Final Crash (Final Fight bootleg) repair - need help

yavuzg

Member
Joined
Aug 20, 2013
Messages
259
Reaction score
6
Location
Ankara, TR
Final Crash (Final Fight bootleg) repair - need help

Hey guys,

I generally do not touch to bootleg PCBs but just trying to help a friend out here... He desperately wanted a Final Fight and ended up with this bootleg PCB due to lack of funds. Moreover the PCB does not even work :p

this is a photo of the PCB which I got from the internet. I just didn't shot an actual photo of it (I'll do when I'm home)

AA01102.JPG


He asked my help so here we are...

I've checked the couple of ROMs and the game turns out to be Final Crash.

IMG_2253_zpsy7wvgds4.jpg


I came to the point to get the sound section working (replaced some caps which were knocked off and fixed the amplifier IC's broken legs) and put a missing xtal (used an 11 Mhz one salvaged from another bootleg PCB which looks similar to this one).

The result I got is this...

IMG_2240_zpsxvqf9fnp.jpg


The game coins up and starts and the music and sound effects can be heard but the graphics are as in the photo. It looks garbage but some how consistent with the game graphics. You can tell the colors changes as the game progress on the background.

The video is like you're trying to push a 31Khz signal to a 15Khz arcade monitor. But the sync is around 15Khz according to my scope readings...

IMG_2250_zpsjq7cc22a.jpg


Could one of the PALs (most probably near GFX roms) might be faulty? If so anybody knows if the PALs have even been dumped for this hardware, somewhere?

As I said I'm really unfamiliar with these bootleg PCBs, any help will be appreciated...
 
CPS1 bootlegs are notorious to be badly engineered, most of times they 'spit out' not perfect SYNC signal like you have in your board.Anyway, I think the 11MHz quartz you used is not the right value.I can see your scope measured 14.04KHz on SYNC pin which is out of range that most of monitors can accept.Please, post a hi-res picture of your board.

P.S.
MAME driver of Final Crash reports that main CPU clock is 10MHz.Does your board carry a single oscillator/quartz to generate the master clock?If yes, check what you get on pin 15 of the 68000 CPU.
 
Hi.

Can I get a look at the parts side of the JAMMA connector?

Yesterday I "repaired" my old SF2 bootleg because the lazy bootleggers never grounded pin 14. They instead expected operators to redundantly ground it on their cabs. I guess to save money not making a trace?

If your video is not grounded it could lead to "sync" artifacts like you are getting. Try checking continuity from pin 14 to pin 1. If there's no connection then that's your issue.
 
The PCB has two quartz crystals. One of them was already there and the value of it is 27.115 Mhz (Says X-TAL2 on the silk screen)

The other one was missing. I replaced it with a 22.1184 Mhz crystal. I mis-remembered it as 11Mhz in my previous post...

Now, I know these boards were produced rather in generic purpose that not all the components indicated on the silk-screen necessarily be populated. But the location of that missing crystal seems it must have been populated because there were broken legs in the solder holes and all the other components surrounding it were populated (i.e. the capacitors, resistors and even they have used a resistor as a cable tie to keep the xtal in place)

Anyway, I found a similar, although not the same, scrap board from my junk/spares basket. On that PCB there were two crystals and one was rated as 32 Mhz the other was 22.1184 Mhz. So I snatched it from that junk board and soldered in place of the missing crystal. I really do not know what must be the *real* value of this crystal. There are some pictures of similar/same PCB on the internet but not detailed for me to read the value of that crystal...

Anyway, the actuals pictures of the PCB are coming. My phone's battery just died, letting it recharge now. :)
 
Last edited:
Hi.

Can I get a look at the parts side of the JAMMA connector?

Yesterday I "repaired" my old SF2 bootleg because the lazy bootleggers never grounded pin 14. They instead expected operators to redundantly ground it on their cabs. I guess to save money not making a trace?

If your video is not grounded it could lead to "sync" artifacts like you are getting. Try checking continuity from pin 14 to pin 1. If there's no connection then that's your issue.

Good point! This PCB was not grounded also...

BUT, my Video GND is connected to normal GND in my JAMMA rig but I double checked it anyway, it is grounded...
 
I'm pretty sure the 22.1184 Mhz oscillator you put in is not the right value for this bootleg.This causes wrong SYNC signal.Can you figure out where the clock produced by this oscillator goes?After being generated by the oscillator ( or a quartz, they are different), the periodic signal is inverted and divided and then routed to the board to.Most likey this oscillator generated the clock for the video timing and not for CPU.Check if the frequency on pin 15 of 68000 CPU is 10MHz.
 
Actual pictures of the PCB attached...
 

Attachments

  • IMG_2260.JPG
    IMG_2260.JPG
    378.8 KB · Views: 41
  • IMG_2261.JPG
    IMG_2261.JPG
    280.7 KB · Views: 28
  • IMG_2262.JPG
    IMG_2262.JPG
    293.8 KB · Views: 25
  • IMG_2266.JPG
    IMG_2266.JPG
    328.7 KB · Views: 26
I'm pretty sure the 22.1184 Mhz oscillator you put in is not the right value for this bootleg.This causes wrong SYNC signal.Can you figure out where the clock produced by this oscillator goes?After being generated by the oscillator ( or a quartz, they are different), the periodic signal is inverted and divided and then routed to the board to.Most likey this oscillator generated the clock for the video timing and not for CPU.Check if the frequency on pin 15 of 68000 CPU is 10MHz.

I'll trace back the sync signal. I'll update if it is somehow related with the new crystal.

Btw, the clock of 68K shows 10.87 Mhz...
 

Attachments

  • IMG_2267.JPG
    IMG_2267.JPG
    187.3 KB · Views: 16
OK.Judging from location of the 22.1184 MHz crystal, it should generate the clock for the 68000 so it should be right but check if you get a 10MHz signal on its pin 15 to be sure.Now you should trace back SYNC pin back from JAMMA connector and figure out where video timing signals are generated.Most likely you will find the V-SYNC and H-SYNC are mixed tpgether (usually by an AND gate of a 74LS08)
 
Last edited:
OK.Judging from location of the 22.1184 MHz crystal, it should generate the clock for the 68000 so it should be right but check it you get a 10MHz signal on its pin 15 to be sure.Now you should trace back SYNC pin back from JAMMA connector and figure out where video timing signals are generated.Most likely you will find the V-SYNC and H-SYNC are mixed tpgether (usually by an AND gate of a 74LS08)

You are right, I traced it back, passed 100Ohm resistor, an inverter and till the point they got joined, a 74LS00 (Quad NAND). It took about half an hour because the trace routed all over the PCB and jump to the upper PCB via the ribbon cables :)

I'll continue to follow...

Btw, I just changed that missing crystal with a 4.4336 Mhz one just for fun, got the exact same CYSNC (14.08Khz) but the PCB didn't boot (i.e. the game didn't even started on the background). Changed it back to 22Mhz crystal.

Also measured the clock of 68K on pin 15 as you said, it is 10.87 Mhz. This does indeed got changed with the crystal value. When there is no X-TAL1 (the missing one) the clock is around 8Mhz. When there is this 22Mhz crystal the clock for 68K is 10.87Mhz.
 
Last edited:
Small update...

I found out that the crystal near the CPU (which was missing) is indeed the source for the CPU CLOCK.

The other crystal (27Mhz) is the source for video signals. I also changed that one with a 32Mhz crystal the horizontal sync became 16.5 Khz.

Still, the video problem is there and doesn't seem to be related with the sync signal itself.
 
Small update...

I found out that the crystal near the CPU (which was missing) is indeed the source for the CPU CLOCK.

The other crystal (27Mhz) is the source for video signals. I also changed that one with a 32Mhz crystal the horizontal sync became 16.5 Khz.

Still, the video problem is there and doesn't seem to be related with the sync signal itself.

That's what I told you from the beginning :)
The 32 MHz is not the right value I think, try to lower the value until SYNC is around 15.6KHz (or adapt the divider circuit using the 32MHz one)
 
@caius, yes you told me... :)

But let me tell you something more funny...

I'm totally confused after I visited my operator friend during lunch break today. I asked if he has this kind of bootleg PCB so that I'll check the crystal values and perhaps snatch its PALs and if not locked may be dump them.

He showed me a pile of s***t in his warehouse. I found may be 5 of these boards looking "similar" i.e. two PCBs connected together with ribbon cables, without a CPLD in the middle and with two crystals at the very exact locations etc.

NONE of them has a standard!

One was even has 25Mhz in place of XTAL1 (the one close to the CPU) 10Mhz (!!!) for XTAL2...

The others has different values (10 Mhz, 27Mhz, 25Mhz etc.)...

He also said the problem I'm facing *might* be related with what he calls "sync PAL" which he was refering to the PAL close to the sound section area.

Anyway, the lunchbreak was not enough for me to find a working one. Tomorrow I'll try to find a working one from him and just to conclude this topic and time spent on THIS board I'll cross check things with it.

If all fails, I'll rip off the EPROMs, CPU and RAM's and thrash this PCB and give my friend the working FF...
 
P.S.
MAME driver of Final Crash reports that main CPU clock is 10MHz.Does your board carry a single oscillator/quartz to generate the master clock?If yes, check what you get on pin 15 of the 68000 CPU.

caius, you were right all the time :)

the crystal (27Mhz) which was already on the PCB was a complete B.S. Probably it was there from a previous repair attempt.

After checking the complete video generation section starting from the JAMMA connector (7407s, 74LS273s, RAMs etc which turned out to be OK) I noticed that an 8Mhz crystal in place of 27Mhz one generated a more stable video, which now you can guess whats going on...

IMG_2354_zpsl8l7pn0n.jpg


Then I ignored what I've seen on all the other bootlegs at the operator's warehouse, and remembered you mentioned about the MAME source code, I snatched a 10Mhz crystal from a scrap UMK3 PCB and found a 24Mhz crystal in my parts drawer...

aaand bingo!

IMG_2382_zpsdyprgv4c.jpg


thank you guys...
 
Good.So, summing up, you used a 10MHz crystal for the 68000 (the one just near the CPU) and a 24MHz for the video timing (the one on the sub-board)
 
Oh, btw, for reference, the 24Mhz crystal should go to XTAL1 location (i.e. closer to the JAMMA edge and the 68000 cpu) and the 10Mhz crystal should go to XTAL2 location (the one closer to the middle lower edge of the PCB).
 
OK so they were totally wrong values.And I can see your board is actually a Final Fight bootleg and not Final Crash, these are quite rare to see.
 
Good.So, summing up, you used a 10MHz crystal for the 68000 (the one just near the CPU) and a 24MHz for the video timing (the one on the sub-board)

No, funny design, the one closer to the CPU infact not even connected to the Clock pin. :)
 
Back
Top Bottom