Got this Konami Block Hole (export version of Quarth) from my friend ‘mastercello’ :
Board was marked with “sprite issue” and indeed it’s what I got once fired it up:
Some sprites (like main ship, see the above picture) were garbled and totally wrong.There are eight 27256 EPROMs containing all the sprites so I went to probe them with a logic probe and I found that the common address line A10 (PIN21) of four of them was not prperly toggling compared to the healthy lines A9 and A8 for example.This was confirmed also by my logic analyzer:
I was able to trace this faulty address line back to the sprites generator ASIC marked ‘051960’:
So my suspects was that this ASIC would address the four sprite ROMS wrongly due its weakened address line.To confirm this I lifted the involved pin:
and its behaviour was the same as before when connected to sprite ROMs (on the left the an healthy address line, on the right the faulty A10) :
At this point, I was pretty sure the ASIC ‘051960’ was faulty in this address line so I played the card of replacing it:
And I was right, this restored graphics completely.Board 100% fixed.
Another board from my friend ‘mastercello’ and another quick fix.
We have on the bench a Raiden DX (Seibu Kaihatsu) PCB in good shape:
Board booted but colors were wrong and sprites colored blocks:
This Seibu hardware doesn’t use many components since most of logic has been condensed in some custom ASICs which have mostly graphics functions so I focused on them.Like I usually did I probed the tightness of each pin with a small needle and found some lifted pins of the one stamped ‘SEI252 SB05-106’ (QFP208 package).
A simple reflow did the trick and fixed graphics completely.
This repair log is proof that we must always start from the simpliest things!
Another faulty board from my friend ‘mastercello’, this time an original Toaplan Out Zone :
He told me that it worked once and then nothing.Actually when I first powered it up I got this :
The screen with the wavy lines is common on all the Toaplan PCBs which run on similar hardware.It means that the system is being intialized but it should last until the CPU starts to properly execute code.In my case this screen was permanent so there had to be a problem in the main code execution.First thing I did was to check CLOCK, /RESET and /HALT lines of 68000 main CPU, they were fine.Then I dumped the two 27C010 programs ROMs and got from my programmer a warning about a poor contact on pin 17 (which is a DATA output) of the one labeled ‘TP018_08’:
A closer inspection of the device revealed many oxidixed pins.A bit of fine sandpapering is what the EPROM device needed and enough to fix the board completely.
My friend ‘mastercello’ sent me this Donpachi PCB for a repair:
The board was in good shape but had a GFX issue where background layer was partially missing with lines across it:
Studying the hardware and with help of MAME I could identify the devices containing the data of the two background layers, they are two 42PIN 8 Mbit MASK ROM @U54 and U57.Probing them with a logic probe I found that the one @U54 has all its data lines stuck LOW or HIGH, also the two control lines /CE and /OE were stuck LOW though they were properly connected to a near ASIC which generates these signals.Obviously this was totally abnormal.As confirmation of my theory I was able to reproduce ths issue in MAME using a dummy ROM file of the background layer data instead of a good one:
This was enough to convince me to desolder the MASK ROM and read it in my programmer resulting in an empty dump.Direct replacement for this 8 Mbit 42PIN MASK ROM is a 27C800 EPROM though the silkscreening under the chip sais ’23C16000′:Fitted a freshly programmed device restored graphics completely.
Got this Pretty Soldier Sailor Moon PCB from my friend Ifog:
He told me that gameplay was fine but some of the audio/voice samples were missing or played wrongly.This was confirmed once I fired it up, here’s a brief audio recording:
As beginning of my troubleshooting, for first, I identified the audio digital circuitry:
As you can see from picture above, PCM samples data contained in the two MASK ROMs (40 and 42 pins) are read by two OKI MSM6295 PCM voice synthesis chips (while music is generated by a YM2151).So, I went to probe the MASK ROMs and found that the one @U47 had some address lines stuck as well as the control lines /CE and /OE.
Most of the address lines are directlty connected to the respective OKI MSM6295 chip so I replaced it but with no luck.Instead, I could trace the /CE and /OE lines back to a GAL16V8B @U28.Probing the outputs of this GAL revealed they were all stuck LOW.So. I tried to read it in my EPROM programmer which failed reporting an error:
I also read and disassembled it with my dedicated hardware, it was like empty :
This confirmed definitively that the device was bad.At this point the only possibility was taking a good GAL chip from another PCB or dumping and reproducing it.
Luckily my friend Ifog had a same board with a good GAL and he was able to provide me the binay dump which I took care of analysing and reversing it (see my latest PAL updates here).Once programmed a blank GAL16V8 device and fitted it on PCB, all the PCM samples were restored.Job done.
