Donkey Kong repair

hindered

Well-known member

Donor 2015
Joined
Jun 4, 2009
Messages
1,796
Reaction score
32
Location
Kalamazoo, Michigan
So, I have a DK CPU board that I hooked up to a known-working DK Vid board. Everything seems to be working fine, except the monitor is super washed out. I have 4 other Nintendo boards that all display fine on this monitor, but this board it looks as if the brightness has been turned all the way up. Adjusting the brightness knob on the monitor does not help.

Some background, this board was converted to DK3 and had a bunch of hacks done to it. Three resistors on the CPU board were cut (68ohm) and a bunch of resistors were soldered in various places (of course I did not document where). I've removed all the added resistors, and replaced the 68ohm resistors.

What's the next step? I have a working DK CPU and the schematics, do I just need to go component by component and make sure there weren't other hacks done that I missed?
 
Ok, I just compared this board to my working CPU board and all the ICs match up properly. So, I decided to start metering resistors. I get to the 3 68 ohm ones I replaced, turns out they're 68K ohm because I am bad at ordering things on the internet. Whoops! Would that explain the symptoms I described?
 
68...... 68K..... What's the difference? Only three zeros. lol

I know, right? I feel pretty stupid.

Does it make sense that this would make the video signal super weak, assuming the 3 resistors have to do with video output? I can post the board locations later when I get a minute.
 
Swapped out the 68k for 68, video works fine now.

There were no digital sounds, checked the 10k pot, it's fine. DAC and 7k were both new. Turns out the board had the wrong socketed chip at I think 8H... I swapped it out with one from a parts board of the correct model and now all sounds are back. Hooked it up to my working video board and had a 100% functional pcb. :)

Next up, the video board. :) I think it's just an issue of replacing the 2148 ram.
 
Replaced the missing 2148 ram on the video board, and fired it up. Looks kind of like the attached pictures (I'll try to take real pictures soon, don't have access to my camera right now) taken from http://www.brasington.org/arcade/tech/dk/. Problem is, those are bad 5B and 5C, which are on the CPU board, which is known working.

So, I dug further and saw that the 2114 ram at 2p and 2r on the video board might be the issue, and I happen to have about a brazillian of those laying around, so I took out the originals, socketed them and put in new ones. No change. Put the old ones back in, no change.

Then I noticed that there was a socketed chip at 1E, which is not normally socketed. This made me think it was part of the DK3 hack, so I took a 1E (mb10136 IIRC) off a parts board and put it in -- no change.

I'm officially out of ideas at this point -- any suggestions?

EDIT: Not sure if this is at all related, but the other symptom of this board is that there is no "jump" sound at boot -- all the in game sounds work 100% however.
 

Attachments

  • bad5b-10.gif
    bad5b-10.gif
    2.9 KB · Views: 23
  • bad5c-5.gif
    bad5c-5.gif
    5.3 KB · Views: 19
Last edited:
Bump again. Just picked up a logic probe. How can I use this to test the vid board?

EDIT: pulled out the multimeter and checked some suspect traces. Found a bad trace and ran some wire to fix it. Also replaced a few cracked 1.0uf caps. Board's behaving differently now but still broken. It actually seems a bit worse now. :)
 
Last edited:
Here's some pics finally. DK, the lady, and jumpman all seem to be OK. The playfield is trashed/garbled. Keep in mind this is a known working/repaired CPU board. All issues are related to the video board. Again, these symptoms look a lot like the bad 5C on Braze's page, but the CPU board is known working. I'm assuming there's components on the video board that feed the CPU board (or vice versa) and I can track down what on the vid board (if anything) affects 5C on the CPU.

Guess I've got these schematics, and this logic probe, and can go from there.. any guidance would be appreciated. I'm a bit out of my depth but willing to learn.

Edit: These pics were taken after removing the jumper which repaired the trace outlined in the previous post.
 

Attachments

  • coined_up.jpg
    coined_up.jpg
    88.9 KB · Views: 9
  • DK_climbing.jpg
    DK_climbing.jpg
    96.2 KB · Views: 11
Last edited:
And.. "progress". Sort of. Here's some pics after re-jumpering the broken trace from 4S to 3S. Playfield works in columns, garbled in other columns.. No more sprites though -- DK is not visible while climbing during game start, and no jumpman/lady/barrels/fireballs/etc either.

Notice the HSS Kit boot screen.. one of the tests is "BAD" but I can't tell which.
 

Attachments

  • dk climbing.jpg
    dk climbing.jpg
    98.9 KB · Views: 5
  • high score screen.jpg
    high score screen.jpg
    94.2 KB · Views: 5
  • HSS test.jpg
    HSS test.jpg
    96.7 KB · Views: 5
More progress.

First, character sprites are back. Must have been a bad connection to the ram in 6p or 6r, or a poor ribbon cable connection... They came back on their own after doing some reseating.

Based on the screenshot of the HSS kit here http://www.brasington.org/arcade/hs/dk/boot.shtml it appears that the BAD is one of the video ram. I swapped 2p and 2R, and 6P and 6R, and the same ram still tests BAD. So, I figure it's a bad trace or a bad component feeding the "bad" ram.

So, I get out the logic probe and pretend to know how to use it. While poking around, I found that if I short out the two pins in the upper left hand corner of 2P or 2R, the screen changes dramatically (the "garbage" columns as pictured in the previous post are replaced with properly rendered, but incorrect, columns. see pictures). So, I figure something that feeds 2P or 2R is bad. All direct traces into the RAM checks out fine. I looked at the schematics but don't really know how to proceed.

Anyone?

Edit: Looking at the 2114 datasheet here http://www.arcadechips.com/download/2114.pdf indicates that by shorting those pins, I'm applying Vcc (which I think should be 5v) to Address input 7. Does this mean that for whatever reason, current, or the incorrect current, is not making its way to A7 on the 2114 ram?

Double Edit: Sprites are gone again. Iunno.

Triple Edit: Swapped out the ram for a couple of 8114s. Now instead of garbage columns, I'm getting columns that are all 0s. Not sure if that's an improvement or a step backwards. I'm assuming a step backwards, because with the other ram, it was at least pulling out bits of information, just getting them in the wrong place. These ram are getting all 0s.
 

Attachments

  • shorted_pins.jpg
    shorted_pins.jpg
    96.2 KB · Views: 16
  • shorted_pins2.jpg
    shorted_pins2.jpg
    92.5 KB · Views: 11
Last edited:
Back
Top Bottom