How Can I Calculate Power Requirements from a Sensor Datasheet?

[lucy_b14]

Hi

I am trying to calculate the power required to take a measurement using a sensor. Datasheets do not provide the power requirement for devices (in watts). Can anyone give me any advice on how to find the power requirement using information provided in a typical datasheet? I know I could just use the equation P=IV, but surely it's not that simple.

I'm not sure if it makes a difference, but the sensor will not be used continuously, just one measurement needs to be taken.

Thanks for your help

L

 

[Pumblechook]

What power? DC, AC, high frequency, Where from? Where to? Roughly how much power?

I think posters would get a better response if they explained more.

 

[MATLABdude]

Hi

I am trying to calculate the power required to take a measurement using a sensor. Datasheets do not provide the power requirement for devices (in watts). Can anyone give me any advice on how to find the power requirement using information provided in a typical datasheet? I know I could just use the equation P=IV, but surely it's not that simple.

I'm not sure if it makes a difference, but the sensor will not be used continuously, just one measurement needs to be taken.

Thanks for your help

L

Hi Lucy,

Parsing a datasheet is a skill that takes a while to acquire and even longer to hone. Could you provide a link to the datasheet in question?

 

[lucy_b14]

Hi

Sorry, I guess I was a little sparse on detail - I was trying to keep it simple!

I am basically designing a wireless autonomous sensor device powered by energy harvested from the surrounding environment. This energy will need to be stored in a battery, so I need to find the amount of power (DC) that I need to harvest in order to select an energy harvesting technique (e.g. vibration, light...etc).

The sensor device will probably take a temperature measurement or something similar (what the sensor measures doesn't really matter). The following datasheet shows the kind of information (minimum voltage and current on page 3) that I'm trying to determine the power requirement from:

http://focus.ti.com/lit/ds/symlink/tmp175.pdf

Hope this is a bit more clear

L

 

[Proton Soup]

unless I'm missing something, it looks like most power consumption will not be through the device's power input, but through the digital interface pins. and if all pins are held high, should be about 1uA per pin when not in use. yeah, it's not straightforward.

 

[MATLABdude]

Hi

Sorry, I guess I was a little sparse on detail - I was trying to keep it simple!

I am basically designing a wireless autonomous sensor device powered by energy harvested from the surrounding environment. This energy will need to be stored in a battery, so I need to find the amount of power (DC) that I need to harvest in order to select an energy harvesting technique (e.g. vibration, light...etc).

The sensor device will probably take a temperature measurement or something similar (what the sensor measures doesn't really matter). The following datasheet shows the kind of information (minimum voltage and current on page 3) that I'm trying to determine the power requirement from:

http://focus.ti.com/lit/ds/symlink/tmp175.pdf

Hope this is a bit more clear

L

As Proton Soup says, the majority of the power consumption will probably be when you pull the SCL and SDA lines low (even more so if you use the One-Shot feature). When ball-parking (order of magnitude or so) power for something like this, I typically ask:

1) How much of the time will it be active and consuming maximum current? If this is a tiny fraction of the time (<5% or so of the time), and the "active" current is lower than the quiescent (or sleep) current, then you can safely ignore "on" operation.

2) If 1) Use the quiescent (or, if you're using One Shot, the sleep) current and multiply it by your operating voltage. The proper formula is: Power consumed (in a device) = sum(voltage_in * current_in) - sum(voltage_out * current_out). Since voltage_out is zero, this reduces to the operating voltage multiplied by the input quiescent / sleep current.

In the event that the first assumption isn't true, I'd recommend going with the worst-case scenario and assume that, when the device is on, it's drawing maximum current for the entire duration (instead of going on and off, as in the case of a digital input). In this case, you should *also* factor in the power dissipated in the pull up resistors when SCL/SDA go low (actually, this number should be orders of magnitude larger than the power consumed in the device when SCL and SDA switch low).

In the back of your mind, you should also be asking if you can supply the current for when the device is on. For instance, I had circuit brownouts once when I didn't properly take into account the need for 1/4 A for an RF transmitter which only required a few mA quiescent. Not a big problem, just annoying (and something you can confirm if you've got a decent power supply).