Current consumption problem

[fedimakni]

Hello,

I measured the current consumption of a specific iot device i designed (just composed by uC+sensor+passive components) quantity is around 500 pieces. i plot the current consumption in sleep mode in google sheet but i found that the distribution is so big some devices consume under 10uA but also there are other which consume more than 20uA which in some cases more than 3 times. (picture attached)
I didn't understand the reason behind this wide range.

Could anyone help us with some useful information about the possible reason and how to deal with this?

Thank you very much for all your kind response.

[Marco]

Did you do a repeatability test for sanity checking? (ie. after you're done with the main batch of measurements, pick a few devices you already measured and measure them again.)

[phil from seattle]

Also, how are you measuring the current draw? Is it over a period of time or just an instantaneous reading?  Perhaps the device, even in sleep mode, has variable draw.   

[Terry Bites]

You need to time average your measurements. The devcie will draw varying currents as its internal operations change.

[ajb]

In addition to the questions about your measurement method:

(just composed by uC+sensor+passive components)

When you say "uC", do you mean really just a microcontroller IC, or do you mean something like an Arduino or teensy?  Really no other ICs, not even an LDO?  What kind of sensor?  What is the range (min to max) of the quiescent current specifications for your components? 

[David Hess]

Check the specifications in the datasheet for the devices that you used.  Sleep current might vary over that range.

[thermistor-guy]

Hello,

I measured the current consumption of a specific iot device i designed (just composed by uC+sensor+passive components) quantity is around 500 pieces. i plot the current consumption in sleep mode in google sheet but i found that the distribution is so big some devices consume under 10uA but also there are other which consume more than 20uA which in some cases more than 3 times. (picture attached)
I didn't understand the reason behind this wide range.
...

Check if you have any uC or digital I/O pins configured as inputs, which are floating. If they are unused bi-directional, try configuring them as active outputs. Floating I/O pins will increase the current draw; the excess current will vary from one device to the next.

I've seen a related problem in a low-power application, that used a telephone line as a power source for a uC-based circuit. The operating current budget was 70 uA max under worst-case conditions. Our prototype software failed to initialize unused I/O pins to outputs, resulting in excessive quiescent current. The extra, unwanted, 20 uA current draw threatened the whole project. The solution was to initialize the uC carefully, immediately after power-up.

[David Hess]

Check if you have any uC or digital I/O pins configured as inputs, which are floating. If they are unused bi-directional, try configuring them as active outputs. Floating I/O pins will increase the current draw; the excess current will vary from one device to the next.

ESD damage can also cause that kind of variation in leakage current of floating I/O pins.

[SiliconWizard]

How can we figure out anything with so little information?

[fordem]

How can we figure out anything with so little information?

THIS is a question that I have asked myself on an almost daily basis since joining these forums - the original poster fails to clearly define the issue and then various respondents debate solutions to an unknown situation.