Might want to check the 12V supply. The board makes 5V from the 12V supply and it's pretty unlikely that both boards are flawed.

Let's face it, WPCs are a bad design. Far fewer DE & Gottlieb machines from 90s have reset problems than WPCs. There isn't enough margin between the watchdog voltage and voltage regulator output.

Also, power supply technology has advanced since the 90s and a switching regulator (as used on the Kahr board) is superior to the factory LDO. That's why I generally replace the regulators on WPC PDBs with the EZSBC equivalents. I completely agree that a Kahr board won't solve all your power supply woes and it's best to do a formal diagnosis, but the board itself is not going to hurt anything and might work fine for quite a while in some situations.

LM323 is a linear regulator with a ~2V dropout. If the input voltage drops below 7V, the output voltage would drop to input - 2V (although there is a capacitor on the output side as well), but it would not go to zero and does not cause a reset by itself. The reset is caused by the watchdog detecting when the MPU 5V line drops below 4.7V.

And yes, given how pervasive this problem seems to be, it is unquestionably a design flaw. TTL logic typically has +/-10% margin on VCC and designers could have used a regulator with an adjustable pot set slightly above 5V to compensate for IR drop.