original,Wide;",)
https://github.com/MiSTer-devel/Arcade- ... ga.sv#L193

The new one on cosmic core ("H0OJK,Aspect ratio,Original,Full Screen,[ARC1],[ARC2];",)
https://github.com/MiSTer-devel/Arcade- ... ic.sv#L209

I will check later for sound
THanks again for your time on this core

I tried for sound It is louder with the RBF in comparison of previous release for The Tower of Druaga.
But I don't think it needs to be changed because the Arcade have loud music generally.
Maybe someone in the future could test it against real hardware to see if it is similar or not.

Thanks again !

Final change for the moment, the DAC implementation has been moved into the sound chip - as per Gyurco's changes for Mist. (mame still implements it as an independant DAC, but his version seems more logical and does not need the silence hack)

I may look at the IO side of things again later, with a view to adding 56XX and 58XX modules (and any other needed) to make it more like the real hardware and remove most of the "if game == GROBDA" type conditions.

Although with different combinations and input mappings for each PCB there may still need to be variations, but they can be moved to the top level for ease of access.

Hey Macro,

FYI - I just tried your more recent zip update attached but not still working for me unfortunately.

I'm slightly n00b to the scene so Just verifying:

added
Arcade-Druaga.rbf to /media/fat/_arcade/cores

overwrite
Added grobda(W,New Ver.).mra on /media/fat/_arcade

ihaveamo wrote: ↑Tue Dec 20, 2022 10:07 am

added
Arcade-Druaga.rbf to /media/fat/_arcade/cores

You need to rename to Druaga.rbf and remove the old Druaga-<DATE>.rbf.

Also notice that update_all will add the old rbf file again.

You could also create a directory names "Cores_Test" and in this directory create a directory named "cores".
You put MRA on "Cores_Test" and rbf in cores directory in this case update_all will not modify it.

Thank you Marco !! The fix is merged