What's new

Does Anyone Own a Gundhara PCB?

If anyone who owns one of these and is willing to sell it, I'd be happy to buy it and have it sent direct to TS so he can get his work done on this.
I appreciate the offer but there is no need. plasticfactory was kind enough to send me his and I was able to fully document everything I needed.
 
Sorry I haven't updated sooner I'm still trying to organize all of my data here and make this hardware do other interesting things. I'll say I've got a good 95% understanding of how the later Seta Gen 1 hardware (the ones with 4x QFP Graphics Customs) works at this point... I'm still trying to crack that last 5%

But it's only fair that plasticfactory shared his board with me to research so I'll share all of the information I learned from the board with all of you.

Here is the updated RAM mapping differences:
gundhara_zr_ram.jpg


Here is the difference in the way that Sprite Addressing works between Zombie Raid, Gundhara and Blandia.
On a whole Sprite data is 32-bit and most games support up to 16Mbit of data (that means 2x 27c800 ROMs up to address A18)
Blandia and MS Gundam use 4x 27c800 ROMs (meaning 32Mbit of data, using up to address A19)
Gundhara uses 4x 27c160 ROMS (meaning 64Mbit of data, using up to address A20)

Here is how they're all mapped
sprite_addr_blandia.jpgsprite_addr_gh.jpgsprite_addr_zr.jpg

Zombie Raid is further unique in that it uses so little sprite data they added extra hardware to squeeze it all into a single ROM chip, but since a single ROM is only 16-bit and it needs 32-bit data the extra hardware bank swaps the chip and loads half of the data on each read into a pair of flip-flops as a temporary memory store... it's honestly kind of annoying and I'm sure they did this because the Mask ROM cost was more expensive than the flipflops,

But it's still workable. The two bank switching flip flops on zombie raid need to be removed, then the two addressing flip flops used by Gundhara need to be added, and since the address lines on the u200 and u201 sprite sockets on Zombie Raid were all messed up to support the weird sprite logic I used the u63 and u64 sockets with some 27c322s, rebuild the A19 and added the A20 address lines alone with socketing U62 addressing logic chip so that I could pull pins that Zombie Raid intentionally grounded.

It's ugly but the result is this:
View: https://www.youtube.com/watch?v=g5EjqhZWCIg

^Audio is fully working too I just don't have JAMMA audio on my PVM bench setup at the moment
 
Last edited:
While I'm at it does anyone have an original of any of the following:

Daioh
Kamen Masked Rider Club Race
Mad Shark
Magical Speed
Mobile Suit Gundam
Rezon

There are a few assumptions I'd love to confirm on these boards to determine the differences from other similar PCBs. Looking at a multi and want to include as many games as I can.
 
Last edited:
While I'm at it does anyone have an original of any of the following:

Daioh
Kamen Masked Rider Club Race
Mad Shark
Magical Speed
Mobile Suit Gundam
Rezon

There are a few assumptions I'd love to confirm on these boards to determine the differences from other similar PCBs.
original mobile suit gundam? i do
 
original mobile suit gundam? i do
Awesome. This is kind of an oddball PCB on this hardware, it'll probably be the last game I tackle.

I have Daioh, but it’s that weird prototype PCB that is much bigger than the final.
Interesting.
From the conversion perspective the Daioh Prototype is like a completely different game, they changed the memory map around quite a bit.
does yours have the hand-made "kick" connector sub-board?

Downtown?
Thanks for the offer, I do own this one myself.

Interestingly MAME recently split the Seta Gen 1 driver and separated out a bunch of the early 8MHz games
So even though system 16 lumps them together. MAME categorizes the "Downtown Hardware" as a separate platform that includes
Twin Eagle
Downtown
Caliber .50
Arbalester
Meta Fox
US Classic Golf
and Thundercade/Twin Formation
 
@imbord3rlin3 @plasticfactory
For:
- Mad Shark,
- Kamen Masked Riders Club,
- Magical Speed
- MS Gundam
I need the pin routing for these two chips confirmed:
sokonuke_irq_addr.jpg


and here's a map of where those chips are on the PCB
sokonuke_taisen_game_chp_map.jpg


The image above was the routing I worked out from my Sokonuke Taisen Game PCB and looking at pictures I'd assume that the above game are mostly (hopefully exactly) the same.
My assumption is that all of the above games are the same except for MS Gundam as it has more PCB modifications than the others and a slightly different IRQ map according to the MAME driver, though my hope is that MAME is wrong and the hardware isn't actually different.

Basically what I'm looking for is confirmation of the IRQ addressing on these PCBs. The IRQs are one of the trickiest parts of conversions on this hardware and one of the things that MAME doesn't have entirely right, but if you get it wrong on hardware the game wont boot. I'm hoping it's the same for all the above mentioned game.
 
Damn, you're fast! I did a search and grabbed the wrong link, try now.
lol I was already in the thread typing a reply.

the picture from the first link, that gray colored sub-board near the JAMMA edge is the hand-made kick connector sub-board I was talking about. USA version of Daioh used 6-buttons and since the JAMMA edge only supported 3 it had a kick connector for the other 3 buttons.

I hope I didn't buy a conversion at the price I paid, but fuck it if I did, that game is pretty good!
So that's a legit Seta Gen 1 prototype PCB, but I've seen prototypes of several other games with the exact same PCB. so there's a good chance that Seta re-used these boards when developing other games. Given the inclusion of the kick connector sub-board I'd say it's probably as legit as it gets from a prototype perspective.

I've seen pictures of that board before (maybe even your exact board) and I understand how the circuit works enough that it could be re-created on another PCB.
Maybe you can confirm some of my assumptions for some added confidence
1. is that the sub board sits on top of U42 sharing it's pins
2. is that there is a single fly wire from that sub-board going to U41 Pin 7
3. that the NEC UPD71054 chip (U121) has pin 11 disconnected from the PCB and tied to 5V through the jumper wire on Pin 24

One other thing that might be interesting is to know the jumper configuration for
JP151, JP152, JP153, and JP52/JP55
They should all be straight-forward A-B or B-C jumps but JP52/JP55 is weird
JP55 should all be tied together and then there should be a single jump from one of the JP52 pins to the JP55 pins, so is it A, B, or C on JP52 that is connected to JP55
 
@imbord3rlin3 @plasticfactory
For:
- Mad Shark,
- Kamen Masked Riders Club,
- Magical Speed
- MS Gundam
I need the pin routing for these two chips confirmed:
sokonuke_irq_addr.jpg


and here's a map of where those chips are on the PCB
sokonuke_taisen_game_chp_map.jpg


The image above was the routing I worked out from my Sokonuke Taisen Game PCB and looking at pictures I'd assume that the above game are mostly (hopefully exactly) the same.
My assumption is that all of the above games are the same except for MS Gundam as it has more PCB modifications than the others and a slightly different IRQ map according to the MAME driver, though my hope is that MAME is wrong and the hardware isn't actually different.

Basically what I'm looking for is confirmation of the IRQ addressing on these PCBs. The IRQs are one of the trickiest parts of conversions on this hardware and one of the things that MAME doesn't have entirely right, but if you get it wrong on hardware the game wont boot. I'm hoping it's the same for all the above mentioned game.
not gonna lie my dude, thats a little over my head. how bout i just mail u the board. im not missing it.
 
While I'm at it does anyone have an original of any of the following:

Daioh
Kamen Masked Rider Club Race
Mad Shark
Magical Speed
Mobile Suit Gundam
Rezon

There are a few assumptions I'd love to confirm on these boards to determine the differences from other similar PCBs.
I have a Rezon pcb if you need any info from it. . . .
 
I have a Rezon pcb if you need any info from it. . . .
Yes! there are a few pieces of info I'd like though because the PCB is so different than others I don't have any specific chip numbers or pins I can give you.

So if you're up for a multi-meter scavenger hunt, here's what I need!

There are 2 addresses that I need to figure out how they're expressed in hardware:
Code:
map(0x500004, 0x500005).w(FUNC(seta_state::ipl1_ack_w));
map(0x500006, 0x500007).nopr();

First of all I don't trust that MAME has correctly identified 0x500004 as an IPL1 Acknowledge, some other games they've identified IRQs like this incorrectly.
secondly they've the address at 0x500006. I suspect that this is the watchdog, or it could very well be a different IRQ.

So ultimately what I want to know is, for those two addresses which IPL is it, or are either of them the Watch dog?

Here's my attempt at explaining how the circuit works:
I'll attempt to explain, and this will probably be confusing, but after this explanation I'll try to put some simplified steps to unraveling all of this. :D
Here is what I expect:
-One of the PALs generates an enable signal for the 0x50#### address range. Then that will go to a Demultiplexer somewhere (usually a 74LS139 on these Seta Board) which will direct to the more specific (0x500004 or 0x500006 ranges)
-Assuming it's a 74LS139 that Demultiplexer there's probably only 1 or 2 of these on the whole board.
----The PAL output (which will be on the pin 11 through pin 20 side of the chip) will go to either Pin 1 or pin 15, on the demultiplexer
----On that same demultiplexer either Pin 2&3 or Pin 13&14 will be attached to A1 & A2 (Pin 29 and Pin 30) of the 68K CPU
----The remaining pins are the outputs, one will go to the the coin counter circuit, one will go to the video registers circuit, and two others will go to our mystery addresses

Now the Watchdog circuit for most of these seta boards is reset using a 74LS161 chip. There should probably only be 1 of these on the whole PCB and it's usually somewhere between the CPU and the front edge of the PCB the reset for this is pin 1. BUT we don't know for sure that either of our two mystery addresses are a watch-dog. so this could be a red herring

IPLs are on the 68K (Pin 23 for IPL2, Pin 24 for IPL1 and Pin 25 for IPL0). there's usually a flip-flop chip between the IPL pin of the 68K and the place where it's triggered. Depending on what else connects to those fip flop will determine what that IPL is for (such as Vblank, or sound interrupt, etc.) There are a few different flip-flop chips that could be used here. I've seen 74LS74 as well 74LS107. and there are others that could potentially be used.

How I would approach this:
1. I would start identifying a 74LS161 chip (I think U11 based on Rezon pictures I'm seeing)
2. Figure out where any demultiplexers are (I think maybe U15 or U41 based on pictures, there could be others)
3. Figure out where pin 1 of the 74LS161 chip goes (good chance it's one of the pins of those demultiplexers, which chip and which pin?)
4. Confirm that the addressing pins of the demultiplexer go to where we think they do (PAL output on Pin 1 or pin 15, A1 and A2 on 68K on the next two pins down from the PAL pin).
5. From here we'd probably work out next steps based on information we gathered in the above steps. but one of the other pins of the demultiplexer will go to a flip flip (maybe U42?) and maybe a different pin on that flip flop goes to an IPL etc.

I typically create a spreadsheet and map out the chips to easily keep track of which pins I've identified and where they go.

for instance here is the demultiplexer and flip-flop pin map for Zing Zing Zip:

zing_zip_addressing.jpg


and here's the one I put together for sokonuke:
sokonuke_addressing.jpg


I suspect that Rezon will be similar to both of these with some subtle differences.

Zing Zip above might be a bit challenging to follow because there are jumpers in the mix so that the IPL address can change based on the jumper configuration.
 
  • Like
Reactions: nem
lol I was already in the thread typing a reply.

the picture from the first link, that gray colored sub-board near the JAMMA edge is the hand-made kick connector sub-board I was talking about. USA version of Daioh used 6-buttons and since the JAMMA edge only supported 3 it had a kick connector for the other 3 buttons.


So that's a legit Seta Gen 1 prototype PCB, but I've seen prototypes of several other games with the exact same PCB. so there's a good chance that Seta re-used these boards when developing other games. Given the inclusion of the kick connector sub-board I'd say it's probably as legit as it gets from a prototype perspective.

I've seen pictures of that board before (maybe even your exact board) and I understand how the circuit works enough that it could be re-created on another PCB.
Maybe you can confirm some of my assumptions for some added confidence
1. is that the sub board sits on top of U42 sharing it's pins
2. is that there is a single fly wire from that sub-board going to U41 Pin 7
3. that the NEC UPD71054 chip (U121) has pin 11 disconnected from the PCB and tied to 5V through the jumper wire on Pin 24

One other thing that might be interesting is to know the jumper configuration for
JP151, JP152, JP153, and JP52/JP55
They should all be straight-forward A-B or B-C jumps but JP52/JP55 is weird
JP55 should all be tied together and then there should be a single jump from one of the JP52 pins to the JP55 pins, so is it A, B, or C on JP52 that is connected to JP55
I'm sorry, I didn't see this reply and the questions. I'll dig out the board to answer these questions.
 
Back
Top