Space Invaders Pinball Machine (Bally, 1980)

Location: Near Morrison, CO.
Symptoms: Blows fuse (F4).

The machine would immediate blow the fuse (less than 1 second) upon powering up.  Fuse F4 supplies the 43 volt solenoid circuits which are prone to broken wires, shorted coils, etc.  So, I expected to find a short somewhere on the playfield.   I isolated the power supply by removing all of the connectors except for J2 (cabinet) which supplies the wall AC power to the power supply, but this connector also supplies solenoid power to the knocker and the coin door.  The fuse still blew.  I checked the cabinet wiring to make sure there was no short, then focused on the power supply itself.

It turned out to be a shorted bridge rectifier between the AC and the minus (-).  I brought the power supply back to my shop for repair.  I looked up the part number for the bridge and found it on Marco’s site.  For better or worse, the bridge I received was slightly larger.  Not only were the pins further apart, it was also thicker.

New bridge rectifier (black square) above an original working bridge rectifier just below it.

This posed a problem as there are two bridge rectifiers in the circuit and both are mounted on the bottom side of the PCB and are heat-sinked against a metal bar on the power supply chassis.  If one is thicker than the other, one will be heat-sinked, the other will not.

I decided to shift the bar over, so that existing bridge would contact the bar and the new bridge would go directly to the chassis.

With the metal bar, both rectifiers will contact the chassis when the board is remounted.

I added some new heatsink grease to both bridges.  I powered the circuit and all was fine.  This power supply was only used on two machines, KISS and Space Invaders.

Back at the customer’s location, the game powered up and worked fine.  I replaced a few burned out bulbs.  The NiCad RAM backup battery mounted on the CPU board has partially failed. I am debating whether to replace it with a 1 farad super cap, or the typical 3 cell AA holder.  The super cap requires the machine to be powered up periodically so that it can recharge.  The AA batteries don’t require recharging, but there is no nice place to put the battery pack and they have to replaced every few years.

 

Funhouse Pinball Machine, Williams (1990)

Location:  Greeley, Colorado.
Symptoms:  Many.  Bulbs out, trap door not working, a pop bumper not working, both sling-shots not working, flipper sticking, etc.  Needed general servicing.

The biggest mystery with this machine was that every GI (General Illumination) bulb was burned out, except for 3.  I suspect that at some point in the past, there was a GI short to a solenoid supply which blew most of the bulbs.

As I was replacing some of the controlled lamps, I discovered some that were burning very brightly.  This raised a red flag that one of the rows or columns in the lamp matrix was stuck on.  The easy way to check this is to go into the test menu and run the single lamp test.  If more than one light comes on at a time, then the row or column driver is shorted or bad.   It turned out to be Column 8 and the TIP107 transistor.  All of the Column 8 lamps would light when any of the other columns were on.  After replacing the transistor, the controlled lamps all worked except one.

The one lamp that didn’t work was the “Open Trapdoor W/Flashing” light.  It was lighting when any other column or row was lit. In other words, the only time it acted normal was when any thing in its own column or row was lit. At first I thought it was a shorted diode.  The diode tested OK with the meter.  But the meter doesn’t test the diode at greater than 3 volts.  So I snipped it and the light should have gone dark, but it didn’t.  It turns out, someone had soldered the wires onto the socket incorrectly and bypassed the diode.  Once I corrected the wiring, it worked fine.

As for the sticking flipper, it was a bad link that was catching on the end of the solenoid.  I replaced the links, plungers and rubbers on both lower flippers.  The upper flipper looked good enough to leave alone.

The other solenoid problems were all related to broken wires.

The trap door wasn’t working because the end of the spring broke on the solenoid that latches it up.  I found the spring in the bottom of the machine and made a new hook with my pliers, and reattached it.

The game had an L4 version Game ROM (U6) installed.  The most recent official version is L9, plus there is an L9.05H, which is a home version with some additional features.  The difference between and L4 and L9H requires a move to a larger ROM (1 Mb to 2 Mb), plus one of the sound ROMs (U18) had to get upgraded to L3.

The L9H ROM images I downloaded from ipdb.org didn’t really match the documentation.  The U6 ROM image was a 4 Mb image, and U18 was a 2 Mb image.  In this Funhouse machine, both of these parts were 1 Mb so I couldn’t erase and re-program.  I was out of 2 Mb EPROMs.  So I ordered some blank 2 Mb parts from Marco and received 4 Mb parts instead. Ugh!  I looked at the schematic for the sound board and saw that it would support a 4 Mb ROM.  So, I programmed both images into 4 Mb EPROMs.  This works fine for U6, but for U18 it doesn’t work without loading the code into the upper half of the EPROM to make it look like a 2 Mb part.  Although the sound board hardware supports the 4 Mb ROMs, the firmware in the other sound ROMs doesn’t.

A note on CPU jumpers for larger ROMs:  I think by default, the early WPC games had the W2 jumper installed.  W2 supports the 28 pin ROM families and will still work with a 1 Mb 32-pin ROM because the associated pin on the 1 Mb ROM is not used.  But anything larger (2 Mb or greater) will need the W2 jumper removed and the W1 jumper installed.  The W1 jumper supports all of the 32-pin ROM families.  So when using a 1 Mb ROM, the jumper can be in either position.

The same is true for the sound board with regards to U18.  Jumper W4 supports the 28-pin families and jumper W3 supports the 32-pin families, with the 1 Mb ROMs not caring which position the jumper is in.  But keep in mind that at least for Funhouse, the sound system firmware doesn’t support anything larger than a 2 Mb ROM in U18.

After correcting dozens of small problems, the pinball machine is looking and working great!

Wurlitzer Model 3300 (1969)

Location: Henderson, CO
Symptoms: Both record lifting arms jammed into the record carousel.  The carousel was unable to rotate.  Trip switch not working.

I spent a good half-hour trying to unjam the record arms from the carousel and realized there was no hope without total dis-assembly.  Both tips were bent over facing the rear of the jukebox.  I’m not sure how they got that way, possibly when the jukebox was moved to the basement.  I’m surprised the plastic tips hadn’t broke.  The only option was remove to the record playing mechanism from above the carousel, then remove the carousel.  I advised the owner that this was going to take several hours.  I got the go-ahead.

Once everything was disassembled, I noticed the three rollers that support the carousel were frozen.  A generous dousing of WD-40 didn’t solve the problem.  I finally took a pair of pliers and rocked the rollers back and forth sideways to break whatever corrosion or gunk kept them from turning.  That worked.  Then another dousing of WD-40.  I followed up with regular oil and all three rollers were spinning freely again.

The tips on the record lift arms needed to rebuilt after being jammed.  They are spring loaded and should open up as the arm rises into the carousel.  I disassembled each tip, cleaned, readjusted, and lubricated with silicone.

I put the mechanism back together and did a couple of trial runs with no records installed.  Everything seemed to be working great.

Once again, another bad trip switch.  I encounter these a lot on later model Wurlitzers.  As I have described here in this log before, I broke open the cover and flushed with contact cleaner.  It’s pretty much all one can do, as you really can’t get to the contacts themselves with any kind of abrasive. I reassembled the trip switch, adjusted it, and played several records.  Seemed to work okay after that.  On the jukeboxes I have done this, they are still working after nearly two years.  But still, it would be nice to find a replacement switch.

Amplifiers, Amplifiers, Amplifiers!

In an odd quirk of fate, I’ve had a rash of jukebox amplifiers to work on, as well as some Seeburg control centers.  The amplifiers are a Wurlitzer 532, Seeburg TSA1, Seeburg SHP3, and an amp from a Rockola 424.  The 532 is the oldest and the SHP3 is the newest.

Testing a Wurlitzer 532 amplifier.

Wurlitzer 532

At the customer’s house, the 532 was exhibiting all kinds of problems.  The sound was popping, cutting in and out, and most of the time was severely distorted.  One of the items that was causing all of the popping was a bad connection in the octal socket between the mechanism and the amplifier.  The power to the amp runs out, and back in, through the connector (except for the tube heaters which are on all of the time).  Also the amp has a rectifier on it which supplies power to the mechanism.

Each female pin on the octal socket is shaped like a “U”.  Over the years with the plug being unplugged and replugged, the top part of the “U” spreads apart and doesn’t make a good connection with pin. I took a sharp ice pick and poked it down into the edges of each hole and bent each arm of the “U” so they would be closer together.  That solved that problem.

However, there was still a lot of distortion.  I opted to bring the amp back for a bench test.  With a bench test, I can inject a known signal and trace it through the circuit with an oscilloscope to see where the signal breaks down. In this case, it was bad at the output of the first tube in the chain.  It wasn’t a bad tube, but a bad electrolytic capacitor connected to the cathode of the tube.  The amplifier will be getting re-capped (replacement of all electrolytic capacitors).  It’s overdue.

Seeburg TSA1

At the customer’s house the amp was distorted especially in the lower frequencies.  Changing the volume had no effect.  It seemed to be in both channels.  There wasn’t good stereo separation.

I brought the amp back for bench testing.  There is a coil in the amp that connects the left and right channels together.  I’ve seen this in other jukebox amps of the late 1960’s, and it doesn’t make sense.  After I disconnected it, I discovered the distortion was only in a single channel.  And it took me a while to discover it as I had to drive the input fairly hard for it to show up.  And it was more prevalent at bass frequencies.  It looked like someone was taking a bite out of the lower side of the sine wave (it wasn’t at the peak).  After checking the bias and the caps around the preamp stage where the problem was, I concluded it had to be the transistor.  I replaced it, and the problems cleared up.  I chalk that one up to weird transistor failure.

Rockola 424 amplifier

The service manual calls this a 40276A.  The amplifier itself has 40218 labeled on it.  Regardless, one channel was completely dead.  The mono switch didn’t have any effect, so the problem was in the output stage.

I brought it back for bench testing. I quickly discovered the 1200uF capacitor used in the feedback circuit had a broken lead.  This amp had been previously worked on by somebody else and they had replaced the capacitor without securing it well.  The vibration of moving the jukebox to Colorado from the east coast probably caused it to break. Since the lead was broken right at the capacitor, it needed to be replaced.

While checking out the amp, I noticed some electrolytic capacitors where leaking.

Electrolyte leaking from capacitors.

Those 100uF capacitors got replaced, as well as the main filter capacitor (2940uF was replaced with a 4700uF).  After cleaning the preamp wirewound pots (not easy or effective), the amp is working well.

The Seeburg SHP3 amplifier will be covered in a future post about a Frankenstein jukebox.

Haunted House (Gottlieb pinball machine, 1982)

Location: Wellington, Colorado.
Symptoms: Would blow fuse as soon as ball landed on lower playfield.

The problem was obviously a short related to something that powered the lower playfield.  I inspected the wiring and coils on the lower playfield and didn’t see any issues.  With the machine off, I checked for shorts in coil and flipper power wiring and didn’t see any.  I finally found the short with my meter in the playfield illumination wiring, and it turned out to be on the main playfield.

On the underside of the main playfield, there are 8 lamps that light the lower playfield. These lights only illuminate when the ball is on the lower playfield, and are controlled by a relay.  Next I did a visual inspection of each of the 8 lamps, and as usual, found the short on the last one I checked. 😉  The center contact was shorted against the side of the socket, which is typically where lamp circuits short out.

The owner had replaced all of the lamps with LEDs.  One important thing to note:  On Haunted House and Black Hole, these 8 illumination lamps are powered by 28 volts, which will blow out regular 6 volt LEDs.  The original bulb number is a #313. You either have to use an LED specified to replace a #313 bulb, or you have to modify the lighting circuit to operate on 6 volts (a procedure is here).

 

Wurlitzer 2104 jukebox (1957)

Location: Lakewood, CO
Symptoms: Needle not contacting record, plays wrong selections.

We selected a record and the needle hovered about 1/8 inch above it.  I moved the tone arm and felt it binding.  I put some silicone lubricant on the tone arm pivots.  I also noticed the tone arm audio cable seemed to be a little tight which added to the binding.  I loosened the cable clamps and gave it another 1/4 inch of slack.

Usually when a Wurlitzer is not making the correct selections, the problem is either dirty switch contacts in the keyboard, or a lubrication issue with the pre-selection arm or disk at the bottom of the mechanism. Since this was probably the cleanest unrestored jukebox I’ve come across, I didn’t suspect dirty switch contacts.

On the ’04 models the pre-selection disk (I don’t know what Wurlitzer calls it) is used to differentiate the letter selections.  I call it pre-selection because this disk the first thing that begins moving when making a selection.  The disk rotates into position determined by a couple of solenoids, then fires one of the numbered solenoids visible around the edge of the mechanism. This disk rides on 3 nylon wheels, which often stop turning with age. I lubricated the nylon wheels and got them turning again.  The jukebox now made the correct selections.

One additional item was that I cleaned the volume control with potentiometer cleaner.

Star Trek: The Next Generation (Williams pinball machine, 1993)

Location: Littleton, Colorado.
Symptom: Left cannon/gun shooter wouldn’t find its home position.

This one had me stumped for a while.  If I executed the diagnostic test for the gun shooter, it found the home position normally, without fail.  Yet when powering up/booting the game, it just cycled back and forth.   I checked and adjusted everything associated with the home microswitch.  I finally concluded there had to be a bug in the firmware.

It turns out I was partially correct.  There is an optical sensor located at the front of the gun to sense when the ball is in position and ready to be fired.  If this sensor fails, the revision LX-3 firmware ignores the home switch, will not home the gun, and it gives an incorrect error message. This post helped steer me in the right direction.

I burned (programmed) a new game ROM with more recent LX-7 code.  In the revision history, there is mention of “Enhanced the broken gun launcher opto compensation.” I don’t know if it fixes the problem of the incorrect error message or not, because I went ahead and fixed the opto sensor problem.

To troubleshoot the opto sensor problem, I first checked the voltage across the IR transmitter located on the left side of the cannon. It was about 1.3 volts, and this normal. Next I checked the voltage across the phototransistor located on the right side of the cannon.  It was reading zero volts, with or without blocking the light beam.  A properly working sensors would have approximately 12 volts across it with the light beam blocked, and between 0 and 1 volt with the beam unblocked.

Next, I checked the continuity through the wires down to the first connector under the playfield.  It turned about both wires were bad.  I noticed that the shooter solenoid wires to the gun had previously been replaced as well. It is a common problem that flexing wires thousands of times will break them.  I suspect they are broken where they go through the playfield.

I ran two new wires from the phototransistor down to the connector and left the old wires in place. That fixed all of the problems.

ST:TNG is one of my favorite pinball machines in terms of theme execution.  The audio clips are great and integrate well into game play.

Twilight Zone Pinball Machine (Bally, 1993)

Location: Littleton/Denver, CO.
Symptoms:  Flipper problems, ROM checksum error, Proximity Sensor error.

I started with the ROM checksum error.  I re-seated the game ROM in U6 and the other socketed parts, with no improvement.  I programmed a new 27C040 with the game code (decided to try the 9.4H version) and replaced the ROM and the problem was solved.  I verified the code in the old ROM and there were several bits that had reverted back to 1, which was causing the checksum error. This probably caused the game to lock up occasionally, which was one of the complaints.

The proximity sensor senses when a normal metal ball is delivered to the shooter lane (as opposed to the Power Ball which is made of ceramic).  It was simply out of adjustment and not located close enough to the ball to sense it.

The flippers had the most problems.  Neither of the left flippers were working properly.  The top flipper was kicking out, but not holding and the bottom flipper was vibrating and buzzing and weak.  At first I thought the two problems were related and that there may have been a driver problem on the Fliptronic board.  But that wasn’t the case.

The upper flipper had a broken wire on the holding coil. I resoldered it the coil and it solved that problem.

Fliptronic I Flipper Button Opto Board

The lower flipper had a more difficult problem to find.  I hooked the oscilloscope up to the flipper driver output of the Fliptronic board.  It was rapidly pulsing off and on.  When I disconnected the flipper connector, the problem went away and there was steady power to the connector.  I checked the coil resistance and found no problems.  It turned out to be the flipper button opto board.

Normally when pressing the flipper button, the optical path opens and the phototransistor begins pulling the signal down to ground from 12 volts.  When the voltage falls below 5 volts, it triggers the flipper driver via the processor.  A good opto circuit will pull down to below 1 volt.  This was only pulling down to about 4.8 volts, which was too close to the threshold of 5 volts.  When the flipper engerized, the ground probably shifted a little causing the circuit to think the voltage had risen above 5 volts, thus shutting the flipper off. As soon as the flipper shut of, the voltage would fall back below 5 and energize the flipper again.  A vicious circle, which caused the output to pulse off and on very rapidly.

I tried cleaning the opto interrupter, but that didn’t help.  The opto interrupters for this board are no longer available (as far as I can tell).  The only option was to replace the board.  Great Lakes Modular offers a totally redesigned flipper button board that utilizes magnetic sensors instead of optical. The pinball machine owner opted to give it a try.  I ordered one and installed it.

The board needs to be adjusted once it’s installed.  The LEDs on the board really help this process.  However, to adjust it, you need to slide the magnet into different positions and I found it really sensitive.  Just a 1/16″ movement will cause it to go out of adjustment.  Great Lakes recommends taping the magnet after getting it adjusted.  I opted to glue it.  We’ll see how long it lasts. Once adjusted, it had the same feel and response as the right side flipper button.

 

Seeburg M100C Jukebox

Location: Centennial, Colorado.
Symptom: Visible smoke coming from jukebox.

The owner said the smoke was visible at the front of the jukebox.  This could have been the mechanism or the selector switches, or even something from down below.  I visually checked the latch solenoid, but I didn’t notice anything wrong.

I checked the fuses and none were blown.  I carefully powered up the jukebox.  The jukebox was in the middle of playing a record when it was shut down, so it continued playing the record when I powered it up.  All seemed normal.

After a while, smoke did appear and it was coming from the latch solenoid behind the selector switches.

This is a common problem.  The solenoid isn’t designed for continuous use.  Under normal circumstances, a person deposits coins, the solenoid energizes, the person makes selections until their credits are used up, then the solenoid is de-energized.  That process probably takes no more than 30 seconds.

In this case, the jukebox had a small pushbutton switch on the rear that gave three credits every time it was pressed.  It was being used at a party and someone probably pushed the button a number of times establishing the maximum number of credits, then the person didn’t follow through and use all of the credits.  Eventually the latch solenoid overheated and shorted.

Victory Glass sells a solenoid protector that is plugged in between the Wired Selection Receiver and the selector keyboard. It has a small switch on it that selects between coin operation and free play.  The advantage to this solenoid protector over the solutions provided from various internet sites (where the latch solenoid is wired to the hold relay), is that if a letter is selected without a number or vice versa, the solenoid eventually times out and releases after about 30 seconds.  This is great if you have small kids around who might be randomly pressing buttons and then walking away.

 

 

 

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.