my new defender, pink background?!

gwarble

Well-known member
Joined
Feb 18, 2011
Messages
2,465
Reaction score
222
Location
Petaluma, California
my new defender, pink background, no "initial tests..."

Hey everyone,

I picked up a new defender yesterday! decent condition, board problems and G07 chassis was pulled, switcher hacked in but pulled also, original ps intact and supposedly worked but was bypassed to diagnose board issues...

anyway, i haven't gotten my nose dirty yet digging into it, but i threw a switcher back in (wiring was nicely labeled!), moved it close to another machine and borrowed the monitor... i get the same result that the guy i got it from told me about, boots into defender like:

- red only rug pattern (x2), not multi colored like my robotron
- bootup sound
- blank screen for 3 or so seconds
- another bootup sound
- four leds blink twice
- defender attract mode with a pink background...

what should be a black background is all pink... obviously in the signal and not some monitor issue... and i get two rug patterns, blank, attract mode... in mame i get like 7 rug patterns, "all systems indicate ok", then attract

again i'm hoping someone knows off the top of their head, i haven't dug into it yet except:
- voltages are within 10%
- quick reseat of chips/connectors
- auto up or down, hitting advance goes to blank screen without any diag (but there is some weird hacked wiring around those switches so they need investigation)


my first guess is rom problem... but maybe someone can narrow it down closer with these details...

thanks, in case you were wondering, $75 delivered
- gwarble
 

Attachments

  • DefenderFail1.jpg
    DefenderFail1.jpg
    98.4 KB · Views: 27
  • DefenderFail2.jpg
    DefenderFail2.jpg
    97.2 KB · Views: 44
  • DefenderFail3.jpg
    DefenderFail3.jpg
    85.5 KB · Views: 20
  • DefenderFail4.jpg
    DefenderFail4.jpg
    97.2 KB · Views: 24
Last edited:
the cab...
 

Attachments

  • DefenderCab1.jpg
    DefenderCab1.jpg
    94 KB · Views: 23
  • DefenderCab2.jpg
    DefenderCab2.jpg
    95.7 KB · Views: 27
  • DefenderCab3.jpg
    DefenderCab3.jpg
    98.7 KB · Views: 22
My first guess would be to test the video output transistors and then working back through the video circuit components.

The other thought would be to check the video cable and to reflow the connectors.

BTW: Nice pickup.

ken
 
thanks for the insight... i'll start checking there

i'm concerned that because of my weird rug pattern/bootup sequence and lack of service mode that there is something else going on... i have the 4 leds but no "all tests indicate..." screen which is weird... does anything in the video circuit affect this stuff?

thanks
- gwarble
 
a little more info to help help... i fell asleep on the couch with pink screen defender running and woke up to a black background defender looking pretty normal.

a power off/on goes back to pink screen, and cpu reset also goes back to pink screen...

and once the background has gone black i'm able to get into the system check via ADVANCE, so thats cool...

time delay changes like this in my limited experience are usually either bad/old caps and or cold/cracker solder connections or bad connectors/wiring... in this case the previous owner reflowed the solder on the cpu and rom boards in what looks like a clean job, i didn't notice anything questionable... i plan to redo it myself to be sure but narrowing down the circuit sure would help... also some insight on the caps on the board


one other problem i noticed randomly after a few of many resets or power cycles: sometimes the monitor is shifted down about 1/4 of the way... so i see the radar at the top, wrapped from the bottom of the screen like a monitor sync issue... i can still see the monitor defects (tear/curl etc) in place at the normal position on the screen, so it looks like its in the data somehow (another reset always clears it, only seen it a handful of times)

thanks again for any help, i saw another thread here with a similar described problem and no solution...

- gwarble
 
Last edited:
Thanks for the look/help...

The switcher is putting out correct voltages (i don't know how picky defender is but all are within 1/4 of a volt or less) and the cpu board is receiving the correct voltages at the connector and at the ram chips... the rom board is fine at the connectors but i didn't actually check the voltage at the ROM chips, i'll do that this eve

switcher should be fine, i can try another one, but the previous owner had the original power supply give the same problems, then he hacked in a switcher to diagnose it... so i can (and plan to eventually) go back to the original power supply which should still work, but i'm leaning towards the voltage not being the problem as i and he have seen the pink screen on the linear PS and two different switchers

thanks
- gwarble
 
Defender is pretty picky about the +5vdc, and most will not run under +4.90vdc.

+4.75vdc will not run in most cases at all.
 
I think it was ( from memory, ill check again)
+5.02, -4.97, 12.07
which sound close enough to me
There was also no change in voltages when it started acting normally after running a while


When normal, test mode works, in color ram test there were no colored bands, but i noticed a few faint single-pixel vertical black lines

Thanks
- gwarble
 
i'm going to say conclusively that the voltages getting to the board are fine...

its getting quicker and quicker and "warming up" into a black background... less than an hour on and its almost always fixed its own pink background...

a reset still usually shows pink, but a few times i get a normal bootup procedure if its been on a while

i reflowed the ROM board, inspected... one ROM had 2 almost broken legs... i ghetto-fixed them, all connections appear good (connectors, chip-board connections, ribbon cable...

no improvement, cpu board is next, i'll pull, reflow, and carefully inspect it

ideas?

thanks
- gwarble
 
thanks for the idea/input, but i'm positive it isn't the monitor... you can kind of tell from the crappy pictures (it looks like a data problem, not a monitor color or sync issue), but since then i've also run it on my appleII test monitor and now i've popped in a WG from another cabinet i just picked up, and its the same result (both of those monitors tested fine on other setups)

also when it "warms up" and the pink goes away, you can see it change that its something in the signal...

an update to spurr any other ideas, it takes less time now to "warm up" and get over its pepto-bismal look, usually within the first 30minutes of being on... and then after its already "warmed up" turning it off and back on it usually (but not always) starts pink but reverts to normal sooner

i've now reflowed the solder on ROM and CPU boards, moved all the RAM chips around to different spots, and reseated anything i could. My next step will be to acquire a ROM or cpu board and swap them out and narrow it down, although i'm surprised with all the pro's on here, and how long this game has been around, that someone doesn't know what a pink screen actually means, so i could tell which chip, capacitor, resistor or transistor is the culprit, or at least which circuit to investigate

from the defender theory manual, it appears that the color ram is filled with data from the ROM on bootup, and then colors are address and the actual color is chosen from the colorRam... so my untrained guess is the data is getting into the color ram but failing upon retrieval until "warmed up"... what that means at the board level i don't know

thanks again
- joel

PS FYI: i inverted the signal with the "cut trace/jumper and video pin 7" trick to use this on the apple monitor which only has neg CSync input, but i re-jumpered this and is back to seperate positive hsync and vsync... in case you were wondering. there's no way its the monitor :)
 
I would be surprised if the ROM board would change it. I have a Joust MPU board I am currently working on that the colors are off. In the past I have changed the color RAM out without fixing the issue.

The one board I did fix the color issues on, I just changed out the 74LS374 (chip 1A) which drives the digital to analog color drivers. I think I changed something else out too, but I don't remember what it was.

ken
 
EDIT: ack, defender doesn't have a chip at 1A (as joust does)! but theres a set of resistors, maybe we're getting close
does anyone know if those resistors/diodes do the digital/analog conversion that 1A would accomplish on later CPU boards?
EDIT AGAIN:
ok that same chip, 74LS394 is used at 1M, 3M, and 4M on the "later" defender according to the books...
=======
awesome, thanks for the info... i'll check out connections to that chip, looks like i have to desolder it out and put in a socket? have a recommended source for that chip? or a recommended way to test it?

also, just so we all remember its not only a color issue... when the background is pink, the "initial tests" and "advance" settings don't happen/work and the bootup rug pattern is light and dark red only, not the multicolored rug we all know

thanks
- gwarble
 
Last edited:
PS, boards connectors, ram on board, also varied voltage from 4.9 to 5.1 on the switcher with same results

Thanks for the help
 
Did you check the Transistors (Q1, Q2, and Q3) like YellowDog suggested?

If you're not sure if they are good, you could swap two of them around and see if your colors change.

If Red is the one giving problems, then you could check IC 2C.

Check your video Decoder Proms (2 and 3) on the CPU board as well. I had a board that had double images and it turns out that two of the legs on the Video Decoder Prom 3 were bent over. Easy to check and eliminate.
 
EDIT: ack, defender doesn't have a chip at 1A (as joust does)! but theres a set of resistors, maybe we're getting close
does anyone know if those resistors/diodes do the digital/analog conversion that 1A would accomplish on later CPU boards?
EDIT AGAIN:
ok that same chip, 74LS394 is used at 1M, 3M, and 4M on the "later" defender according to the books...
=======
awesome, thanks for the info... i'll check out connections to that chip, looks like i have to desolder it out and put in a socket? have a recommended source for that chip? or a recommended way to test it?

also, just so we all remember its not only a color issue... when the background is pink, the "initial tests" and "advance" settings don't happen/work and the bootup rug pattern is light and dark red only, not the multicolored rug we all know

thanks
- gwarble

Sorry, my bad. Defender drives the color values directly out of the color RAM. You can get the transistors 2N4403 from Jameco, Unicorn Electronics, GPE, etc. They are cheap (like $0.10 each) so it is always good to have a few spares on hand (and some 2n4401 transistor also).

The chips at 1M, 3M and 4M are totally different chips. Don't mess with them, they don't have anything to do directly with the color output. They are the buffer chips between the pseudo RAM bus and the microprocessor bus. If you are getting past the rug test, these are OK.

ken
 
thanks, you guys are awesome!

so if the transistors Q1-Q3 were bad, would specific colors be out or problematic? i don't really know how to test them, nor if i can be done on board or they have to be pulled... Visual inspection looked ok, i can swap them around, BUT, the problem i'm seeing isn't really around one color or another, its like the wrong data is being sent for colors... if that makes sense... its not the same symptom as if a color was acting up on the monitor at least

same thing with the decoder chips 2 and 3... i pulled them out, visually look ok, but i didn't swap or check the chips themselves


any recommendations to test the transistors? or an idea which one could be problematic if the black screen is pink/purple and the green aliens are black?

thanks
- gwarble
 
Back
Top Bottom