Gunbird PCB (or other psyko games on this style pcb)

kemical · Post by **kemical** » Thu Mar 10, 2005 1:54 am

Can anyone check if their roms labeled 4 and 5 have 2 of the legs connected with a small peice of wire soldered across them?

kemical · Post by **kemical** » Thu Mar 10, 2005 3:52 am

actually i guess they might bypass something.. without them connected the game doesn't boot up, just sits on a garbled static screen.. 5 seems to be for sprites and 4 is backgrounds...

with them connected I get the backgrounds turned on and sprites.. but my sprites are garbled..

Post by **system11** » Thu Mar 10, 2005 10:54 am

Sometimes this is done to make a larger eprom behave like a smaller one, or sometimes to tie the CS or OE lines high. End result very much depends on what type of chip it is, and which legs are soldered together. If you're really curious and are able to post the chip type and which legs are joined, I can tell you what it's doing. Note leg numbering starts at 1 in the top-left, and ends with the highest number on the top right. Top = where the dot or cutaway is.

kemical · Post by **kemical** » Fri Mar 11, 2005 1:40 am

They are both 40 pin eproms, with legs 40 and 39 connected on each (getting +5v), the # on the chip is mbm27c2048-10 and 12

I'm wondering if the one that seems to relate to sprites showing up or not needs to be erased and written to again.. with the legs not connected on #5, sprites just show up as white boxes.. but with the legs connected the sprites show up, just with the individual stored tiles offset wrong or with the wrong color

d78 · Post by **d78** » Fri Mar 11, 2005 7:56 am

I'm pretty sure those 2 eproms are program roms.

And my board also has a small wire bridging 2 pins on both eproms.

If you are unable to get it working, I'd be interested in a particular IC on that board...

Post by **system11** » Fri Mar 11, 2005 11:54 am

On a 27C2048 eprom, pins 40 and 39 are Vcc and /PGM. Vcc is basically the positive voltage (5v). /PGM is the pin that controls writing to the device, in a similar fashion to a RAM chip. Note that /PGM instead of PGM means that it behaves inversely - if the pin is 'raised high' (essentially on) then it actually disables writing.

If you solder Vcc and /PGM together, you tie the /PGM line high, and disable device writing. Usually this is done at the PCB level, there'd be a track underneath the ROM socket doing the same job, but on boards where they might use different sizes with different pinouts (example - a 27C4096 has an address line where the PGM pin is on a 2048), you see mini hacks like this appearing.

So now you know (probably more than you wanted).

raiden · Post by **raiden** » Fri Mar 11, 2005 12:21 pm

may I ask where you got that knowledge from? Is it availlable online somewhere? In books? Or do you happen to understand electronics enough to see through the inner workings of a PCB? I find that utterly intriguing and fascinating.

Post by **system11** » Fri Mar 11, 2005 1:25 pm

raiden wrote:may I ask where you got that knowledge from? Is it availlable online somewhere? In books? Or do you happen to understand electronics enough to see through the inner workings of a PCB? I find that utterly intriguing and fascinating.

I checked the datasheet for 27c2048 chips - they contain a pinout, theory of operation and so on. The hardwiring high/low of pins done at PCB level is usually done with tracks (check underneath a few rom sockets one day, you'll probably find a few where pins are connected together this way) - but there's no reason an insane designer couldn't give those pins their own track back to some other component that keeps the line high/low.

I don't know as much as I'd like to about this stuff - what I learned was from asking other people who knew, and lots of learning the hard way with a broken board, a big stack of datasheets, and a logic probe. Actually checking data itself coming in/out of chips is beyond me, never needed to get into it (if for example you have jailbarred sprites, and the buffers look ok, you can usually put safe money on the sprite ram being faulty), but understanding basic logic and whether pulses are doing what they should isn't that hard for anyone to pick up a grounding in.

Take for example a 74LS04 - a very simple logic chip that basically inverts signals. Look at the datasheet:

http://www.datasheetarchive.com/datashe ... 09675.html

There's a pinout, a thoery of operation, the internal structure (many simple chips have the same function repeated multiple times - the LS04 actually contains 6 inverters), and the most useful part - the function table (sometimes called the truth table).

Of course some chips are far more complicated than others, but when you're down the simpler end of the spectrum, and even when looking at more complicated ones, using the datasheets, and having a basic idea of what you -expect- to see will allow you to spot anomalies. My first repair was an Up'n Down PCB. One day, it just stopped. Froze. Black screen. I didn't know what to do, but some great newbie advice from some friends in ukvac soon had me checking CPU and crystal clock lines.

Here's my repair log so far (out of date now, but hopefully shows the benefit of a little clear thinking, and a basic understanding of logic electronics) - the Up'n Down entry is on there:

http://homepage.ntlworld.com/jamesg22/repairs.html

Oh - and it's MUCH MUCH easier to learn this stuff with older boards. There's more space to work in, the circuits and components are far simpler, and frequently you can even get the schematics - which saves a LOT of tracing work. With some modern boards all you have is a mass of custom surface mount chips, all totally undocumented, and if you're really unlucky you cant find the traces as they're under the chips, or worse - embedded in a multi layer PCB so you cant even see what connects to what else. Repairing modern boards past non-custom or obvious failure, is pretty much a lost cause - even for people doing this stuff for a living.

raiden · Post by **raiden** » Fri Mar 11, 2005 5:46 pm

Wow. Although I probably don´t understand half the words you used, this opens up a whole new world to me. Have you seen the japanese guy who rebuilt a Galaga PCB? I´ve always admired him like hell, but had no idea how you can learn such a thing, even with all the descriptions he gave on his page. I feel like this PCB fascination is deep enough to keep me busy for the rest of my life.

dboeren · Post by **dboeren** » Fri Mar 11, 2005 6:50 pm

Very informative! I'd love to learn more about diagnosing board faults, so often a "tested & working" boards has minor faults on it, or an untested board is really close to being usable.

I've gotten some decent results just by checking for good socketed chip connections and cleaning the edge connector contacts, but I still have a couple boards that are *almost* working.

Raiden II: Game plays fine, but background graphics are garbled. Levels 1-4 have very small amounts of garbling, Level 5 is full of garbage.

Saint Dragon: Sound is too loud and distorted, also joystick 1 doesn't register when pressed right.

I'd really like to get Saint Dragon fixed to be playable. I just took it out last night, cleaned it up, and was pleased at how close it seems to be. (I had marked it BAD a long time ago and hadn't touched the board for a few years). After reading your page, I am planning to revisit it and pay particular attention to the miniature daughterboards to check for broken pins that could account for the lack of p1-right joystick. The sound I'm not sure on, guessing it could be a bad capacitor or something? The volume pot on the board doesn't seem to work properly either.

The only spare parts source I have is a messed up Bad Dudes board (game & sounds OK, graphics are bad) that I will happily steal anything from since I have another Bad Dudes that works perfectly.

I have a multimeter and soldering station, but no EPROM burner.

lmn4096 · Post by **lmn4096** » Fri Mar 11, 2005 9:21 pm

Ah finally !
thank you Bloodflower, I saw your page a long time ago but lost the link.
I tried to found it for month with no luck.

very informative. thanx again.

Post by **system11** » Fri Mar 11, 2005 9:42 pm

Raiden II sounds like roms - you need to read them to check them...

St Dragon - control is very likely the daughtercards, sound could be a blown amp or faulty capacitors (or worse - op-amp). If the pot isn't working, try turning the other one

It has vol left/right pots, only one channel comes out of the JAMMA interface. Switching it to mono might be a good idea too, if it isn't already.

kemical · Post by **kemical** » Sat Mar 12, 2005 12:46 am

thanks for the info bloodflowers

I bought this pcb as not working for about 10$, It had no sound and I was able to fix that (there was a leg broken off another eprom) but the garbled sprites I'm not sure what that is, everything else with the game seems to be playable and normal.
I'm thinking of getting a logic probe, I have a couple other boards with problems and messing with electronics stuff is enjoyable.