Hi Bruce,

the "4013" needs steep edges at its clock input, otherwise the outputs can be become undefined! So, I would recommend the following circuitry:



Using an additional "40106" also allows you to spend the SET input of "4013" a proper level shifter. Just fabricate by the help of NPN a suited /FORCEON signal. As long as the micro gets no power, the NPN will be turned-off and by this the SET input of "4013" will be low. To force an "on" the micro must turn-on the NPN.

The circuit works as follows:
When pushing the switch, the 220nF cap is quickly discharged and the output of left "40106" goes high. When releasing the switch afterwards, the cap becomes charged again, but at a much lower speed. So, this part of circuit provides a reliable debouncing and guarantees steep edges at the CLK input.

A full analysis of this debouncing scheme can be found here.



Some hints:

- The parts at RES input of "4013" provides a power-on reset, which might be useful after changing the battery. Omit it if you don't need it.

- The 10k resistor limits the current through your silicon rubber switch.

- The two 1k5 resistors limit the input currents, if Vbat is accidentally shorted (during battery change or removal).

- The "40106" and "4013" are powered by Vbat, of course.



Some hints on the output cap of LP2992. Use a 10µF/6.3V Y5V cap like the one here.

As you can see from the impedance plot





this cap fullfills the ESR requirements of your regulator! "Better" ceramics are worse here!!!

Isolate the succeeding decoupling caps (at the micro, for instance) by the help of 1R resistors, so that the low ESR of these decoupling caps will not violate the ESR requirements of LP2992 again through the back-door.

Kai