Joe, what SCR? Our 88-ohm example relay draws only about 150 mA, so the coil can be connected between the output pin and +V. Once the timing cap voltage rises above 2/3 Vcc, the output goes low, the relay pulls in, and it stays that way until power is removed.

A diode from the timing cap to +V ensures that it's discharged at the beginning of the cycle. Depending on the application details, there may need to be a dummy load R from +V to ground to give it something to discharge into.

The .01 uF bypass cap on pin 5 has a 33 microsecond time constant, so it has no effect at all on the relay output.

All of which is purely academic, because I would never connect a 555 directly to vehicle power, which exhibits transients that would knock your socks off. Like others here, I'd prefer a more robust solution.

Honestly, in this case I think I'd avoid silicon altogether and modify LarryC's approach: Use the R-C network on a pilot relay to drive the power relay.

The pilot relay would be a reed type with an adequate contact rating to energize the power relay and a 5-volt coil. The series resistor would then be just slightly higher than the coil resistance, maximizing the "R" (which, in turn, minimizes the expensive "C"), and guaranteeing operation at low battery voltage. It also reduces the voltage across C, further reducing its cost and size. I'd add a diode across the series-R to discharge the cap quickly when power is removed. Since the cap discharges through the low-resistance power-relay coil while the pilot relay is still closed, it happens quickly.

Example: The COTO 9081-05-00 reed relay ($1.66) has a 500 ohm coil. Choose series-R = 620 ohms, 1/4W and C = 2200 uF, 10V (0.4" diameter x 0.8" long, $0.39).

The pull-in delay is 1 second, and the dropout (of the pilot relay) is about 0.7 seconds.

Including a 30-amp automotive power relay such as the Potter&Brumfield 1432782-1 ($1.82), and a couple of 1N4001 diodes, the total parts cost is less than $4.50.

Cheap enough yet? [Linked Image]

Edit: The P&B relay has 30A contacts, not 40A.

[This message has been edited by John Crighton (edited 06-30-2006).]