[Warning – long post ahead]

(I wrote this up a couple years ago, which is why it reads like an active project, but it's not.)
-----------------------------

This seems to be a common problem on Gottlieb EMs, and low and behold, I ran into one myself. I was able to correct it, so thought I’d pass along what I found, in case it helps someone else. (Plus I can find this again later after I forget what I did a long time from now…)

The Symptom:
You play a game and are on ball 5. The ball drains, but kicks out again to the shooter, yet the game is over. The Game Over light is on and the playfield is dead. You shoot the ball to let it drain back to the outhole, and wonder why is it kicking out the extra ball? The game plays fine otherwise, stepping through each ball in sequence.

What’s going on here?
So the ball kick out is fairly straightforward: the ball return coil fires when a switch is closed on the ball return relay (usually the O coil) and score motor switch is closed (the score motor runs; this was switch 4C on the schematic I was using). OK, so since the ball is kicking out, those two switches must be closing. Since it does that for the first five balls in the actual game, it must not be problem. Something else must distinguish when to kick out/not kick out a ball, and in particular, to NOT kick it out after the last ball.

Since the ball return relay (O) is in the path, and is a relay that is activated with conditional logic, maybe there’s something more to it that.

Looking at the logic for the (O) relay, there’s the ball return switch that needs to be closed. That makes sense, since that’s the leaf switch for the ball in the outhole. In line with that are a few score motor switches (three on the schematic I was using), then way off to the right side of the schematic there’s a normally closed switch called the 6th Pos. “Ball Count” Unit. This is the top switch in the switch stack on the ball count stepper unit. It is normally closed, until the ball count stepper steps to the “6th position”, which is the “Game Over” position. What this means is that when the ball count stepper finally reaches the 6th position, this switch opens, and the (O) relay can no longer activate. If the (O) relay can no longer activate, then neither can the ball return coil (which in turn means the ball won’t kick out again).

Well, that’s interesting, but if the ball is still kicking out after the game is over, that means this switch must still be closed, allowing the (O) relay to activate, etc. Gotta dig deeper.

So, the goal would be to make sure the (O) relay does not activate after the last ball drains (ball 5 for a 5-ball game). And it looks like the intended way to ensure that it doesn’t is for the 6th position ball count unit switch to become open. So then what causes that switch to open? Well, getting the ball count unit to step to the 6th position (which is the “game over” position).

Looking again at the schematic, the Add “Ball Count” Unit coil is the one responsible for stepping the ball count unit forward. It has two separate paths that activate it: one is common (shared) with the path that activates the (O) relay, except that instead of the ball return (outhole) switch, the “trough switch” instead completes the circuit to activate it. This is the switch on the underside of the playfield that lies in the path from the outhole to the shooter; the ball rolls over it after being kicked out of the outhole before depositing in front of the plunger.

The Distinguishing Clue:
To really see what’s happening in this case, you can prop up the playfield (first remove the pinball!), and start a game. Everything will reset up until kicking out the (non-existent) ball. If you then manually trigger the outhole switch (I like to use a wooden pencil for this – no metal!), the ball return kickout will trigger. But – notice that the ball count number will not be at 1 yet (on the backglass). Then manually trigger the trough switch (simulating the ball rolling over it). Triggering the trough switch will activate the ball count unit, and the ball count will now be at 1 on the backglass. You can now run through all five balls manually, by just triggering the trough switch – that’s the switch that is causing the ball count unit to step forward.

But see what happens after ball 5 now. Simulate ball 5 draining by triggering the outhole switch again. The ball return kicks out another ball! But note – Game Over is not yet lit. Now trigger the trough switch again. Bam! Game over. Playfield dead. This is the problem behavior. However, note that the ball didn’t really kick out AFTER the game was over; kicking out the sixth ball is actually what caused (triggered) the game over. So why didn’t the game end on ball 5?

The Cure:
Since we know that the (O) relay needs to not activate in order to not kick out the extra ball, and we know that the (O) relay won’t activate if the 6th position ball count switch is open, we need to find what else will advance the ball count unit that is NOT the trough switch. We know the trough switch path is working (because balls keep kicking out), so there must be a path to advance the ball count unit without kicking a ball out – and there is.

The other path that advances the ball count unit is the one that checks if the “last” ball has been played, and which is based on the ball count unit itself. There are five things in this path, and if one of them is not behaving, then an extra ball will kick out instead, triggering the trough switch again, and thus finally ending the game (but with a dead ball in the shooter lane).

The five items are:
- the zero position “ball count” unit switch
- The 3-ball/5-ball adjustment switch
- The ball count stepper unit positions 3 and 5 (position 4 is lumped with 3)
- Score motor switch 1A
- Switch on relay (O)

In my case, I tried both the 3-ball and 5-ball setting, with the same result, which told me the problem was with one of the items common to either ball count setting. I cleaned the zero position switch in the ball count unit (which closes after the ball count unit steps to ball 1), score motor switch 1A, and the switches on the (O) switch stack (did them all since I was there). I also cleaned the rivets on the ball count unit, and the spring switches. One additional thing I did was to slightly realign the wiper disc on the ball count unit to make sure the spring switches “rode” on a less worn spot of the rivets when it stopped (I did this by ensuring it stopped slightly earlier than it had been, which got it more on high spots).

I suspect it was the zero position switch that was the actual root cause in this case, but I can’t be sure, since I cleaned/adjusted more than one thing at the same time. But, the game now ends correctly, with no extra ball kicked out.

So there you have it. Maybe this will help sort out this issue on another game sometime.

Yes, there can be other causes as well. Another one I hit was due to switch tabs touching together inside the score motor wire harness (but it was also feeding an extra ball too early as part of game start, as well as one after the game).

It's useful to know whether a game "develops" an issue (i.e., has been working correctly and then exhibits a quirk) versus "it's always been this way" (i.e., at least relative to when you're working on it). The former can be narrowed down more quickly (I think), since there at least was a good baseline, so something likely fell out of adjustment, whereas the latter opens up to lots of possibilities (modifications, incorrect reassembly, hacks, etc.).

As it turns out, on a Gottlieb, only the dogbone in the score motor switch position 2 is in the 'L' position; all the others are 'S'. Just happens to be a handy thing to know/spot check when going through a game (and when disassembling/reassembling a score motor). And awesomeness when you catch one in the wrong spot!.

The lifter arm that sits inside the dogbone bracket, which looks like this:

Dogbone bracket

The dogbone (lifter part) fits into the S or L slot, and is usually black. The dogbone bracket is usually silver.

I've no idea why they're called that, but that's the name used in the Gottlieb parts catalog.

The dogbone brackets are only used on A or C levels (the dogbones ride on the upper and lower cam edges); the B and D levels are for the posts that stick up from the cams. So there's no reference because there are none used there.