hi,
That looks OK to me, I would increase the 1K to 2k2.

Did you follow what I said about the 12V protection diodes in the 12V battery rail.?
E

EDIT:
This simple LTS sim shows the switching action, the timing is not set for hours.!! for this test.
The 10meg resistor is ONLY for this test, it leaks away the voltage on the 10uF.

 

Did you follow what I said about the 12V protection diodes in the 12V battery rail.?

Errr. No. Is it a diode in line with the power feed? Or across it?

 

hi,
For vehicle mobile equipment I add these components,
E

 

Thanks all, that's given me some more ideas to try...
Components on order

 

One issue I see is if Vbatt droops below the voltage on the charged 10uF cap, you'll be exceeding the Absolute Maximum Rating on the inputs (its Vdd on the TS555). Vbatt can droop considerably when starting the car (or disconnecting from power). Consider a silicon diode (1N4004) in series with Vbatt to Vdd and a low leakage Shottky diode from (+) capacitor to Vbatt. The failure mode originally described kind of makes me think whacked inputs or latchup.

Just my .02

 

I've modified the circuit with the suggested diodes and 10 ohm resistor and it's been working OK since the weekend.

Also the time period got so erratic, sometimes staying on overnight, that I changed the capacitor to a 470uf and put a 7.5M resistor across it, now it stays on for 1hour 10 minutes pretty much every time.

 

Why use a relay? Just hook the resistor and led to the drain of the mosfet.