Generic Galaxian - Resurrection from the Dead

[Van Diesel]

Hi folks

Once again I feel I should publicly thank Mr_goshki for selling/sending me a big box of boards at a fair price for me to wade through and get some fixing experience back (since I've been out of this lark since about 1993!). Many thanks to you, Sir.

Right . . . this page will act as a place holder as I wade my way through the board and attempt to get it to 'Rise from it's Grave' as the Altered Beast saying goes.

Here we have a two-part Galaxian / Moon Cresta board. I assume originally from a cocktail cabinet, so it could fold. If this board has a more formal title, please let me know.

Somewhat of a resurection as you can see - No CPU, No RAM, No ROMS, No Colour PROM . . . It's as dead as Python's Parrot . . . but the essense of a possible worker is there . . . and it's not nailed to it's perch.
Good to see that the two ribbon connectors are still in place.

Ok, first thing's first - I've made a big order of bits from China and Poland which I will detail fully on the next page when they arrive with me, including prices paid. I intend to burn a few ROMs including the Galaxian TEST ROM (which I've never used) to get some experience and have a shed-load of fun or frustration in getting this board working again.

I'll need a Colour PROM '82S123' - If anyone has one for sale or knows of a source for the standard Galaxian Prom (not sure if it is the same colour program as is used on a Moon Cresta), please let me know.

More to follow . . . stay tuned . . . don't touch that dial . . .

Van
Van Diesel2020-08-26 17:15:48

 

[jonhughes]

I repair the odd Galaxian PCB so have a range of spares.

I know I have the video RAM, an unwanted daughter board that would match very well to your board, the sprite ROMs (not RAMs) and I'll look out for the bipolar chip.

I have the daughter board in my hand so just need to track down a bipolar for you.

 

[Van Diesel]

That's extremely kind of you, Jon.
I'll PM you . . .

 

[Van Diesel]

RIGHT! Galaxian resurrection Part TWO
(Similar game but with different colours and with slightly different sound effects . . .)

First of all, I did some work on the board just by looking at it and seeing if I could find any physical faults or problems . . .

PHYSICAL FAULTS

1)
Three 10nF decoupling capacitors broken from around the boards -
Typical problem with a board that's been moved around a lot - They get
damaged easily. I replaced them.
2) The GREEN and SYNC lines were cut close
to the output of the edge connector. Huh? Snakes alive, why would you do that? Bizarre
- Anyway, I linked them both across to the output with some PCB wire.
3)
C39 broken. This is the pass-through capacitor feeding the volume
control, so it's carrying the sound. With this capacitor broken, it
means no sound in the game! I replaced it.
4) 5V power being fed to the board through a
single thin PCB wire. Huh? Gadzooks, this looks original too! The wire looks
completely inadequate to carry the current needed for the entire board, so I
got rid of it and linked over with two wires more capable of carrying
the load on the 5V rail.

Also, I wrote the voltage rails onto the board itself. NOTE: I certainly wouldn't do some of these things on to an important or original board . . . but this is a 'worker' board I can use for fun stuff . . . like using it to make an interesting Fix Log for people to read, or for people to make a realistic judgement about the state of my mental health.

5) Resistor R77 missing (removed by previous owner). This is one of the resistors at the output of the colour PROM, responsible for the resistor ladder for the correct colour levels. Strangely, although this resistor is in the manual schematic symbolically, it's not mentioned as being R77 or given a value. It is a 470 Ohm resistor. I put a spare in the board.

Mentioning the colour PROM . . . I needed one. A spare was kindly sent to me by UkVAC member JonHughes (thank you, Jon). Jon tells me that although it 'works', it's not programmed for the correct colours used in Galaxian. That's fine - It's something I can use to get a decent picture to begin with. I intend to program a correct colour PROM myself (by hand!) later on . . .

------------------------------------------------------------------------

ADDED LATER - SILLY BOOTLEGGERS!

If you look at the picture above, you will see that three of the resistors are non-standard . . . they are dark green in colour. These resistors were the wrong values! They were 22 ohms and should be 220 ohms. The bootleggers put in the wrong values!
These incorrect resistors will cause some of the colours being displayed far brighter than they should be. I changed all three for the correct values. Now the resistor ladder is correct 220 / 470 / 1000.

-------------------------------------------------------------

IC SOCKETS

You may notice that the socket for the PROM in the above picture is an 18 pin, but the MMI6331 is a 16 pin device. This was a typical standard install across this board from original, obviously in an effort to save money (only buy one type of socket).
If you look carefully, you can see I gummed up the two holes with Nail Varnish . . . some purple glittery stuff I wear at the weekend when I frequent the bars in Norwich looking for sailors.
Also, I think it adds to the magic of the game, don't you? . . . and prevents anyone else from inserting the chip in the wrong way in the future of course!

Whilst examining the board, I noticed that the NMI line had been cut and then reconnected.
I'll make a mental note of this in case it's something that requires attention later on . . .

OK . . . now for other stuff . . .

PARTS PURCHASED INCLUDING COSTS.

P2101A RAM x 6 . . . . I only needed 2 chips, but I thought . . . what the hell - £13.05p
UPD2114LC x 5 . . . . I only needed 4 . . . one for my spares . . . - £4.23p

Z80A CPU - I only needed 1 . . . 1 bought 2 . . . - £2.92p
2716 EPROMS . . . I bought five - I only need 3 - £4.26p
MMI6331 PROM . . . (I'll program the correct colours myself later!) - I bought 3 - £8.75p
74S201N Attack RAM . . . I bought 6, although I needed 5 - £6.34p
(These latter parts were a small gamble, as they were 'untested' from an old Frogger board . . . time will tell if they are good or not . . .)
Speaking of which . . .

You'll notice the magical glittery nail varnish again Hey ho <saucy wink!>
These ICs will need heatsinks later on . . .

I programmed the two graphics ROMs that Galaxian needs, gave them tidy labels and populated them in the correct locations on the PCB . . .

Once again, I am indebted to JonHughes for providing me with a Galaxian Daughterboard and a ROM set at reasonable cost. I programmed the TEST ROM myself . . .

I just need the CPU to arrive . . . and then . . .

WHAT'S GOING TO HAPPEN WHEN I SWITCH IT ON?!
The next thrilling installment coming soon . . . don't touch that dial!

Van
Van Diesel2020-09-16 19:48:46

 

[Van Diesel]

You wanted it . . . and it's HERE - Galaxian Resurrection part THREE!

'Further into Unknown World . . . well . . . unknown PCB!'

I know that you are all waiting with antici . . . . . . . . . . pation to find out what's going to happen when I power this board up . . . but just before then, a few checks.
I checked the the two power rails were not shorted out. . . they weren't . . . all good to go.

POWER UP!

Well . . . it's not dead.

Actually this is rather pleasing for me as it tells me that a great deal of the board is in working order. Remember that this board had no known history whatsoever. Anything was possible.

This was with the Galaxian Test ROM installed at the correct position (7H) but I had no experience with what to expect, so this was quite a learning experience.

It seemed to me that there were at least two individual faults apparent.
1) The program didn't seem to be running
2) A graphics issue.

At this point, my (in)experience had me poking around the two 2101 Rams (Object Ram) as the outputs appeared to be stuck low. I didn't really find any obvious problems, and it wasn't really active i.e. being addressed by the CPU. Further about this later on . . .

I checked the CPU - Not resetting. It seems that the test ROM code is running, but looping around continuously.
I checked the CPU buses and the associated buffers - All seem OK.

GALAXIAN TEST ROM

As I understand, the very first thing that the test ROM does is to check the Work / Game RAM at positions 7N and 7P. If it can't access it at all, then it does nothing. This is what it seems to be doing . . . or rather not doing?. . . whatever.

My search continued to the addressing of the Work RAM.

It seemed that the outputs of the 3 to 8 decoder at 8N (74LS138) were stuck. At no stage did the outputs change, and the RAM _CE line to the AND gate at 2N never got triggered from it.
I changed this IC . . . let's see what happens now . . .

'Uffed'?

Could it be that the board is now telling me to 'Get Stuffed'?
How rude!

Actually what is happening now is that the code in the TEST ROM is now running! Huzzah!
There is a graphic issue which prevents the full text being displayed. The 'uffed' is for 'RAM stuffed' i.e. knackered, kaput, broken, fubar, not working.

This 'uffed' message was displayed twice while the code was running doing it's tests, so I deduced that the board cannot access both the Video RAM and the Object RAM. Perhaps this is a major addressing issue affecting both these parts. I am fairly certain that the actual RAM itself (purchased recently) is good.
The test program cannot directly test the Attack/Sprite RAM, so it displays moving graphics in between the two tests for a few seconds for you to see if there is anything missing or obviously wrong on the display.

At this point I went poking around several points on the PCB to see if I could work out what the problem was, but everything I tested appeared fine.

Whilst fiddling around, removing and replacing the Object RAM I noticed something interesting.

I circled it in red - A tiny, tiny cut in the track underneath the turned pin socket. I can only guess that this was caused when a previous owner of the board replaced the sockets (or RAM ICs) with new sockets and inadvertantly slipped with his tool.
No disrespect to him. I've inadvertently slipped with my tool many times. I'm sure we all have . . .

I checked the schematic and this cut was of the CPU Address bus line A8 which feeds further up the board (to form the Video RAM address bus). My meter showed that it was indeed 'open circuit' so I used some PCB wire underneath the board to bridge the cut.

Power back on and what do we get?

I know it looks like 'DINT WORRY' but please be aware that my current test monitor is suffering from Foldover at the top. The SCART input can't cope with the input from the board. This isn't a board fault. The image is fine.

Let's replace the Test ROM with the Galaxian ROM daughterboard and see what we get . . .

Hell yeah! She's a good-un!

Incorrect colours of course due to me using a test PROM, but I knew that would be the case.
The text '1UP HIGH SCORE' should be in white, not blue . . . the 'We are the Galaxians' should be in red etc

Tapping the edge connector inputs with a ground wire and I was able to coin-up, start the game, move the Galaxip and fire - Oh yeah!
There were no obvious missing or garbled sprites during the attract mode so it looks to me like all the Attack RAM is in good order too.

Next part . . . fixing the sound (which has issues) and programming a PROM with the correct colours (by hand).

Coming to you really soon . . .

Van

Van Diesel2020-09-16 19:34:34

 

[jonhughes]

Great stuff - hopefully it's just the amp and then you'll be up and running in odd colours.

If it's anything else like missing sound effects let me know and I may be able to help.
jonhughes2020-09-16 19:37:21

 

[patzik]

Don't know if will end up being an economical repair, but makes for a great read, so thank you.
Interested in programming the prom by hand, can't wait to see how you will do that.

On to part four...

 

[yoganuggy]

Nice job so far. You are certainly enjoying the hobby.

It's
unfortunate the replacement part costs add up. Great fun to repair as a
hobby but you can see why it's better for most to buy a new board what with the time, effort and cost of replacement parts.

Some
boards I've spent more on parts that it would cost to buy a replacement
board, but then you would miss out on the repair fun highs and lows
roller-coaster. It's working YES!!!!! but then as you dance around the
room you realise there is no sound NOOOOO!!!

 

[Van Diesel]

Hi again folks and friends

Once more, an update to the fix! But before then, I'll answer a few questions from my adoring fans . . .

Don't know if will end up being an economical repair, but makes for a great read, so thank you.

Thank YOU, Patzik - I'm going to try to do more fix logs in the same style - Relaxed - Hopefully informative - Slightly comedic and definitely fun!

It was never going to be an economical repair, but it is unclear what PCBs are worth nowadays.
I've been offered faulty Galaxian boards from between £20 and £80

.

Some
boards I've spent more on parts that it would cost to buy a replacement
board, but then you would miss out on the repair fun highs and lows
roller-coaster.

You are absolutely right YogaNuggy. I doubt a Golfer spends much of his time thinking about his Club Membership costs whilst he is enjoying his 18 holes. It's not all about money . . .

Interested in programming the prom by hand, can't wait to see how you will do that.

I'm glad I piqued your interest . . . well without further ado . . . .

Like THIS!

MMI6331 Colour PROM Programming by Hand

First of all, I built a programming board. This is on single-hole board rather than stripboard and I'm using the resistors to stretch between tracks (hence the look of the circuit).

The first 'hack' of this board didn't work as the information I was being presented with was inaccurate (the details from the Datasheet were wrong) . . . but I got it right the second time and the details included here are the ones I succeeded with.

When programming the PROM, I needed a source of 5.5V applied to VCC on the chip so I used an LM317K with appropriate resistors to generate this. I used 5 of the row of 8 DIL switches to set the address (A0-A4) on the right, and gave a gentle 'tap' on the switches on the left to pop each fuse as I programmed each bit (D0-D7).

Like an EPROM, the outputs of this PROM are set as '1' as default from the factory and you have to pop the fuses (tap the buttons) in order to set the '0's.
The datasheet recommended programming pulses of 10uS . . . but I knew that my finger taps would be a fair bit longer than that even being super-quick, so I reduced the programming voltage to a lower level (21V) to compensate a bit.
It worked perfectly. Programming is done with the chip in the disabled state (_CE set high).

The truth table paper shown to the right was taken from the Galaxian file '6l.bpr'. I used a piece of PC software called 'HxD Hex Editor' in order to output the binary values, although the offsets were output in decimal so I had to mentally convert a lot of it. No matter.
The blank spaces where the binary is not shown is a programmed '0' so effectively it will involve pressing all 8 buttons one after another in order to set this.
Of course, one bad finger press or incorrect switch setting and it's 'Game Over' . . . you can't go back . . . no chance of erasing.
I took great care, I managed it all fine!

It didn't take me that long and . . . it worked perfectly! In fact, I programmed another MMI6331 for another board and it worked fine as well.

Programmed PROM inserted on the board - Labelled with a biro on Tippex.

(Please ignore the monitor fold-over to the right of the picture - This is not a board fault)

And there we have it - The right colours!

Last part - SOUND!

Van

Van Diesel2020-09-28 20:45:45

 

[Judder]

That's a brilliantly funny writeup