What fixed your game today / general repair log

All of the ROMs went in and out of the sockets real hard (so to speak) Chip legs were dirty and the sockets were the Atari gold pincher ones. They work when the board is in pristine shape – but this one was dirty and the chips were not clean..

I hate those sockets so much. At least the shitty midway ones are easy to pull off and desolder, but the Atari ones suck especially with no solder mask on the boards. There's always damage underneath them from folks getting a little too eager with screwdrivers when pulling eproms.
 
The little gold pincher ones? I find them the easiest to remove. I pop the covers off exposing the row of pins with the paper strip, then I get my small pointy cutters under it and snip off the row as flush as I can get to the PCB.. After that - desolder gun from the back and 90% of the time it clears the via and sucks the remaining bit of pin out.. The trick is to cut flush and not leave enough width of the pin that it can't pull through.. I'm not a fan of replacing 'all' the sockets.. but this board needed it..
 
The little gold pincher ones? I find them the easiest to remove. I pop the covers off exposing the row of pins with the paper strip, then I get my small pointy cutters under it and snip off the row as flush as I can get to the PCB.. After that - desolder gun from the back and 90% of the time it clears the via and sucks the remaining bit of pin out.. The trick is to cut flush and not leave enough width of the pin that it can't pull through.. I'm not a fan of replacing 'all' the sockets.. but this board needed it..
You suck the small remaining pin into your solder sucker??
 
You suck the small remaining pin into your solder sucker??
Yup.. pretty much.. sometimes they just fall out and sometimes the go right into the gun with the solder.. no big deal. The gold pincher ones are the only ones this method really works for me.. The legs are very small and pull through most of the time.. Other sockets the legs are too wide on the parts side and you can't cut them close enough to the PCB to matter.. for the most part.. I think I had all of the MC sockets off and cleared in under 20 min.. (which is fast IMHO - I hate replacing all the sockets)
 
The little gold pincher ones? I find them the easiest to remove. I pop the covers off exposing the row of pins with the paper strip, then I get my small pointy cutters under it and snip off the row as flush as I can get to the PCB.. After that - desolder gun from the back and 90% of the time it clears the via and sucks the remaining bit of pin out.. The trick is to cut flush and not leave enough width of the pin that it can't pull through.. I'm not a fan of replacing 'all' the sockets.. but this board needed it..

I just melt from the back side while pulling on the paper strip and they come out cleanly without cutting and potentially nicking traces.
 
A 6100 90 deg tube swap.

Needed to get a MH (tempest conversion) running. The 6100 tube was completely toast. Red/Green guns were 0 and Blue was faint. I honestly think the blue was just someone turned up the HV to see the lines because the Maze is not blue. Plan was to fully test this 90 deg swap theory. So pulled the monitor and started.

I pulled out the 100deg tube and removed the yoke and rings. And I took the neck board over to the table to change out the female connector. I pulled a CR-23 connector from a defunct 19K4600 chassis that was available and put that in it's place.

1734698792122.png

The connector was a match for the board with the exception of PIN 4 on the socket. PINs 3/13 are not used, and are a blank spot on the connector. There is no hole on the board for PIN 4. So I bent the pin under the socket and installed it.

1734699568156.png

Dropping the tube in and putting the yoke/rings on and reconnecting everything was standard issue. I didn't bother trying to set their position other than putting the rings in the correct place on the neck. Connected it up to the game and turned it on and was greeted with an image where the Z axis appeared to be shorted on. Since the game had been running with a blue image I thought maybe the board had an issue. So I dragged the monitor over to a nearby Black Widow. Same deal. Yanking out the connector on the bottom of the neck board (color inputs) had no effect and proved that something was going on with the monitor. Didn't have enough time to deal with this where I was so into the car and home it went.

At home I was able to reproduce the problem on my Star Wars and so I started debugging. I pulled another CR-23 connector from my stash and swapped it out on one of my known working neck boards. Tried this configuration with the same results.

At this point I went back to the original neck board and set the unit up in a configuration where I could start poking around. I looked at the following
  • Heater voltage was correct,
  • I could disconnect either the Brown or Black single wire from the deflection board and the image would go away (expected of course)
  • Tried a different deflection board as a sanity check on heater voltage
  • Looked through the color gun circuits on the board thinking maybe I've been shorting something
  • Validated HV (was 1k low) but correcting this made no difference
Ok, I've exhausted the little stuff so time to back out from the problem. The tube I selected for the swap was sitting in the shop from a previous "repair" on a Cheeky Mouse game. (flyback died and no reproductions available). So I was assuming it worked. I've got a tube in my Star Wars that I know works so I proceeded to pull apart my SW while continuously swearing because this is the 4th time I've done this in the past month.

Out goes Cheeky Mouse burn, in goes Tapper burn. Power up and same Z axis maxed out image. Ah, so what' the hell, it's clearly something with this setup. Decided to start comparing tubes and schematics.
  • According to Tubular both of the 90 deg tubes share the same pinout/heater/voltage specs but the Anode was 100v diff (negligible)
  • Comparing to the Schematics I found issues....
1734702005859.png1734702073225.png
  • Looking up the OEM tube I found it has the same pinouts as the 90 deg tubes? Ok WTF.?
    • Turns out it's errors in the Schematic for Printing 2 of the 6100 manual. Printing 3 is corrected: [edit: most of the pins were incorrect in printing 2, I just highlighted 2]
1734702372029.png
  • Ok, so that means that the tubes are compatible and something else is going on. But what?
  • Well the only thing sitting between the Tubes and the Chassis is the connector. Maybe it's not a direct match?? Placing the 2 types of connectors side by side you'll note that pins 4 (we saw that above) and 12 are NOT populated in the original CR-24 connector but they are populated in the CR-23!
1734704236647.png1734704458510.png
  • If we look at PIN 12 we note that it's connected to the GRY wire, which is GND. Obviously PIN 12 is NOT supposed to be used at all on a 6100 so I pulled the connector off the board, folded PIN 12 under (like I did for PIN 4) and reinstalled.
1734704822969.png
  • After the above change I connected up the monitor and it works!!!!
1734704952374.png

So a successful 90 deg tube swap into a 6100.. You can read more about the fitment and results after this debugging session in the linked thread.
 
Last edited:
From the Missile Command above..

One is a failure
Two is a coincidence
Three is a trend...

Decided to run the board overnight a couple more times.
Good thing I did.. Woke up to the board appearing to be watchdogging.. except it wasn't. No resets going to the CPU. In test mode it would partially draw the screen showing RAM good, MAP good and ROM good.. then crash before it drew the rest of the screen .. for the most part.

Looking at the schematics - first thing that popped into my head was interrupts..

image-1.png
More 7474's .......

img_6975.jpgimg_6976.jpg
The 7474@E7 Pin 3 was corrupting the circuit and messing with the interrupts.

I replaced all 5 x 7474's with the matching date codes..
There was a 6th one - different manufacturer - so it stays.
Board works (again)!
 
Recently i resolved a similar issue. Maybe this will help if you havent already been down this road.

Check the verticle bars on the playfield, and the slides the horizontal Telescoping bar mounts to. I recently replaced one verticle bar that was slightly bent. The old and new parts are just bare SS rod (on my model some are chrome) lubed with lithium grease. Grease gets tacky with age and will attract dust and grit. I cleaned the vertical bars with fine steel wool and metal polish, then lubricated accordingly. Cleaned the slides so everything glides smoothly. The Telescoping bar has special requirements per the manual, silicon for it I believe.

Additionally the guide wheels at the top need to be greased as well so does the gearbox nechanism.
 
From the Missile Command above..

One is a failure
Two is a coincidence
Three is a trend...

Decided to run the board overnight a couple more times.
Good thing I did.. Woke up to the board appearing to be watchdogging.. except it wasn't. No resets going to the CPU. In test mode it would partially draw the screen showing RAM good, MAP good and ROM good.. then crash before it drew the rest of the screen .. for the most part.

Looking at the schematics - first thing that popped into my head was interrupts..

View attachment 788554
More 7474's .......

View attachment 788555View attachment 788556
The 7474@E7 Pin 3 was corrupting the circuit and messing with the interrupts.

I replaced all 5 x 7474's with the matching date codes..
There was a 6th one - different manufacturer - so it stays.
Board works (again)!
Long burn ins are a pain but are necessary if you have flaky ICs. I had the same issues with an Outrun board that would pass days' tests then fail with a different fault. It took several long burns but the board was solid after weeding out the bad ICs.

These were designed to be run for very long periods, I'd try and burn it in for several days, you'll know…

p
 
Long burn ins are a pain but are necessary if you have flaky ICs. I had the same issues with an Outrun board that would pass days' tests then fail with a different fault. It took several long burns but the board was solid after weeding out the bad ICs.

These were designed to be run for very long periods, I'd try and burn it in for several days, you'll know…

p
My experience so far.. when a board has sat on a shelf for 15-20 years.. chips die this way.. I have a 3 day process (spread over 6-7 days..)
Run 12 hours day 1
Run an automated power cycler day 2... 30 min on, 60 min off - 16 times
Run 12 hours day 3

If anything fails it may get 1 or 2 more of these.
I have the bench space and they run overnight or during the work day typically.
 
Atari Quantum - Random collapse.

If you moved the deflection board harness, the screen would go back to normal.

It appeared to be the deflection board ground, so that was repinned. Problem persisted.

Re-flowed the board connector. Problem persisted.

Replaced the board connector and field connector. Problem persisted.

Noticed that one of the two large resistors appeared to be loose. Reflowed both of these large resistors, and game is working but now the image is too small.

Removed the 10K trim pot from the small resistor, now image is too large.

Next action is to use resistance decade box and find the right resistance, and then install the correct resistance network to get the image back on the screen.
 
Last edited:
Q-Bert running the multi-Q board - no sound.

Examined sound board - a 2 amp slow blow fuse was blown. The fuse was to the 30 volt supply on the sound board.

Replaced the fuse. Sound works.

It was THE FUSE.
 
Game PCB: Neo Geo MVS 2 Slot

Symptom: Left & Right speaker low sound

Steps:
-Tested headphone jacks, worked with no issue
- Validated JAMMA connector wiring for left and right speakers (different pin out for MVS system)
- Validated Left and Right speaker wiring
- Validated speaker pot 1K ohms volume
- Performed Audio circuit cap kit replacement, did not fix issue
- Suspected audio output Power AMP Chip from schematics (HA13001)
- Verified 12V at pin 10 Vcc input
- Observed different op amp impedances than in component data sheet (pins 3&12, 4&7)
- Observed high 8M ohm readings for ASO protection circuit (pins 12&6, 7&6) **not sure if this is correct way to read it, though a new part was 1M ohms instead**

Repair: Replaced Power Amp chip, sound issue resolved with sound back to 100% loud as hell.
1000024757.jpg

1000024760.jpg1000024753.jpg
 
Last edited:
Tempest - text blurry and colors on the edge were off.

Adjusted the focus pot on the flyback which resolved this.
 
Fixed my punchout board. The symptoms were flickering background and slightly miscolored text on the title screen. The text had some white spots.
PXL_20250110_131120216.MP.jpg

With it being background graphics issues, I immediately suspected the bak board. Since little mack and the opponents looked fine, I figured that everything up to that point was fine. Background graphics are controlled by 4A and 4B for the lower monitor. Both roms tested good on my eprom programmer. I really didn't feel like probing the board so I used a 28C64 to see if that had any change. When I replaced it in 4B, that resolved my issues. Solution: flakey 4B 2764
 
Fixed my punchout board. The symptoms were flickering background and slightly miscolored text on the title screen. The text had some white spots.


With it being background graphics issues, I immediately suspected the bak board. Since little mack and the opponents looked fine, I figured that everything up to that point was fine. Background graphics are controlled by 4A and 4B for the lower monitor. Both roms tested good on my eprom programmer. I really didn't feel like probing the board so I used a 28C64 to see if that had any change. When I replaced it in 4B, that resolved my issues. Solution: flakey 4B 2764
Is tested good in eprom mean verified at +/-5 or 10% vcc? I have seen similar on vs. board where reprogramming (not erasing) solved the issue and allowed proper verification.
 
Is tested good in eprom mean verified at +/-5 or 10% vcc? I have seen similar on vs. board where reprogramming (not erasing) solved the issue and allowed proper verification.
What I meant is that when I tested the eprom with the gq, it tested as having the punchout rom on it. I tried rewriting it but no dice. I can't test within tolerance because I don't think the cheaper modern programmers can do that.
 
Not today, but this week. Walked by Centipede and it looked like this.
1736624632997.png
Found this thread, had to wait on new LS153 because out of the 1000's of IC's in the decades old office I inherited, I was out.
In the short 9 months I've been doing this job I have learned to solder, test boards, and replace caps and resistors. This was my first shot at working on something this old. Took my time getting the old chip off as I had limited tools for the job. In the end the copper braid got the job done. New part installed in less than a minute and she back up and running.

It seems one of the last techs at my arcade made off with most of the good tools, so I'm slowly building back up. Getting an LP560 per recommendations from here and the TPG2 so I can start wading through the dozens of CRT's crowding my work space. A lifetime of working on cars, putting together computers, and fixing the fight sticks I broke playing Street Fighter seemed to have been a good background for this job. Still hard to believe I get to putz around with video games all day and get paid for it.
 
Back
Top Bottom