86 Generator Differential?

I agree with Pwrtothepeople. 86 relays are very fast, even at a fraction of rated voltage. Seven milliseconds is possible. If it is possible, I would replace the lamp with a separate over-rated resistor(s) to draw the same current and an LED lamp (with parallel inverse diode). The metal clad MIL spec 10 watt resistors are good.

The bulb was a 120PSB and the socket is a GE104PXG42, no resistor.

Chemman, I see 2 possibilities here. First, the 120PSB bulb has a filament resistance of 4800Ω, which is why is doesn't need a resistor. If the filament broke and reconnected in a shorter length (which happens occasionally in incandescent bulbs), the filament resistance would be greatly reduced. Since that bulb is the only resistance in the circuit, the resultant current could be enough to trip the 86.

The other possibility is the bulb & socket contacts of . Your bulb has a telephone slide base. The bulb must be inserted into the socket in a specific orientation for contact, but it's relatively easy to get the lamp misaligned to the point where it will still light, but there is a relatively small clearance between the bulb contact on one side and the socket contact on the other side. Dust or humidity could then bridge enough to allow current flow in a 125V circuit. If you look at the bulb contacts you may find indication of overheating.

Personally I think bulb misalignment is the more likely answer. I have replaced many of these sockets over the years due to the bulb contacts shorting out the socket and damaging it. For a critical circuit such as this, I recommend either bayonet or screw-type sockets which maintain positive separation of the contacts.

If it is possible to change lamp socket I suggest 24V MES E10 green LED lamp which draws 15 mA with two ARCOL HS10 resistors in series 1 of 2200 ohm, one of 4700 ohm total 6900 ohm. I put two of these on a generator breaker position open/close application 20 years ago. Still working with same lamps....