XCore Exchange

would a piezoelectric sound trigger be appropriate?

Well my idea would be to just sample the ADC in sleep and if a high enough voltage is seen it starts sampling at full speed and process the data, or just use a low power opamp to wake it up using a interrupt so it can sleep all the time.

Thinking out of the box here litterally. Would it be possible to split the device into a sensor with lowpower transmission device (bluetooth or similar depending on range) sending data to a repeater/GSM unit somewhere that run on mains power to lessen the stress on battery resources ?

/Mikael

PS Guess the stress will land on device cost instead, but if all seems impossible maybe the approach can solve something, dunno DS

mike34 wrote:Thinking out of the box here litterally. Would it be possible to split the device into a sensor with lowpower transmission device (bluetooth or similar depending on range) sending data to a repeater/GSM unit somewhere that run on mains power to lessen the stress on battery resources ?

/Mikael

PS Guess the stress will land on device cost instead, but if all seems impossible maybe the approach can solve something, dunno DS

This is possible - as you said, pressure might fall on device cost. Nearest mains is close (<5 metres), just can't be used to directly connect to sensor or sensor battery.

If the best we can do is 1 month, that's still potentially manageable. I'd just prefer it to last as long as possible...

There is a customer willing to pay £50 a unit for 50 units for a trial...

Is there a Wi-Fi connection available or potential to install one?

Andy wrote:Is there a Wi-Fi connection available or potential to install one?

No, not really - neither in the specific case nor the general case. A version could be made that used wireless instead of mobile networks, but this would be a different class of device (with better power properties!).

Doing it wireless like that doesn't have much advantages as the power hungry part is the MCU/DSP since it needs to be ready all the time. How about a phone line being available. It can be used to power the device and it can use it to dialup in to the internet at say midnight when nobody is using the phone line.

Regarding audio processing:

If you use Audio and a microphone, remember that the electrical signal that comes out from the microphone is very different from how the humar ear and our Brain percieves the sound.
The humar ear has a heavy attenation of low frequencies specially at smaller Sound Preassure Levels. Sounds that humans do not even notice can clip an pressure sensing microphone, for an example infrasounds e.g. the backgroundnoise measured outdoor in [V] is usually much worse than we percieve.
If you design a threashold solution "in you head" based upon what you hear with your ears - you might end up with something that doesn't work IRL with a microphone.
It might be possible to use low power analog filters prior the detector to save energy, so the DSP can fall to sleep between the (sound) events.

It is probably also good to know that there is several different types of microphones, and each type can be sensitive to the preassure field but also the particle velocity of the propagating soundwave or a mix between preassure and velocity depeding on the design.
Lack of such knowledge can give straaaaange results.

Or just use a small DC blocking cap

is it possible to considet a client/server design?

several sensor nodes with, for example, low power XBee mesh networking module (or other low cost IMC RF modems, even point-to-point) connecting to master server node with connection to mains, with RF modem and GSM modem?