Sega System 32 - Bad IC16

PCB problems and fixes
Post Reply
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Sega System 32 - Bad IC16

Post by neocps1 »

Hi,

Anyone fixed a system 32 motherboard with bad ic16 ?
Ic16 is the shared ram between main cpu and sound cpu.
It also serves as sound cpu work ram.

Most likely, when ic16 is reported bad, it is not really.

thank you
User avatar
leonardoliveira
Please Continue...
Posts: 692
Joined: August 30th, 2012, 5:53 am
Location: Brazil
Initials: leo

Re: Sega System 32 - Bad IC16

Post by leonardoliveira »

Reported bad could just mean the CPU doing the memory test could not reach it.

However, if you can verify that it did actually reach the RAM chip then the chip is bad.

So the proper course of action is trace the path from the said RAM to the CPU which handles memory tests (likely the main CPU) and have all traces checked/parts involved with memory address and data decoding verified. 8-)
Image
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

i am working on it. oddly, ic 16 error look to be pretty common
User avatar
leonardoliveira
Please Continue...
Posts: 692
Joined: August 30th, 2012, 5:53 am
Location: Brazil
Initials: leo

Re: Sega System 32 - Bad IC16

Post by leonardoliveira »

Has anyone successfully repaired that fault yet?
Image
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

i found one place where someone said it fixed the problem but couldnt remember how.

my first diagnostics indicate the ram is good, so does the z80. also the z80 is running since it goes crazy when i remove the ram.

the z80 look like it is stuck in a loop waiting for a value (rd and mreq active. iorq and wr arent).

the ram is shared between the main cpu (v60) and z80.

cs, oe, ce2 and wr look ok.

74ls244 around z80 look ok (only tried piggy backing a good one). 74ls157 decoder also.

maybe the z80 is stuck in early init waiting for some device which prevent z80 and v60 to communicate through shared ram.

this device may be a sound chip.

frustrating really :D
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

pal 315-5441 is good.

too bad there are no schematics for this board...
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

z80 data and address bus are directly connected to shared ram (8 bit bus).

this bus is not shared with v60 work ram bus.

v60 operates at double frequency than z80 and v60 has 16 bit external data bus.

so i suppose there are buffering chips between two buses to communicate.

z80 shared ram control signals are connected to 315-xxxx.
i assume it is good otherwise z80 wouldnt run.
User avatar
Asayuki
Please Continue...
Posts: 431
Joined: August 29th, 2015, 10:16 pm
Location: Remuria, Germany

Re: Sega System 32 - Bad IC16

Post by Asayuki »

What about the Z80's BUSREQ# pin? If I wanted to access the Z80's RAM externally I would definitely use DMA along with buffering. Maybe tracing that back would bring you either to the incriminated buffer or to some further hint from the scope.
My 15kHz cabinet Peplos will never power up, with any item, and I am quite proud of that.
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

Good idea ! Also, the WR signal coming from the chipset look like a bit weak. Do you think adding
a pullup can help ?
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

Asayuki wrote:What about the Z80's BUSREQ# pin? If I wanted to access the Z80's RAM externally I would definitely use DMA along with buffering. Maybe tracing that back would bring you either to the incriminated buffer or to some further hint from the scope.
Z80 WAIT input signal can also be used for this, isn't it ?
User avatar
Asayuki
Please Continue...
Posts: 431
Joined: August 29th, 2015, 10:16 pm
Location: Remuria, Germany

Re: Sega System 32 - Bad IC16

Post by Asayuki »

Using WAIT# would be far-fetched although possible if the circuitry is built in a certain way.
The most notable difference is that the address bus of the Z80 remains active during a WAIT cycle (http://www.z80.info/zip/z80cpu_um.pdf page 13), along with MREQ#, RD# and M1. In your particular case this would prevent the 68k from addressing the shared RAM.

EDIT: Z80's WR# afair is push-pull with threestate; I don't see why they shouldn't have done the same in the chipset. How high in voltage is it? Does it have rounded edges or other distortions? Does the WR# signal already have a pull-up somewhere?
My 15kHz cabinet Peplos will never power up, with any item, and I am quite proud of that.
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

I don't see any pull up for the WR signal. It looks a bit weak, so I think I can give a try adding one.
What resistor value can I use ?
User avatar
Asayuki
Please Continue...
Posts: 431
Joined: August 29th, 2015, 10:16 pm
Location: Remuria, Germany

Re: Sega System 32 - Bad IC16

Post by Asayuki »

Typical pull-ups are 10k. You can get lower but I would advise against a value lower than 2.2k.
IMHO it's not going to make much of a difference though.
My 15kHz cabinet Peplos will never power up, with any item, and I am quite proud of that.
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

making brief contact between A12 and A7 while in attract mode makes sound fulling working.

of course it doesnt fix the problem.

i changed both 74ls244 ic14 and 15 used to buffer the z80 adress bus for v60 access and now my trick doesnt work anymore. ic16 still bad in self test of course :D

i think i fixed something but it is not enough
User avatar
Asayuki
Please Continue...
Posts: 431
Joined: August 29th, 2015, 10:16 pm
Location: Remuria, Germany

Re: Sega System 32 - Bad IC16

Post by Asayuki »

Interesting. Are those address lines still toggling around? Well, actually you should double-check them all. You can scope-compare buffer inputs vs. buffer outputs qualifying with the enable signal. Also the data bus should be buffered as well, possibly bidirectionally. That should also get a good check up. I think you might be closer than you think to solve this issue. ;)
My 15kHz cabinet Peplos will never power up, with any item, and I am quite proud of that.
neocps1
Banned
Posts: 83
Joined: May 5th, 2012, 8:55 am
Location: France

Re: Sega System 32 - Bad IC16

Post by neocps1 »

thank you :)
Post Reply