Monday, March 29, 2010

Good news, bad news?

I feel a great victory has been won, but at the same time, I'm worried that things may be worse than I thought.  We left off at my display.  Player three had been successfully replaced with a $10 piece taken from a junk pile.  All that was left were figuring out the "A" and "J" segments of my alphanumeric displays.

I read the guide on checking segment problems on a system 11 machine.  I was held up because I had neither a logic probe nor an oscilloscope to diagnose problems.  Luckily, I was able to borrow an oscilloscope.  Also, I called my friend Eric M. who has experience setting them up.  He was an enthusiastic partner in getting these display problems resolved.

We were both pretty certain that if we found the solution for one segment, it would be the same for the other.  Boy were we wrong.  On with the story:

First, I determined that the display board for my F-14 had come out of some other Williams pinball machine from that era.  However, I wasn't able to figure out which one.  I checked the schematics on for other machines from that era but they were too hard to read to be certain.  Bad scans?  Bad source materials?  Can you read this?

I couldn't.  I figured that since you could take a display board from almost any system 11 pinball machine and swap it out, even if the board schematics were different, the connector pins had to be the same.  (J1) was the common denominator between all of the boards for the "A" and "J".  So we set out to identify and document the paths for the "A" segment (J1-19) and the "J" segment (J1-12).  Then, once that was done, we would start to test each link along the way.

This is me figuring out where each line started and ended.  Eric took notes.  I stopped documenting at the "IN" side of the 7180 chip.  This is important to know for later.

We took notes and then Eric set up the scope.  Here are Eric's disembodied hands setting up the o-scope:

Because there are two different voltages, we had to adjust the scope twice.  Everything up to the 7180 chip was low voltage, 12v I think.  Everything after the 7180 was -100volts.  At first we were a little worried about the scope handling that voltage.  Eric had never used it for more than 5 volts.  We read the manual and our fears were laid to rest, we were OK with -100 v.

First we tested known good segments so we knew how the wave should look.  This is what we saw on good segments.

This was a known good pulse.  So we expected to see something like this on the "A" segment:

 But we didn't!  We got a good pulse all the way through the 7180 "out" pins on the "A" segment!  That was a relief.  I thought I was going to have to spend $20 replacing the hard to find 7180 chip.  Since the chip was fine and the resistors after the chip tested within spec, I must have a bad trace between the out of the 7180 and the glass display.

We figured we would get to that during the "repair" phase.  Next we set our sights on the "J" segment.  We were hoping to find that it was the same thing -- a bad trace.

See our careful notes, well, we knew what they all meant at the time.

We tested the "IN" for the "J" segment on the 7180 and the signal was bad.  It wasn't gone, it was just really really weak.  That was a relief because we figured that the problem must be with the cheaper 4050 chip. 

We checked the out (Pin 6) on the 4050 and the signal was weak.  That must lock it right?  Wrong.  We checked the "IN" on the 4050 (Pin 7), weak.  Ah, so then it must be a bad trace from the J1 connector to the 4050!

I buzzed out pin 12 to pin 6 and there was no continuity!  But then we noticed that I was checking the wrong pins.  It gets confusing when you're flipping a board over back and forth.  I checked again and the trace was OK.

Then we thought it was a bad ribbon cable!  That can happen!  It's the next thing in line.  We tested it and, nope, that was OK too.  We kept testing signals until we were completely off the display board and on the actual "J" segment pin of U41 on the main board.

The weak signal is coming off pin 3 of U41, the 6821 chip on the system board!! 

Disaster!  That chip is huge!  How much is it?!!?

It turns out that it costs about $6 - $7!  What a relief!  I'll order one tomorrow with a new socket!  Hopefully that's the end of that problem!

We then returned our focus to the "A" segment.  We checked every trace and found, with a Digital MultiMeter and a magnifying glass, that the very last trace before display pin 1 was broken.  It was very hard to see.  We couldn't get it on camera.

I took tweezers and pulled the trace off the board.  The trace had lifted from the board and I didn't want it shorting on anything.  Here is the trace pealed away from the board:

With Eric manning the camera and being a third hand, I soldered in a jumper.

Completed work:

And does the "A" segment work?  You bet it does!

I'm down to fixing just one more segment, and we're fairly certain that U41 - a 6821 chip - is the problem.  I'll update you when I get the new chip in!

To conclude.  I said it was important that I stopped checking traces at the 7180 chip when I was tracing out each leg of the circuit.  I should have gone all of the way to the display.  Next time I'll check each trace segment end to end.  Had I done that in the first place, I wouldn't have needed all of the fancy equipment to fix the "A" segment.

We also guessed wrong that the "A" and "J" problem were the same.  They actually couldn't have been further from each other.  One was a bad trace at the display end of the circuit and one was a bad chip all the way back on the system board.  I'm glad we tested each separately otherwise we would have wasted more time chasing an assumption.

We both got a real high from tracking down these problems. 

Is it wrong to hope something is wrong with a machine so you get to fix it?

No comments:

Post a Comment