T2 resetting

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Got a T2 recently (full write-up is pending) and put the board on my bench (it's mainly jamma).
It's currently resetting every 2 seconds.

Checked the clock and that's ok.
The Watch Dog signal at U39 pin 11 is constantly low.
U39 pins 15 & 16 are High and pulsing low every 2 seconds (reset)
U99 (large custom) reset point (N15) is high pulsing low every 2 seconds (reset)

So what would cause the reset signal to trigger?
What should I check / do next (verify / burn roms etc)?

Nes4life2016-07-25 13:18:33
 

FinalFight

Active member
vacBacker
Feedback
73 (97%)
Credits
770CR
I dont know if it helps but The MK and T2 pcbs are very picky with voltage they have to be dead on otherwise resetstegrey2016-07-25 13:43:43
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Thanks. That was the excuse I needed to hastily add my voltage displays to my jamma harness.
It's at 5v but still resetting.

I think I'm going to go for a full validation / reburn of the roms next.

 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Took off all the EPROMs in preparation for reburn and 3 of them left legs behind! I was pretty gentle but I guess they're just further gone then I thought. I'll post them off tomorrow and then replace the sockets still retaining the legs. Good desoldering practice.
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Ok got all the Eeproms reburnt thanks to t-m. Still resetting. Double-checked the schematics and the Watchdog is active low; so it is watchdogging. Typically what does the watch dog circuit monitor and what should I check? Realised my oscilloscope is 15Mhz so can't check the clock signals that way.

Which ICs should I validate / check the pins of?
Can anyone highlight particular info on the schematics so I can learn as I go?
 

porchy

Active member
Feedback
5 (100%)
Credits
994CR
U89 and U105 are your main program ROMs. Typically the watchdog will reset if the CPU can't execute the code properly or the code itself let's the watchdog timer run out e.g ram test fails etc
Check the program ROMs sockets and maybe do a continuity check of the pins
 

Ace`

PlanB contributor #1
Feedback
40 (100%)
Credits
4,038CR
I wouldn't trust the wires on those cheap chinese harnesses, especially for gnd, 5v and 12v they're really thin wires with thick plastic around them to make them look like a decent gauge.
 

simonden

Active member
vacBacker
Feedback
20 (100%)
Credits
1,042CR
Same as Ace' and Purity, those harnesses can be real crap. Also don't trust those 2 pin voltmeters, they need calibrating for every different load (screw on back) so unless you have it calibrated for that PCB, those readings will be wrong. Only good to solder to a single test PCB and calibrate for that, otherwise no good for a test rig.
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Thanks all. I'll get a better jamma extension in future.
I checked the continuity of U89 and U105 and the pins are fine.
Managed to follow a few of the address lines (A0 to A4) back to the U99 custom.

I'll check the reset button now
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Reset button is fine. Reset remains at 5v and goes to ground when button is pressed (ignoring the regular reset signal every second assumed as a result of the watch dog activity).
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Just tried it in my jamma cab and it's still resetting.
Could another IC be bad and be putting junk on the address bus or data bus or is there something else more fundamental I should check / replace first?
 

t-m

Active member
vacBacker
Feedback
2 (100%)
Credits
333CR
Judging by the state of the eprom legs check anything else that's in a socket including that PLCC if you can remove it.
 

limpy

User
Credits
132CR
just a idea.my T unit board only works if it gets 5.16v
5v makes odd things happen reboot being one of them.

remove the battery as the save data my be corrupt too.
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
Pcb is now with Royal Mail on its way to VectorGlow.
Like the old Tango slogan... I need it because I'm weak.

Looking forward to getting this cab going; I've a feeling I'll be calling on the talented Gunblade next.
 

RGP

Meeter & Greeter
Feedback
5 (100%)
Credits
2,039CR
A pity I hadn't seen this thread earlier when you pm'd me the other day.

If you know where the watchdog signal is then you can usually stop it - on most boards you tie the WD signal to GND somehow - bit generalised advice but usually the case. That would let you see a better idea of what's happening.

CPU has to talk to ROM and RAM through something - there are usually two things in the way - a buffer chip between the raw address/data of the CPU and the ROM/RAM - in a lot of cases an LS244/LS245 combination. There are signals that enable the desired ROM or RAM to become active and there are signals that drive the buffer chips direction or just select them.

Sounds like the main CPU is accessing the boot ROM at least and starting to perform the boot tests and then running into trouble, halting and then the watchdog kicks it - the fact its going 2 seconds is relatively significant in that theory.

As others have said above, on Midway boards like these, its not voltage at the edge its voltage at the chips - there's a huge drain on voltage across these particular boards so while you might be sending 5.02 in, it might be dropping to 4.7v later down the line. Meter an LS series chip across its two corners (usually pins 7 and 14 or 8 and 16) to get a decent read on voltage level and adjust to make this above 4.80v at the very least.
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
RGP, you sir are a repair guru, and I thank you for liberally and freely dispensing your knowledge. I know with the help of yourself and the forum I could get the T2 board going again but it might take me a while. When it comes back repaired (or beyond repair) then I'll let you all know what was wrong with it and how it was fixed (or not) the we can all benefit from the knowledge.

Looking forward to the next repair... perhaps I'll see it through to completion myself
smiley2.gif
 

ragnar0k

Newbie
Feedback
1 (100%)
Credits
13CR
Stupid T2 boards! Mine is getting stuck after checks and I lost my head on it since months. Still don't know what to do...
I am also scared that when we re-burn the roms from scratch the encryption is broken and things get worse...
 

Nes4life

Active member
vacBacker
Feedback
12 (100%)
Credits
1,134CR
The main PCB has now been repaired by VectorGlow
smiley1.gif
for which I am very grateful. Great service and communication. He was also kind enough to give me an overview of the repair (which obviously cuts out a lot of the investigation work that went into the repair):

- Probed program code EPROMs and some address lines were dead

- Traced lines back to a dodgy logic chip and replaced it

- checked outputs on the large RAMs and some were dead so replaced two RAMs

- verified EPROMS and replaced two of them

- game booted into the test screen and showed 3 more faulty RAMs

- replaced group of four RAMS so they are all the same type (and likely the 4th would eventually go)

- game booted and graphics were fine

- battery and battery holder replaced

- replaced all three EPROMs on sound board as were corrupted

- replaced large resistor on coil board as it had disintegrated




Further suggested work:

- sound board sockets should be replaced (YM2151, YM3012 and CPU)

- EPROM sockets and 40-pin socket should be replaced
 
Top