Hail to Galaxian... by Namco!

[level42]

Ok stupid me, looks like I "fixed" the existing track for write enable to the wrong pin (14) on 8M. Not with the jumper wire but while soldering some small repairs (gotta love converted boards....).

Ok so, the write enable line of the onbject Rams were connected to two outputs of 8M instead of one so I fixed this. (Pin 14 is unused here so it should not connect to anything). Now indeed I have very nice toggling on the write enable line when running the board !

However I  still have garbage, BUT there seems to be a bit of improvement. I can actually "read" some messages now (like ROM Ok) and f.i, this screen which should be saying
"TESTING RAM DON'T WORRY"

Also, the entire screen should be black. Now part is, but part is filled with B's. At least there seems to be some consistency here.... It feels like the test program is calling a certain character from the graphics ROM (a "blank") but every other line it seems to return a B instead....as if it's getting part of the info, but not all, like 1 bit missing....im thinking maybe one of the graphics EPROMs is not OK, but they do verify OK with Romident, but I've read that under high speed conditions an EPROM. Might behave different than on an EPROM reader....

OK apart form this the object RAM still produces errors while running a RAM test on the Fluke, although now it's simply a fact of that its writing 00's but it's reading AF all the time...well almost all the time, when I loop I sometimes get another value but usually AF, AF, AF.....and seems to happen on various memory addresses within the Object RAM range....

[level42]

Mmmmm, found out that A2 (pin 2) on both 2101 RAMs is constantly low while all other address lines pulse.

This leads back to 74157 at position 5E. The corresponding inputs are grounded according to schematics (pin 10) and the other one (pin 11) pulses away happily.
All other outputs pulse.
I have a feeling this 74157 migt be dead.

[level42]

Found a nice overview for checking Galaxian PCBs here:

http://www.paladingrp.com/brianb/board_fixes/galaxian/galaxian_info.html#object-RAM-troubleshooting

[Etienne MacGyver]

You'll get it working i am sure  

I like to folow your progress, its way beyond my knowledge, but still interesting to read.

Don't know if i linked this before, but Adam over at KLOV does some great repair vids, and one of them is a galaxian repair.

don't know if it helps you in this case, but i like them vids.

He also uses the fluke alot, so that can come in handy also  

[level42]

Yes, great addition to the thread Etienne ! Indeed I watched it before and I really liked his drawing because it gives a nice overview how stuff interacts.

I also saw his (?) Punch Out videos using the Fluke, even if that's not the game I'm (or anyone) is working on, it is still very useful !

His Fluke is a 9010, I have a 9100. The 9100 is more advanced in about every way you can think (so advanced, you COULD program a script that would test a board from front to end in one go....but it would take YEARS to develop something like that.....this was only useful for factories. It can even "drive" a robot tester....

However, the 9010 is so much more used in this hobby that there is more info about that one available. Message returned by the 9100 are different (more clear text compared to abbreviations). It also does the tests better and faster AFAIK, but still....since "everyone" has a 9010 it's sometimes a bit confusing.

The 9010 is definitly already more than what you'd ever need for the hobby, the 9100 is 100x overkill and I use only the basic functions.

The probe is brilliant though, even though I could just as well use a 10 bucks one

I am aware that it is hard to follow by texts only. You'd need to have at least the schematics in front _and_ understand a bit what everything does, but I'm also writing this for myself and hopefully some others can use it in the future.

I plan to make a video of how I work with the 9100A. It's a very cool device, especially for free 

[level42]

Just for fun I entered 74LS157 in youtube.....and look what I found.....

http://youtu.be/Z055lTbPPAg

Thats actually a pretty useful little program and I downloaded it

[level42]

74LS157 replaced by 74HCT157 @ 5E (the only thing the shop near where I was working today had in stock).

The good news: Success !!!! The output (pin 9) is now pulsing now and then when running UUT, this definitely did NOT happen before.
The bad news: the goddamned screen still looks the same and RAM tests @5800-58FF still only return AF......

So I fixed something that wasn't even apparent yet because there's a bigger problem that hid this one .......


Out of frustration I replaced both graphics EPROM sockets, but I didn't suspect them although there were a bit loose....all signals seemed OK and indeed after replacing them no change (but it feels better to have the new sockets in )

[level42]

I posted about my problems on KLOV and Brian Brzezicki posted a link to his AMAZING website:

http://www.arcade-cabinets.com/resources.shtml

His site has BRILLIANT walk throughs of how to work with a Fluke and explaining how the hardware works along the way.

This is a MUST read even if you don't have a Fluke, it will still give a good insight in ow everything works on a classic arcade PCB.

The Galaxian page I linked to earlier is actually a part of this site.

[anunaki]

great link

[Q*ris]

great link

+1!!

[ajhippel]

are the namco schematics similar to the midways?

i want to have a look ;-)

[Laszo]

Well no succes so far but really nice blog post. We are all with you on this one. Keep it up 

[level42]

Actually, Muerto was so kind too send me the original Namco schematics. They are still somewhere between Denmark and The Netherlands though, but I plan to scan them hi-res. totally awesome gift from Michael, I love to have original docs with my machines !!!

One thing that is surely different is the audio output stages, and Andy from the UK discovered that Namco boards have different resistors in the video output stages which results in slightly different colors IIIRC.

Thanks Leo

[ajhippel]

staring at the pic i think the following...

bit 1 (started counting from 0) is always high every 2nd line (or every 32 characters)
that means bit 1 (decimal 2) is always "set" to the "normal" char...

that means in the char-roms the address line a4 is high every 2nd line (or 32 characters).

(the address line a0-a2 are the 8 bytes for the charlines, a3=ok,a4 is high)

going more "backwards" in the schematics -> that could mean: whenever the VRA5 is high, the VRAM1 is high as well

it still can be a lot...

but try checking this:
3k 74157 mux
check for a short on 1a(pint2) /1b (pin 3)  (should be open) - but it seems that this one's working

3h Ram 2114
check for a short on VRA5 (pin2) and VRAM1 (Pin12) (should be open)

on a lot of old boards, the 2114 die... in a strange way

if you didn't replace this 2114 yet...try this one...


the prob could (and should) be around this section (3h,3k,4k)


[edit]
It could also be, that the cpu can not write into the char-ram properly...
check also 5j Ls245 (pin 4)
 

[ckong]

My thoughts exactly.  LOL