Repair Log: Joust MPU board

YellowDog

In Memoriam

Donor 2011, 2013
Joined
Apr 12, 2008
Messages
9,851
Reaction score
158
Location
Humble, Texas
I received a Joust MPU board that was pretty messed up. I replaced quite a few chips and the crystal before getting it to work right. Packed it up and shipped it off. A couple days later I got a PM that it wouldn't sync up after it arrived. Which didn't sound right, as I had my usual 4 hour burn in before shipping it. After trying a few things, I asked the owner to ship it back to me.

Sure enough the sync was way out:

attachment.php


This was the rug test. And it passed.

Here is the game running:

attachment.php


I grabbed the logic probe and started backtracking the clock signal. At first I didn't find anything. So I grapped the HP logic comparator and started rechecking all the invertors and flip flops and all of them checked out.

So I started checking the clock pulse from the crystal and when I touched the inner lead, my finger slipped and I shorted the lead to ground. And the picture cleared right up. WTF??? I pushed the reset button, the picture stayed good. I powered down and it was trashed again. When I touched the inner lead to ground using a jumper cable, it cleared up. THinking it was the crystal, I replaced it. Same issues.

After some experimenting I found that if I clipped the jumper cable across R74, it would boot up and run just fine. So I replaced R74. Picture was trashed. Jumpered, ran fine. So I replaced R74 with a piece of wire. Screen was trashed. Long story short, I could touch 3 or 4 places with the jumper cables to ground and it cleared up. Touch it with a short piece of wire, trashed. I checked the jumper cable with my DMM and found there was a small amount of capacitance that wasn't in the short piece of
wire.

Out of desperation I tried using a 22pF capacitor in a couple of places, and lo and behold, it worked. The best place was from pin 2 of the 74LS04 at 7J (the first invertor) to ground. I soldered the capacitor from pin 2 to pin 7 (ground on the chip) and this is what I saw:

attachment.php


(the little bit clipped on the left is normal with the little LCDs and Williams boards). If I swapped an unmodified 74LS04 in, I got the trashed screen.

The only thing I can think is that when I replaced the 74LS04, that the socket added some small fraction of capacitance that causes the first clock signal to degrade. Adding the additional capacitor seems to offset the effect, strengthening the clock signal. It is one of those fixes. Don't know why it works, just that it does.

The fix:
attachment.php


Weird, but it works. Time to pack it up and send it home.

ken
 

Attachments

  • P1012036.JPG
    P1012036.JPG
    67.7 KB · Views: 252
  • P1012033.JPG
    P1012033.JPG
    59 KB · Views: 249
  • P1012034.JPG
    P1012034.JPG
    57.1 KB · Views: 245
  • P1012037.JPG
    P1012037.JPG
    66 KB · Views: 249
Could be because of the particular crystal being used.

Really odd when you get boards that do these types of funky things! It's like the MV4F version of the Neo Geo which will sometimes just go into self oscillation in the audio section. If you pull the 2 .1uf poly caps from the output it stops. When it's oscillating it pulls massive amounts of current on the 12v line which affects not only the 12v line but also the 5v output from the power supply.

Nice fix!

RJ
 
I actually tried 3 different crystals and they all acted exactly the same. Trust me, that was one of the first things I thought of.

I don't know if there was a frequency difference. If there was, the RAM should have failed the RAM test. The problem was strictly in the video output clocks. The logic clocks seemed to be functioning normally. On Williams boards the signal from the single crystal is broken up and divided down into multiple clocks for various uses.

Basically it seemed like the video clocks were not operating correctly, either weak or non-existant.

ken
 
In this case, it was free. The owner had already paid to have the board fixed. While I don't offer a waranty (these are 30 year old boards after all), I will stand by my repairs as long as the boards were not damaged or abused.

And yes, I understand that I am a lousy businessman. But I do this because I want to see these games enjoyed for as long as possible. They can't be enjoyed if the boards don't work.

ken
 
Thanks Ken!

I would've never figured this one out.

The strange thing is when you originally sent the repaired board back to me, it worked somewhat, it just had shakey video - and then the video sync just became worse and worse until it was just like the pics that you've shown. It was the symptoms of something starting to fail. I replaced the crystal thinking that would be it, but just as you had replaced it 3 times, it wasn't! Maybe that board has just so much corrosion that there is some extra capacitance on the board itself??

Well, I'm glad you showed the board who is boss!

You are the Master!
 
I actually tried 3 different crystals and they all acted exactly the same. Trust me, that was one of the first things I thought of.

I don't know if there was a frequency difference. If there was, the RAM should have failed the RAM test. The problem was strictly in the video output clocks. The logic clocks seemed to be functioning normally. On Williams boards the signal from the single crystal is broken up and divided down into multiple clocks for various uses.

Basically it seemed like the video clocks were not operating correctly, either weak or non-existant.

ken
Right I see. I just wondered if you had looked at the different clicks with a frequency counter to see how far out the video clock was.
I hear what you are saying about free repairs, I sold an Asteroids machine a while back that had played fine for years. The guy got it home, plugged it in and it died.... A few IC's in the final stage if the Y vector output had failed. I fixed it for free.
 
Is that acid damage on the copper area around those chips?

No, but it is battery corrosion. Batteries from the late 1970's on typically are alkaline batteries so that technically it is alkalai damage. That is why the remediation for battery leakage corrossion is to soak the board in a weak acid bath (typically white vinager).

ken
 
No, but it is battery corrosion. Batteries from the late 1970's on typically are alkaline batteries so that technically it is alkalai damage. That is why the remediation for battery leakage corrossion is to soak the board in a weak acid bath (typically white vinager).

ken

Do you think that is causing the problem? Not sure what plane it is, but it looks like it has been eaten up pretty well in that section. Maybe the plane has split a bit and isn't strong enough and is the reason you see a bit of capacitance.
 
That's interesting, thanks for sharing. I think your theory is correct - the load capacitance on the oscillator must have been a little 'off' to begin with (maybe due to the trace damage?) and adding that socket pulled it out of frequency range for the video circuit to sync. Nice job figuring out the fix.
 
It's like the MV4F version of the Neo Geo which will sometimes just go into self oscillation in the audio section. If you pull the 2 .1uf poly caps from the output it stops. When it's oscillating it pulls massive amounts of current on the 12v line

someone told me tantalum capacitors dont oscillate at higher frequencies like electrolytic capacitors might

does that have anything to do with it and would you consider replacing with a tantalum or no ?
 
The oscillation stops when 2 poly/mylar caps are removed... Not tantalum or electrolytic.
 
Back
Top Bottom