cascaded delays
I'm also posting a simplified circuit in response to a PM from Jules. This topic has come up before, in the hay baler thread for example. I also suggested chained monostables to generate the delays (using the quad 558 timer, for ex.), but thought a sequence generator based on a CD4017 (or 4022) might be simpler :wink:. Maybe not.
In the circuit shown below, the 4017 sequences through the delays. The cap stays the same for all timing periods, but a different timing resistor is switched in each time the 4017 increments. This approach works best when the delays have the same order of magnitude. The diodes keep the outputs from interferring with one another. Diode Dx provides a quick discharge path when the schmitt trigger output is low, much like the discharge transistor in a 555 monostable does. The second NAND gate is used as a negative-logic OR, and provides rising edge clocks for the 4017. The external trigger input should pulse low briefly, as in triggering a 555 monostable. For a high-going logic trigger, use the remaining section of the NAND package to invert the signal. The 4017's "0" output keeps any leakage current from Dx from charging up the cap during the idle state. The cap charges from ~ GND when output "1" is active, and from the schmitt triggers' lower trip point for outputs "2" and above. Note that the short discharge cycle adds to the delay of each stage except output "1", and should be accounted for. It's not shown on the schematic, but output "6" (pin 5) should be tied back to the 4017's Reset pin (pin 15).
The outputs should sequence as follows:
0) idle state
1) initial pulse (pulse length wasn't specified)
2) 5 sec delay
3) 1/2 sec pulse
4) 1 sec delay
5) 1/2 sec pulse
6) reset, back to idle state
The necessary 4017 outputs would need to be OR'd together for a single output line.
This circuit hasn't been tested, it's a back of the envelope design and may have problems I haven't considered. To improve it, I'd probably put an edge-detector (XOR) between the schmitt trigger and the 4017s clock. Then, successive 4017 outputs could control both the schmitt triggers' charge and discharge cycles (alternating diode polarities where needed).
A small micro is another good alternative, if you're willing to invest some time in the learning curve. Many of the 8-pin micros have internal oscillators, and can generate any arbitrary sequence of time delays.