Looking at the manual, on page 31, the left/right motor is initially driven by a transistor Q27 (Solenoid Driver 12) on the CPU board. Is that transistor shorted? I wouldn't expect the problem to be upstream from there (the chip that drives the transistor), nor would I expect much else downstream if the motor is staying on continuously. When activated, this transistor completes the circuit for powering the motor, so if it's always on, that says me this transistor isn't doing what it's supposed to.

For the record, I could be reading the diagram at the bottom of page 31 incorrectly, but it looks to me like the following:

1. The motor gets 9v supplied via the shaker motor board. This is the only thing coming from the shaker motor board, but is required. Obviously you have it since the motor is on. This 9v is supplied constantly to the bi-directional relay at 2 points - one for one direction when the relay is off and one for the other direction when the relay is on.
2. Q27, when active, connects one side of the motor to ground, completing the circuit to power it. I suspect this is where your issue is. If Q27 if off, then the motor should be off. It also connects to 2 points on the bi-directional relay - one for one direction and one for the other.
3. Q24 activates the bidirectional relay. When off, the motor would move one direction, and when on, the motor moves the other. This could be a secondary issue if it never moves the other direction, but could also point to a bad relay. Basically this relay is just swapping 9v and the path to ground between the 2 leads on the left/right motor - thus controlling the direction.

All this to say that the motor staying on constantly, as soon as powered up, says to me that Q27 is bad. Nothing should be activating it immediately upon powering up. That is THE component controlling the motor being on/off, nothing else, short of a bad hack or a bad chip upstream from q27 that keeps it locked on, should be allowing the motor to stay powered on.

Probably testing when removed is best. If you don't have another TIP122 (I believe that's what it is) to just go ahead and swap in, you could also swap Q27 and Q24 to see if the motor starts behaving and the left/right relay stays locked on. If the behavior remains the same and the transistor is good, I'd start looking at the chip that drives the transistor, but that gets more complicated. You could at least tell if it's locking Q27 on with your DMM.

Yes - this is the only reason the motor would be constantly powered, as far as I can figure. It would have nothing to do with the left/right/center switches if it's always on the game wouldn't even be checking these initially until it got into t-rex diagnostics which takes a couple seconds to get to when booting.

You can check for voltage at the base pin of the transistor to see if it's getting locked on by the chip that controls it. I'm not sure what the logic level (i.e. what the voltage level is to turn it ON) is on this board, but compare it to the base on a transistor for a solenoid that you know is off (I assume 0v for off).

Obviously testing while the game and boards are powered comes with risks of shorting something out or worse, so be careful.

I don't know how much the Rottendog board differs from stock. I'm only reading the stock manual here, but it appears that the next thing up-stream is a resistor array RA25, which I think is acting as a pulldown resistor for Q23-Q30. However, it looks like Q23-Q30 are each driven by another transistor - Q15-Q22. Q27 appears to be paired with Q19, so check that one as well. Q19 has, I believe, a pull-up resistor in resistor array RA7. These are driven by a 7408 quad gate. Specifically Q19 is driven by the gate chip listed as 2J at pin 6. This is driven by the chip at 5f - which is a 6821 - a Peripheral Interface Adapter (PIA). I can't glean any information about how it works upstream from there, but I actually would rule out the transistors before blaming the PIA. I guess a quad gate could be bad - and there are likely testing procedures for that.

So, if the Rottendog is the same, I'd be checking/replacing Q27 AND Q19 and rule those out before suspecting other components.

As far as I can tell:

Q24 is simply turning the bi-directional relay on or off. It doesn't control the motor being on or off, just left or right.

Q27 is what controls whether or not the motor is on.

The motor in that diagram is a circle with a tilde in it. It has a blue lead and a blue/white lead. These both connect to separate common points on the relay. The relay takes in 9v and a connection to Q27, which allows a path to ground when active. 9v connects to 2 points, and Q27 connects to 2 points. When off, the relay provides 9v to the Blue lead and the Blue/white lead goes to Q27. When the relay is on, the 2 switches in it flip so that the Blue lead goes to Q27 and the Blue/white lead goes to 9v.

You should at least disconnect the connectors that go to the coils/relays/motors for in-board testing. You could be getting different results on your readings from the downstream components.

I think your assumptions are reversed regarding Q27 and Q24. If Q24 is locked on, it just keeps the relay on and the motor will only go one direction (when the motor gets power). The motor's power is essentially switched on and off by Q27. Q27 being locked on means the motor is on, regardless of the state of the relay. The relay (controlled by Q24) is only controlling polarity (and therefore direction). Q24 could be removed and the relay in its off state would still allow the motor to move one direction if the motor is powered.

For the record: I believe the diagram is misleading, by labeling Q27's connection to the relay as (LEFT/RIGHT) BRN/YEL. That isn't controlling left/right - just the path to ground via Q27.

The motor is always getting 9v on one lead and goes to Q27 with the other lead. All solenoid driver transistors give the solenoid/motor/relay a path to ground when activated. Imagine instead of Q27, you had a wire that you could manually ground to turn on the motor. You would only touch it to ground when you want the motor on. If you touched it to ground and kept it there indefinitely, the motor would spin indefinitely.

The RD board is different. It doesn't have pairs of transistors and as many gates driving the solenoids.

Q27 appears to be driven directly by U1 74HCT244, with nothing major between them. RA7 is still used as a pull-down resistor.

Glad you got it working!

I still don't see how that transistor could have locked the motor on, but I suppose there's a potential scenario where the game wanted to get the t-rex to a specific point, so kept powering it and never achieved it.

However, you indicated that the t-rex would move both directions, so I would think the transistor that controls the relay wasn't the issue... so maybe Q24 doesn't control the relay?

It's entirely possible that I'm misreading the diagram or that the diagram has the transistors mislabeled.

At any rate, it probably doesn't merit further exploration if you're up and running now.