Defender (Later Series) Repeating Vertical Lines.

Scucci

Well-known member
Joined
Feb 6, 2003
Messages
4,497
Reaction score
94
Location
Nashville, Tennessee
Defender wasn't running, fixed power supply and decoders. Ran for a day, came back to it the next day and it was down. Found out a 7474 wasn't processing the clock right, got it back up. Came back to it a couple of days later and now I have vertical lines that change colors with the Defender logo/mutants in game.

No RAM errors, no ROM errors, good Decoders, good ribbon cable... Just really starting to reach the end of what I know with this game.

Shorted pin 23 on the decoder to flip the screen for player 2, lines were still there. People insisted it was the 165s, I ran the board on the bench and couldn't see anything obvious, socketed and replaced them all anyways. So, it's not the 165s. With the 165s socketed I tried it with all different combinations to see how the issue looked with certain ones removed... I noticed that the pixels were the lines are were black during the rug... and AWLAYS present, didn't matter which of the shift registers were removed, the lines were always there.

I SHOULD be getting some new headers in the mail tomorrow so I can check again with my decoder boards, but the lines were there with the original decoders and my replacements.
 

Attachments

  • 476787795_10227264339856523_4615066613711197456_n.jpg
    476787795_10227264339856523_4615066613711197456_n.jpg
    458 KB · Views: 6
  • 476798812_10227264339936525_3160672513491400850_n.jpg
    476798812_10227264339936525_3160672513491400850_n.jpg
    409.1 KB · Views: 6
  • 476833045_10227264339336510_6791762525407937410_n.jpg
    476833045_10227264339336510_6791762525407937410_n.jpg
    576.4 KB · Views: 6
  • 477033693_10227264339696519_5375638706481994327_n.jpg
    477033693_10227264339696519_5375638706481994327_n.jpg
    597 KB · Views: 7
Possibly a cut trace between the 4116 outputs and the '165 inputs.
The architecture on these boards basically divides them into 2 sections by synchronizing and sharing the RAM every-other clock cycle

The first is the CPU side: Here, the CPU is generating the full screen image and writing it into the RAM. There are diagnostics via the RAM test program and watchdog to ensure that it's working properly, and it is. I would confidently say that the screen image sitting in RAM is clean (without the line)

The second side is where the stored data in the RAM is written out to the screen. The CPU has no visibility into this circuitry and for the most part (exception is the Video Address lines (8-13) passing in for the watchdog - see image below), has to assume it's working properly.

The 165's would be my first suspect as mentioned above. By doing the 'Screen-Flip' you are testing the mirrored bank of 165's, but not impossible both are bad (unlikely). That would lead me to the 257's where they are flipped through and common to both banks.

If you swapped the Decoder PROM @ 3E, I would probe all I/O's and ensure they are correct.

1740488947111.png

~Brad
 
The 165's would be my first suspect as mentioned above. By doing the 'Screen-Flip' you are testing the mirrored bank of 165's, but not impossible both are bad (unlikely). That would lead me to the 257's where they are flipped through and common to both banks.

Nope. The vertical lines are every 6 pixels, which is the 'length' of the '165 shift registers.

A failure on the mux wouldn't be cyclical like that.
 
Nope. The vertical lines are every 6 pixels, which is the 'length' of the '165 shift registers.

A failure on the mux wouldn't be cyclical like that.
Yes, it does look like 6 pixels; would have to have error prior to the 257's.
 
Do the lines move when you flip the screen?
Possibly a cut trace between the 4116 outputs and the '165 inputs.

Look for a floating input on one if the 165s.

The lines don't move when I flip the screen. The graphics are messed up because I was flipping it like crazy and taking pictures trying to get P1 and P2 shots... Unless they're not supposed to mess up, but I figured it was just because they weren't being refreshed as I was flipping the screen. I tried to get 2 shots with the same line color, but I couldn't.

No floating inputs on any of the 165s.
 

Attachments

  • IMG_20250225_163830265_HDR.jpg
    IMG_20250225_163830265_HDR.jpg
    513.2 KB · Views: 3
  • IMG_20250225_163833173_HDR.jpg
    IMG_20250225_163833173_HDR.jpg
    571.4 KB · Views: 3
Last edited:
MOTHERF******************!!!!!!!!!

I figured it out. Well, this guy figured it out (https://forums.arcade-museum.com/threads/defender-video-problem-pulsing-vertical-lines.242460/). The 74LS04 @ 7P (which I haven't touched until now... and was already socketed... spoilers) didn't have any markings on it other that "HD74LS04" but no makers mark... sooo... quality stuff. I swapped it out with a TI 74LS04 and, what do you know... no lines. I guess the little bugger was just barely hanging in there and the cascades of failures pushed the little f'r over the edge and the timing was good enough to run, but not good enough to keep the graphics lined up.

Okay, so... sorry about the wasted time. I hadn't seen that thread until a couple of minutes ago. I hope someone in the future finds either this thread or the other thread and saves themselves the headache I have right now.

Hot hammer on IC action in the attachments.
 

Attachments

  • IMG_20250225_171509293_HDR.jpg
    IMG_20250225_171509293_HDR.jpg
    783.7 KB · Views: 11
  • IMG_20250225_171519218.jpg
    IMG_20250225_171519218.jpg
    1.3 MB · Views: 11
  • IMG_20250225_171532644_HDR.jpg
    IMG_20250225_171532644_HDR.jpg
    946.8 KB · Views: 11
Last edited:
That's an odd one.

If the 12M was really jittery due to a failing 7P, the ~6M could be off and messing up the load timing on the '165s... otherwise 7P has almost nothing to do with the video output. Most of the game runs on divided down 4M -- only the video shift registers run off 6MHz.

That fact that flipping the screen didn't move the lines show it wasn't just a broken trace, since that would flip horizontally.
 
That's an odd one.

If the 12M was really jittery due to a failing 7P, the ~6M could be off and messing up the load timing on the '165s... otherwise 7P has almost nothing to do with the video output. Most of the game runs on divided down 4M -- only the video shift registers run off 6MHz.

That fact that flipping the screen didn't move the lines show it wasn't just a broken trace, since that would flip horizontally.

There was a SLIGHT "wiggle" on the clock lines, but I honestly didn't think it was anything to worry about since it went from not even starting to run to being fully playable when I replaced that 7474. I have a couple more Defender sets to look at so I know it won't catch me off guard if I run across it again. This whole experience is burned into my brain now.

Thanks again to everyone for helping, I think I'm pretty solid with checking issues with the 165s now as that seems to be a common issue as well.
 
Back
Top Bottom