> Hi everyone.
> My name is Fabrice. I'm 36 years old, living in France.
> I just subscribed to this list.
> I'm currently trying to repair a SEGA vector PCB.
> Game is called "Space Ship" and is an exact clone
> of Cinematronics's Space Wars.
Fabrice,
For reference, this board is an implementation of what we refer to as the
"CCPU" (Cinematronics CPU).
Interesting. The board looks very similar to Rockola's CCPU implementation.
Rockola also used a large edge connector in place of individual connectors.
Cinematronics licensed this design to Rockola and apparently Sega as well.
When I get time, I'll find out if it's coincidence they both use the edge
connector or if the Sega design is descended from Rockola.
> There is no output sent to the X-Y screen.
That's normal (for a faulty CCPU).
> My oscilloscope says that the /RESET signal is periodically fired.
That's also normal (in a faulty CCPU). In the Cinematronics implementation,
they use that signal to light a fault LED.
> I understand all the meaning of the word "difficult" when it applies to
the CCPU troubleshooting ! ;-)
It can be difficult, but it's a matter of understanding the system and
applying the right methodology. I've been repairing those boards for 15
years, and you get the hang of it.
> I can seen important glitches on some of the E14 PROM data bits.
This is what we call a red herring. This is _not_ your problem.
At 2uS/div, your DSO may not be sampling fast enough to catch the complete
transition to the opposite logic state, before the next transition begins.
Remember, you're looking at the inside of a 5MHz CPU. Try 50ns/div or better
and see how it looks.
You mentioned there were glitches, but didn't mention if the device contents
were correct. If the contents are wrong, then I'd agree the part has a
problem.
That said, there are some weird artifacts (transitions) on the output of
E14, and you'll find they are coincident with similar transitions on the
inputs of the device. Most of those output signals are driving combinatorial
logic, which isn't bothered by this behavior; when the signal stabilizes, so
does the entire chain. However, if you're seeing glitches, for example, on
the output of F14, which drives _registered_ logic, that would be a problem.
Regardless, this is normal behavior for E14.
> Does anyone know if its a normal behavior for this chip ?
Yes, see above.
> Maybe address bus is changed on clock's raising edge and data bits
> are used on clock's falling edge... In this case, it would be OK.
Again, not your problem.
If you really want to test the prom devices, the best way for you to do so,
is:
1) Substitution
2) Data contents validation (if subs aren't available)
Signature Analysis would be the first choice, but it doesn't appear your
board will easily lend itself to that methodology.
FWIW, I've never seen a prom in a CCPU that didn't work, but did have valid
contents. Therefore, if you validate contents, the probability is very high
that the device is good. Further, PROM failures on the CCPU do happen, but
are statistically very rare. I base this on having repaired hundreds of
these boards.
I'd offer to fix your board (see cinelabs.com), but you indicated you'd like
to fix it yourself. My advice would be to not do what many other people have
done, which is to take a shot-gun approach, which makes the board look like
a mine field, and frequently introduces more problems than it fixes.
Hope that answers some of your questions.
au revoir,
-Mark
P.S. I renamed your thread, so people searching for information about Space
Ship can tell what this thread is about.
P.P.S. Is your control panel missing buttons? Space Wars has 10 buttons,
used to start the game. I see them on the game flyer, but not in the picture
of your actual game.
---------------------------------------------------------------------------
** Unsubscribe, subscribe, or view the archives at http://www.vectorlist.org
** Please direct other questions, comments, or problems to chris@westnet.com
Received on Sun Jan 10 19:28:13 2010
This archive was generated by hypermail 2.1.8 : Sun Jan 10 2010 - 21:50:01 EST