how to track power comsumption of my tracker

[Amalinda]

I want to draw a graph of how my tracker unit absorbs power to be put into a document.
Total duration is about 60 seconds.

Within seconds there may exist 1-1.5A bursts by SIM908.

minimum current will be 0.1mA by microcontroler.
how to record this sort of reading that has a big gap?

thanks!

 

[kam1787]

how long are the bursts? how many might occur? Are the bursts the same shape/duration?

 

[Amalinda]

here is an approximation by the data sheet, but I cannot rely on it. so i want to measure mine.



in a minute there will be one transmission. and one GPS location obtaining and some processing in the MCU and back to sleep. that is one cycle.
I want to draw the power consumption of one cycle for my paper and I dont understand how to do it.

mainly because the GAP between mA and A current levels..
How am I going to record all of these?
Do you recomend recording all of these via a 0.1OHM resister and an ADC and then converting the ADC values later from voltage to power?

please let me know if you have a better way

 

[BradtheRad]

The range of mA to A is a ratio of 1:1000.
To measure that large a range, your ADC needs 10 or 11 bits.

If this is not possible, then your alternative is to measure current draw while the circuit is idle. State this figure as the quiescent current.

When you make the graph, your scale might be on the order of 2 inches per ampere. On that scale, 10 mA will show as 1/50 of an inch. It seems barely significant.

 

[Amalinda]

true what you say, and I don't like to use logarithmic graphs they are too mathematical and unpleasant to see on anything.
OK, so I guess I have nothing else but to change my requirements for the graph and consider only large power.

Since I anyway know the little MCU idle current and current when performing operations (which can be measured separately, easily) I will measure only the dominant power segments.

The problem is these 2A burst times are in micro seconds range. So I'm wondering if micro second times can ever be captured by an ADC. hopefully they can and I will give it a try.

my circuit obtains does two things.
1. GPS location
2. Transmits it.

I AM SO HAPPY NOT TO DO ANY OF THESE IF I CAN GET A NUMERIC VALUE OF POWER REQUIRED PER TRANSMISSION SO I CAN PREDICT LIFETIME WITH A BATTERY OF CERTAIN CAPACITY.

Is there any other way I can predict that without going through this painful surgery?
I just want to predict the lifetime and I have a 5000mAh LiPo.

if i know the power per one cycle (obtain location + transmit to server) I can easily conclude.

 

[BradtheRad]

If your quiescent current is 0.1 mA, it will not be much drain on a 5000 mAHr battery. Even 1 mA continuous will take 208 days to deplete it.

Although it is diligent on your part to estimate battery life, and include it in your treatise...
There is another factor which can be unpredictable. Rechargeable batteries are known to self-discharge over a period of weeks or months (NIMH type is reputedly faster than nicad). For a size D rechargeable this may turn out to be greater than 1 mA.

As for sending transmissions, by raw math I calculate the draw will be 12 mA average. (2 amperes for say 100 uSec, repeated each minute.)