Bram Stoker’s Dracula Pinball Machine (Williams 1993)

Location: Broomfield, Colorado.
Symptom: Error message indicating several switch rows shorted to ground.

The owner had checked the playfield and the coin door wiring, looking for any obvious shorts.  None were found.  I suspected there was a problem on the MPU board, where all of the columns and rows of the switch matrix connect.

Lower portion of the Williams MPU board (click for larger), U20 is just left of center. Note battery holder above it.

I disconnected all of the switch connectors from the MPU. With the pinball machine in the Switch Edges test routine, I took a couple of jumper leads and a diode and connected a Row 1 with Column 1.  Instead of seeing a single switch closure, the entire Row 1 lit up as being closed.  I repeated with a few of the other rows and got the same result.  This pointed to the column driver chip U20 (ULN2804).

To be sure, I checked the column outputs with my oscilloscope. Instead of seeing a signal pulsing from 12 volts to ground, I saw a signal pulsing from about 2 volts to ground.  I checked that there was 12 volts on the pullup resistors to make sure the PCB wasn’t damaged from leaky RAM batteries.  The 12 volts on the pullup resistor was fine.  I then checked the input to U20 to make sure the upstream chip was functioning correctly.

I replaced U20 with a socket because this is a common problem with these Williams MPU boards.  I then placed a new ULN2804 into the socket.  The board was retested in the machine and everything was fine.

Moulin Rouge, Williams Pinball Machine (1965)

Location: Aspen, Colorado.
Symptoms: A variety of things not working.

The Williams Moulin Rouge is an electro-mechanical (EM) pinball machine from the mid-1960’s.  The owner has owned this game for over 30 years.  However, the game hasn’t been used in over 8 years.

When I first opened the machine, I saw some disconnected wires dangling from the playfield.  These were associated with the left flipper.  I re-soldered these wires to their respective places.

In the backbox, there were a couple of steppers that were sticky and needed to be cleaned.  The ball count stepper and the match stepper.  Even if a pinball machine has been set for free-play, the match stepper is important to have working.  The match stepper is used continuously throughout a game and is triggered by various switches in the playfield.  In this case, it controls various playfield features that alternately illuminate.

The machine wouldn’t power up due to a broken in-line switch that had been placed on the line cord.  I replaced the line cord, which had been previously spliced and installed a better quality switch.  Originally these games didn’t have power switches.  To turn the game off, you kick the bottom of the machine, which trips a “kick-off” switch.  To turn the game on, you press the left flipper button.  Kicking the bottom of the machine is awkward, so I prefer an in-line cord switch.

Once the machine was powered up, it was in continuous reset mode.  This is not unexpected with an EM game that hasn’t been used for a while.  A game will get stuck in reset mode when something isn’t detected as being reset.  In this case, the score reels were not being reset.  They were sticky and not turning freely.  Also, there was a dirty contact that was preventing the “100” digit to not get the reset pulses.  Once the score reels were fixed up, the game would reset.

The next thing I noticed was that the outhole wouldn’t register a ball after the first ball. Therefore it wouldn’t give an end of ball score, wouldn’t advance the ball count stepper, and therefore wouldn’t eject the ball to the shooter lane.  After checking the schematic to see what else was in that circuit and manually tripping the outhole relay, I noticed the scoring motor wasn’t indexing to its home position after either a trigger of the outhole relay or the eject relay. It turned out to be a couple of dirty contacts on the score motor bank that were supposed to keep the score motor running until it was indexed.

Next up, there were many lamps not working.  This wasn’t because the bulbs were burned out, but because the sockets and the lamp bases had corrosion on them.

Next up, the right slingshot wasn’t working correctly.  As the slingshot arm would kick the ball, it would get hung up on the playfield plastic over it.  The plastic was warped. I placed washers under the plastic to raise it up high enough so it wouldn’t interfere with the slingshot.

I replaced all of the rubber pieces on the playfield.

The left flipper needed a new bushing and was missing its torsional return spring.  The right bushing was okay, but both will be replaced on a future visit.

All in all, this was pretty typical work for a pinball machine of this vintage and the years of non-use.  People always ask me for estimates on this kind of work.  It’s very difficult because you have to start fixing things to see what else doesn’t work.  But it’s almost always in the 3-6 hour range.

Star Pool, Williams Pinball Machine (1974)

Location: Broomfield, Colorado

Symptom: Not powering up, not working.

This pinball machine is located in a youth center in a church basement.  No one seemed to know much about it, other than it was probably donated. So it was unknown when it last worked.

Given the age of the machine, before powering it up, I checked the mechanical operation of all of the steppers and visually checked the contacts on the relays and score motor cams.  Several of the steppers were gummed up, but the rest were in good condition.  I rebuilt the steppers and fixed the left slingshot linkage which was binding.

We had to drill out the lock in the back box because no one had a key and we couldn’t find one in the machine.

Once all of the mechanical parts were working freely, I powered it up.  The machine would get stuck in continuous reset mode.  The reset cycle description in the manual was nearly incomprehensible.  So I was on my own.

I checked all of the obvious things, like the score reels resetting to zero, the ball count stepper unit, player count stepper unit, etc.  Everything was being reset and all reset switches were working.  The only weird thing was the credit stepper was stepping up to the maximum allowed credits during reset, when it should be subtracting 1 credit.

It occurred to me to check the coin switches, which I hadn’t checked earlier.  They were a mess.  I think someone thought they could get freeplay mode by bending all of the contacts together on both coin inputs.  Once I got the switch wires and contacts untangled, the machine would reset properly.

I bent the switch contacts on the credit unit to give free plays.

After that, there was still some minor tuning needed and got 98% of it working well.  I discovered a broken switch contact on the Spinner Advance Stepper (EOS). It was broken right at the insulator which prevented me from soldering another contact on.  I am currently researching a replacement switch or contact.

I will update when this is done.

Red & Teds Road Show, Williams Pinball Machine (1994)

Location: Broomfield, CO

Symptoms: Delivers too many balls to the shooter lane.

In this era of pinball machine, the ball trough is monitored by infra-red emitters (LEDs) and detectors (photo transistors).  The detectors are on one side of the trough, the emitters on the other side of the trough.  The game senses the ball when it breaks the beam of light between the two.

The owner had suspected faulty opto boards and had replaced them to no avail.

The test diagnostic (“switch edges”) showed that the Trough Jam opto was not working.  It said there was a trough jam all of the time, even when no balls were present.  This caused the firmware to think the ball was jammed and to try shooting it again into the shooter lane.

Switch Matrix (click for larger and clearer)

Even though the owner had replaced the opto boards, I wanted to start at the beginning and make sure it was working. I checked the signal at the collector of the photo transistor while the owner blocked the light in the trough.  When the light was allowed to hit the detector, the collector measured 12 volts (or close to it).   When the light was blocked, it measured near 0 volts.  This is correct.

Photo-transistor schematic; the red dot indicating the measurement point.

At this point, I thought it would be a good idea to check the other switches in the same row and column as the “Trough Jam” to see if there was a wiring or MPU problem in the switch matrix.  Several other switches in the row did not work.

We spent some time tracing the row wire through the various bundles, not an easy task when 40-50 wires are tie-wrapped together.  The row and column wires are daisy-chained from one switch to the next, zig-zagging across the playfield. We found it broken at the White Standup switch.

I re-soldered the wires to the switch and everything worked.

While the switch edges diagnostic was running, we discovered and unrelated intermittent switch.  We found it easily at one of the eddy current sensors because the LED on the sensor board would blink whenever the connection was lost.  One of the wires was pulling out of the connector.  I tie-wrapped the top of the connector to act as a strain relief and hopefully it will hold.

Police Force, Williams Pinball Machine

Location: Littleton, Colorado

Symptoms:  Weak flipper, lower playfield general illumination not working.

I started with the flipper.  I checked the mechanical aspects of the flipper by manually turning it. Then I checked the end-play by lifting it up and down, and tried to rock it back and forth to check whether the flipper bushing should be replaced.  Mechanically, everything seemed fine.

Next I turned to the electrical aspects of the flipper.  Flippers of this type have two coils wrapped around the solenoid coil bobbin.  One coil provides a strong magnetic field for starting the movement of plunger (slug) into the solenoid.  The other coil provides a weaker magnetic field for holding the plunger once it has been drawn inside the bobbin.  The force required to hold the plunger is less than what it takes to draw it in.

When the flipper button is pressed, both coils are initially energized.  Once the flipper is at it’s ending position, it opens a switch that turns off the more powerful coil and leaves the weaker coil energized.  This allows the flippers to be strong, but prevents them from burning up if someone were to hold the flipper button.  If power were applied constantly to the more powerful coil, it would overheat and start burning.

The switch that opens when the flipper reaches its end position is called the End of Stroke switch (EOS).  I check the EOS switch with the power off and the flipper in its normal resting position with my ohm-meter, and it was “open”, when it should have registered a short.  I cleaned the contacts with some 440 grit sandpaper, and adjusted the contacts to make sure they were closing when the flipper was at rest.

Drawing of flipper unit from the Police Force manual, showing the End of Stroke switch.

I rechecked the switch contacts with the ohm-meter and they were working properly.  We powered-up the machine to verify the flipper was working fine.  Basically, since the switch wasn’t closing, the flipper was only running with the weaker coil.

Next I took a look at why all of the General Illumination (GI) lights in the bottom half of the playfield were not working.  The GI lights are the ones that stay illuminated all of the time and light up the plastics and the playfield areas so that you can see the ball.

First I checked to make sure no voltage was present at one of the light sockets.  Although it was a remote chance, I wanted to be sure that it wasn’t just a case of the lights being burned out (which could have happened if there was a short with a higher voltage somewhere else in the machine).  There was no voltage (be sure to measure on AC scale).  I checked where the power was coming from the Backbox Interconnect board, and there was 6.3 volts there.  So somewhere between the backbox and the playfield, the voltage was getting lost.

I tipped the playfield all of the way up and found the GI Relay on a small circuit board at the very bottom edge.  The connectors that connect the wires to the board were burnt and falling apart.  The connector on the circuit board was burnt as well.

Connector failure is a common problem with GI lighting in pinball machines. Oxidation builds up on the connector pin, or the crimp, and causes the resistance of the connection to go up.  This causes the connector to heat up until it fails. Because each machine and manufacturer uses different types and sizes of connectors, it’s difficult to to keep connectors on hand.  In this case, I didn’t have replacements with me.

The owner didn’t want to wait to get the proper connectors and asked that I solder the wires directly to the circuit board.  I did so, and the lights all worked fine after that.  Normally I am a purist about having the connectors fixed properly, but in this case the connectors only served to make the board replacement easier, which is unlikely to ever need replacing.

GI lamp relay board.

We checked and replaced several burned out lights.  The game was working well and in good condition.

 

Judge Dredd follow up

Location: Denver, Colorado.

A few months later, I was called back to the Judge Dredd machine mentioned in this previous post.

The thermistor in the power line module had burned out.  I replaced it and tested the machine.  I noticed that the shorted switch-row problem had returned.  The nice thing was, it was a solid failure this time and I was able to track it down.

In my previous post, I mentioned the short would go away if I unplugged J212.  This was the case again, and it deceived me into thinking the short was somewhere along this stretch of wires that heads towards the coin door.   In the switch table in the service manual, Switch 24 is listed as Always Closed.

(The switches and lamps in a pinball machine of this vintage are multiplexed in a grid.  This reduces the complexity of circuitry and wiring.  Rather than having 64 separate wires to 64 switches, by multiplexing, 64 switches can be monitored with just 16 wires.)

Somewhere in the machine, there is a diode connecting Column 2 and Row 4.  It turns out it’s on Coin Door Interface Board, which is were the wires from J212 end up (D2).

What is deceiving is that when a column is shorted to ground, any switch that is closed will show that row to be shorted.  Since switch 24 bridges column 2 and row 4, every switch in row 4 was shown as closed, leading me to believe the problem was in row 4.  But when I closed other switches in column 2, the entire corresponding row would show as closed.  And as a verification, I open the diode at D2 and the remaining switches in row 4 began functioning normally.

Once I determined the problem was actually in column 2, and the symptom indicated it was shorted to ground, I unplugged J205 and J212 to isolate the CPU board from the rest of the machine.  I measured the outputs of each column driver and found the output of U20, pin 17 was stuck at ground.

I replaced the ULN2803 with a new part and everything in the pinball machine tested normal.