IoT Pilot POC- Current draw issue- assistance/guidance requested

Thread Starter

Steven Downey

Joined Sep 6, 2017
2
So we've developed an IoT sensor POC that we're going to pilot shortly. It's a small multitech mDot LoRa board with a ultrasonic distance ranging sensor. It will take liquid level measurements daily and send them to our backoffice solution. 99% of the time it will be in sleep mode. The board itself only draws ~40ua in sleep, but the ultrasonic sensor (a maxbotix 7389) is drawing 1.9mA in sleep. So with a 2500mAh battery, the expected life is not the 3-4 years we are trying to get, but rather 41 days.

The two of us working on this are both software guys, not hardware, and we're fumbling around trying to figure out a way to cut power to the sensor. We're looking at using an N-Channel mosfet but i wanted to see if anyone on this forum could provide some guidance.

Thanks in advance

Steve
 

Sensacell

Joined Jun 19, 2012
3,432
Without seeing the data sheets for your sensor it's a guessing game.

The general idea would be to use a P-FET to switch the power supply high side, not the ground.
Make sure all the IO lines going to the sensor are LOW when you power it down.
Take care to test how much time the sensor needs to wake back up, you might find it takes a loooong time- test it.
 

Thread Starter

Steven Downey

Joined Sep 6, 2017
2
Without seeing the data sheets for your sensor it's a guessing game.

The general idea would be to use a P-FET to switch the power supply high side, not the ground.
Make sure all the IO lines going to the sensor are LOW when you power it down.
Take care to test how much time the sensor needs to wake back up, you might find it takes a loooong time- test it.

Interesting. We want the default state to be off, unless we turn it on. We thought a P would set the default to on (gate closed if I have the terminology correct) by default. We want to maintain that state while the module is asleep. So it sounds like we had it backwards?
 

Sensacell

Joined Jun 19, 2012
3,432
Interesting. We want the default state to be off, unless we turn it on. We thought a P would set the default to on (gate closed if I have the terminology correct) by default. We want to maintain that state while the module is asleep. So it sounds like we had it backwards?
If you break the GND connection, then the IO lines need to be in a HIGH state to avoid current flow from the IO- powering the sensor.


Breaking the high-side with a P-FET preserves a solid GND connection, if you use a N-FET and break the GND, the extra resistance in the GND lead can be problematic.

Depending on how you drive the P-FET gate, you can make it default to OFF.
But remember to set the IO to High-Z, or LOW state, otherwise substantial current will flow into the IO pins of the sensor- defeating the whole idea.
 
Last edited:

ebeowulf17

Joined Aug 12, 2014
3,307
Interesting. We want the default state to be off, unless we turn it on. We thought a P would set the default to on (gate closed if I have the terminology correct) by default. We want to maintain that state while the module is asleep. So it sounds like we had it backwards?
There isn't a default state in the way that you're making it sound. The input (gate) to the MOSFET should never be left "floating" in an indeterminate state. You can use pull up or pull down resistors to define what state the gate is in if/when there's no microcontroller signal telling it what to do, and the microcontroller can actively drive the input high or low when it's powered and awake. So you choose whatever default you want no matter which architecture you choose.
 
Top