New Zealand Story Repair Log

[Womble]

Another one of my all time favourites found its way to the repair bench, The New Zealand Story, quite a rare PCB this one.



The board was constantly rebooting and coming up with a SUB CPU ERROR message.

YouTube- Faulty Arcade PCB - The New Zealand Story

There are only two EPROMs on the board, and a load of mask ROMs that probably contain sound and gfx data (usually the game program was held in EPROMs so bugs could be fixed, gfx and sound data was unlikely ever to be changed.) So to rule out a software issue I dumped and compared the two EPROMs to the MAME set, they were fine, so it looked like a hardware problem.

The constant rebooting was due to the watchdog system that resets the CPUs if it sees no activity on the buses, this stops the boards spending hours in a hung state in a cab somewhere, the only problem is that if the fault doesnt clear the system will reset endlessly. When the watchdog is barking it is very hard to see what is going on as everything is constantly bouncing, so I went looking for a jumper to stop the watchdog from resetting the system. Original board often have one to aid troubleshooting and this board is very well made, it even has 2 raised ground loop rails to attach the scope probe to. Often this jumper is marked JP1 and is usually located very close to the reset button, if there is one. This board does have a reset button and the jumper nearest to it stops the watchdog, its called JP10 on TNZS boards though. To be as neat as possible I didnt solder the jumper pads together, I buzzed the pads back to nearby component legs and soldered a loop of hookup wire between the legs, almost invisible when removed.



This shut the watchdog up and left the board stable displaying the "SUB CPU ERROR" message consistently. With everything stable I could go round the RAM chips with the scope, a couple of lines didnt look too healthy but as the board had presumably crashed that is not that uncommon.

The board does have a few of the dreaded Fujitsu TTL chips on it, 245s, 157s and 161s, and going over the 245s I found something odd at U27. The 245s are octal bus transceiver chips, that can allow data to pass from one side to the other, or visa versa, or they can isolate the two sides. The Enable pin was logic high which isolates the bus, so the 5V side of the chip had floating outputs, which was normal, the input side had 6 inputs that were healthy, but A3 and A4 were tied high, now this could be normal if the bytes of data banging up against the chip always had 1's at bit 3 and 4, but its not that likely, knowing the way Fujitsu chips die it was more likely that these pins had just got stuck, and the chip was no longer able to pull the outputs down to logic 0.



So I piggy backed a new 74LS245 chip on to the existing chip and hit the reset button.







Desoldered the dodgy chip and fitted a new TI brand one to complete the fix.

Bloody love this game!!! Is it wrong to own 2 PCBs?

[cools]

Bloody love this game!!! Is it wrong to own 2 PCBs?

I've had all three versions at one point or another

[thegreathopper]

Excellent thread.

Thanks.

[markedkiller78]

There is a lot of love here for Taitos gems

Great thread. It's nice to see the elimination process.

[toki]

This stuff is gold, Womble. Kudos for all the effort you're putting in.

[cRacKh0rN]

A really great (and detailed) post. Nice one fella

[channelmaniac]

Nice amount of details!

[Womble]

Niiiice - have trawled your page many times before, awesome collection!!!

[invzim]

I now have a faulty PCB just like this one (got it as non-working) - how the F do you desolder these DIP's? Never met a more stubborn pcb ever. Is it multi-layer lead-free? Tried the usual tricks of adding some flux, "reflowing" with fresh solder etc. The data pins are not too bad, but the 5V and GND pins are a nightmare.

[jugemscloud]

cools wrote:

I've had all three versions at one point or another

how do you tell the difference i thought there was only original and revision (with daughter board)

[Womble]

I now have a faulty PCB just like this one (got it as non-working) - how the F do you desolder these DIP's? Never met a more stubborn pcb ever. Is it multi-layer lead-free? Tried the usual tricks of adding some flux, "reflowing" with fresh solder etc. The data pins are not too bad, but the 5V and GND pins are a nightmare.

Its normal lead solder but there are large sheets of copper in the board for +5V and ground, these act as massive heatsinks when desoldering so its very hard to get it all hot enough.

[channelmaniac]

With my desoldering station I put the iron on the top side of the board where it touches the leg and the pad... then put the desoldering iron on the bottom side and remove the solder.

Works very well.

Another technique is to crank the heat up to about 750 or 800 degrees but that will cause the desoldering tips to discolor and shortens their life.

[invzim]

Used a cut of syringe needle (from a paste dispenser), put it into the hole and applied light pressure while heating the needle and pad with hot air. Sort of got the job done.

This board has two errors, first one was the sub-cpu error, and original fault is it fails to register the start button. The original repair attempt (not by me), was to replace the 74ls245 womble did. On closer inspection, I saw a busted track under the replaced 74ls245 , probably happened during desoldering. I put jumper wire to bypass the problem, and voila - the board boots fine and registers coin input, sound also working. Lessons learned is to always be very suspicious of previous repair attempts I guess.

Now I'm working on the game not starting. The pull-up resistors array seems ok, and with a logic probe I can see that the signal is high, and is sunk low on input. HOWEVER, I for the life of me can't seem to be able to trace where the input actually goes beyond the pull-up resistors. No traces visible on the top/bottom layer, and randomly testing pins on the board to see (hear) where it goes has produced nothing so far.

[Womble]

Lessons learned is to always be very suspicious of previous repair attempts I guess.

Very much so, the number of times I have found that an earlier repairer has replaced the original chip with another faulty chip, or damaged the traces and flooded the pins with solder hoping to bridge the gap is fairly high. I have no idea if they destroyed the replacement chip with a brutal desolder, or managed to kill it some other way, but other peoples work is usually well worth investigating. The only way to rule it out is usually to remove it and re-do the work.

Now I'm working on the game not starting. The pull-up resistors array seems ok, and with a logic probe I can see that the signal is high, and is sunk low on input. HOWEVER, I for the life of me can't seem to be able to trace where the input actually goes beyond the pull-up resistors. No traces visible on the top/bottom layer, and randomly testing pins on the board to see (hear) where it goes has produced nothing so far.

In cases like this is often more useful to trace other control lines, if the coinup failure is due to a blown track you will never find it with a continuity meter as the lack of continuity is the fault. Your best bet is to buzz through the 2 player start, you will likely find whatever chip it goes to has a floating input, that is likely to be the missing input for the 1P start. I will see if I can buzz through the 1P start on my board tonight.

[invzim]

In cases like this is often more useful to trace other control lines, if the coinup failure is due to a blown track you will never find it with a continuity meter as the lack of continuity is the fault. Your best bet is to buzz through the 2 player start, you will likely find whatever chip it goes to has a floating input, that is likely to be the missing input for the 1P start. I will see if I can buzz through the 1P start on my board tonight.

Most appreciated! Already tried to trace the two working inputs, coin1&coin2 but nowhere to be found, which is really strange. When tracing unknown signals, I use a piece of steel wool attached to a multimeter to help me narrow down where it ends up on a board, before going pin to pin - but no luck so far on this one.

[Womble]

Hmm, well its not quite as simple as I thought it would be, the 1P coin up line goes to pin 9 on U58 which looks like a standard resistor network, but possibly isn't. It might be a dual network, which may explain the "X2" part of its name. There is no info on it but at a guess it consists of 4 units of two pins which, one of which is tied by a pull up to pin 1, the second pin in each pair is probably tied to the 1st by another resistor and that then goes off to the SETA X1-004 chip behind the 3 R nets.

I only left myself 5 mins to check it out this morning, but I would guess that pin 9 is the outward line for the 1P coin up and it goes to the via which is protruding from the right hand side of the X1-400 chip, I didn't have time to trace it further.

Can you take resistance readings from pin 8 to pin 1 of U58, and of pin 9 to pin 1, and also pin 8 to pin 9, we can compare it to my board which might show if your X2-500 is bad, I will try to trace where that pin 9 actually goes.

[invzim]

Will probe around the resistor array(or not) tomorrow after work - hope I don't have a shot custom..

[Womble]

Always a worry, but customs tend to be extremely hardy, a bad track or via is far more likely, 9 times out of 10 when I have a fault that could be a custom it turns out to be something else, the fact your board runs means it hasn't suffered any catastrophic event that would likely kill a custom.

[invzim]

Did some tracing and probing, and the X2-005's are indeed not resistor arrays. The pinout and operation is:

pin 1 - GND
pin 2 - VDD/5V
pin 3 - pullup/ jamma input
pin 4 - high on idle, float when button is pressed
pin 5 - pullup/ jamma input
pin 6 - high on idle, float when button is pressed
pin 7 - pullup/ jamma input
pin 8 - high on idle, float when button is pressed
pin 9 - pullup/ jamma input
pin 10 -high on idle, float when button is pressed

pin 3 corresponds to pin4 etc. All output pins EXCEPT coin1&coin2 goes to the seta X1-004 custom IC, while coin1&coin2 goes to the Taito B53-09, which according to mame is an intel i8742 MCU. Further probing revealed that the 8742 were able to pull the lines to GND on input, while the lines to the seta x1-004 simply floated.

It seems to me the culprit is the X1-004 custom, will probe a bit further tomorrow.

[mastercello]

Did you have had any luck with this error?

I have a gameboard (J.J.Squakwers) with these X2-005 arrays - and it looks like that one of them is broken.
At least i have one input (P1 Down) all the times "pressed" and the main character is on its knees all the time.

P1 Up movement lets just reboot the board - kind of a reset switch.

I was working with a resistor meter over the array, and the pins where the "down" buttons is connected i have a different reading like on the "good" ones.
In fact the resistance on a good one is like 22k on the bad one it is just the half (~10k).

Should i change the array?
Did anybody know where to get them?