What's new

Neo Geo Track Ball Games

Tested also on MV2F, Unibios in console mode
So it can work also with AES ! it's getting interesting


OUT 1 from P1 - PIN10, you need to revert it on firmware in my case and pullup the signal with a resistor.
 
Finish now to add rouble motor using airject signal and led for trackball.
Its really enjoyable, its not 100% perfect and there is a really small chance to get a really small move ghosting, maybe its come from my old trackball.
But damn its really funny to play. i test it a lot and its work fine. maybe next i will do some adjustament.
Now i will make a test pcb to use also with AES and only pad port for all MVS that use out signal from player port.
This is my setup now, ist a mess after two day of testing. :thumbsup: ( i dont show you my desk because its worse)
I will share everything after putting it on kicad, so maybe someone can make some changes and make it even better.
I hope this information is helpful to anyone.
If anyone has run any tests, please share.
I'm always curious to hear everyone's solutions.

IMG_3313.JPG
 
Last edited:
I tried wiring up my trackball circuit and I get the same value on both axis. I fired up my oscilloscope and had a look at the signal coming from CN4.1. It looks like it's an active-low pulse at a rate of 59-ish Hz and the low pulse width is around 52 µs but it spends most of its time high. The 74HC157 chips should be switching in the region of less than 25 ns, assuming I only have that 52 µs window to get the X axis read.

Then it dawned on me, I'm using my HAS 5.x which has button remap so it reads the buttons, optionally remaps them, then outputs them. As quick as that works, I suspect it's adding enough delay that the X axis never propagates through in time, (if I had been sensible, I could've measured this delay on my scope before I packed everything away for the evening!).

I think I need to create a dedicated JAMMA harness so I can wire the trackball circuit directly to the MVS and bypass the HAS.
 
Today i got my PCB from JLPCB, trackball is working fine as i except.
Since out1 signal come from P1 port i never test Aircompressor out from P2 port but only from CN4 and surprise its not come player port :unsure: so no roumble motor... maybe it's because of unibios.
I don't want to disassemble my precious console MV2 to change the bios and anyway using the original irritating maze bios just for that sounds stupid, so I'll settle for no accessories. :)

I will upload project for all in the next days.
 
I had one of the Frank_fjs JAMMA adapter PCBs fabbed and soldered it up with every pin wired straight through apart from P1 U/D/L/R/1/2/3/4 and then wired my prototype breadboard directly to those 8 pins so it now bypasses the HAS:

IMG_4100.JPG


The trackball now works correctly and both axis update independently. Now it's a question of whether to support any of the remaining features. At the very least I'll turn this into a JAMMA adapter that allows connecting a trackball.
 
The Clock, Data, & Latch signals on the MVS CN4 connector hinted that there's some sort of SIPO shift register implemented in the SIT custom. My scope confirmed there's 15/16 clock pulses so I hooked up a pair of 74HC595 shift registers chained together and this appears to be the various light/lamp outputs.

The test menu accounts for 15 lamps but the board seems to clock in a 16th value that always seems to be low. All of the lamps work as expected, high value = lamp on, apart from the flash lamp. This stays high during the charge counting phase and then briefly pulses low which I'm assuming triggers a camera flash effect and then needs to recharge.

The original SIT board has 11 outputs using some sort of TO-220 or similar package IC and then there's a TDA62083 IC which looks similar to a ULN Darlington array which seems to have 6 outputs in use judging by the populated resistors nearby. Does anyone with the original wiring harness know how they're split? My hunch would be the 10 timer lamps go to 10 of the 11 first set of outputs.

Either way, I will probably just bring the shift register outputs to a 15+ pin header for someone to do with as they wish. I've driven LEDs directly from the shift registers as a test but you wouldn't drive incandescent lamps from them without using some sort of driver.
 
  • Like
Reactions: nem
bodgit said:
The Clock, Data, & Latch signals on the MVS CN4 connector hinted that there's some sort of SIPO shift register implemented in the SIT custom.
Click to expand...
Yeah, I documented the protocol there:
Ape Cosplay

Post in thread 'Neo Geo Track Ball Games'

Sorry I never posted my findings. I went pretty far in studying the SIT board.

Wiring & protocol

Xa = V (18 ) = P2 UP
Xb = W (19) = P2 DOWN
Ya = X (20) = P2 LEFT
Yb = Y (21) = P2 RIGHT

$380001 (REG_POUTPUT)
D7 = ?
D6 = ?
D5 = P2 output
D4 = P2 output
D3 = P2 output
D2 = P1 output
D1 = P1 output
D0 = P1 output

D7 = ?
D6 = ?
D5 = Air jet
D4 = Card dispenser ?
D3 = Latch
D2 = Clock
D1 = Data
D0 = Trackball X /Y select

Read $380000

w $380001 0 Trackball X select
r $340000 Joypad port 2 inputs (REG_P2CNT)- read left/right A/B buttons (buttons A~D pass-through, left A...
Click to expand...
  • Like
  • Love
$3800001 for SIPO (Serial Input Parallel Output) + card dispenser? + air jet
Click to expand...
 
You must log in or register to reply here.
Back
Top