Does anyone see a problem with that for latching on the power?

I don't. Is this similar to the circuit you have in mind?

 

I would use AK's circuit in post #17.
It uses just one CMOS IC and one small MOSFET with no 555 required, and is very low power (only leakage current when off, and much less than a standard 555 when on).

Note that (if not apparent) the component labeled "reed switch" is your vibration sensor.

 

Below is the LTspice simulation of AK's circuit.
The output starts pulsing the LED (red trace) when the Vib switch S1 closes (yellow trace).
It stops when the Reset switch closes (green trace).

The only change I had to make, was to use a lower max Vgs(th) MOSFET (logic-level type), since the 2N7000 can have too high a Vgs(th) to properly turn on from a 3V signal.
The one I selected was an arbitrary one in my library. It can be just about any logic-level N-MOSFET with a max Vgs(th) of 2V or less.

 

#21 looks simple and inexpensive. Thx. But I wonder if output is less than 3v. My Ckt needs min 3 v to function.

 

But I wonder if output is less than 3v. My Ckt needs min 3 v to function.

If you change R1 to 3kΩ, the 2N3906 should have a output saturation voltage of <50mV @ 10mA load, so the output voltage would be about 2.95V with a 3V battery.
If you need less drop you could change Q2 to a logic-level type P-MOSFET with a low on-resistance.

 

My Ckt needs min 3 v to function.

Which part of the circuit is that?

 

If you change R1 to 3kΩ, the 2N3906 should have a output saturation voltage of <50mV @ 10mA load

Yeah, I'm working on an update. I temporarily forgot that Vcc was so low. For that, a 3906 gets you Vcc switching rather than GND switching, and more consistent performance at such a low drive voltage.

The 2N7000 Rdson is around 5 ohms, maybe less. At that value, a load current of greater than 20 mA favors a 3906, with a saturation voltage of less than 0.1 V.

ak

 

The CPC1002 SSR in place of the 2N3906 has a voltage output of 2.99 volts at 30ma.

 

The CPC1002 SSR in place of the 2N3906 has a voltage output of 2.99 volts at 30ma.

Nice find. A p-channel FET would get rid of the 2 mA holding current, but there aren't many that are qualified for less than 4.5 V Vgs.

ak

 

Such a great suggestion for using CPC1002N. It is cheap and small. I just hope the vibration sensor gives enough time for FET to close. It vibrates fast. Probably SW18030 (a slow or retarded sensitivity) would give more time. I should give that a try.
Thx

 

I would still like to know why the 3 volt output is so critical.

 

Hi sghioto
The circuit board has a couple of modules, one of which is ESP12. I believe it requires 3.3V but it has been working fine with two AAAs at 3V. I don't think I can go much lower.

 

What is the peak current the switch circuit has to handle?

ak

 

How long does 2 AAA batteries last with all that?

 

Great questions. That has been the biggest issue, with saving battery. It pulls about 90 to 100 ma. To save battery we put the ESP12 to 'deep sleep' after 10 seconds. Preference would be to hold the power for 10 to 15 seconds using the latch to wake-up circuit, then actually disconnect the power until the next vibration is detected. I could use a CMOS 555 to do that. But yes, need 3V, and need to hold power for a few seconds and release to save battery.

 

That has been the biggest issue, with saving battery.

need 3V

You might consider using one Primary Lithium thionyl chloride (non-rechargeable) battery (here), which has a voltage of about 3.6V.
They stay above 3V until near full discharge.
The come in various sizes, with the AA size having a capacity of about 2400mAh.

 

using the latch to wake-up circuit, then actually disconnect the power until the next vibration is detected.

Using the ESP to reset latch:

 

Nice find. A p-channel FET would get rid of the 2 mA holding current, but there aren't many that are qualified for less than 4.5 V Vgs.

What I like about the CPC1002 is the output can switch high or low with high or low input and rated 700ma.