We are designing a circuit to measure the current consumed by a Ultra-low power communication module.
The requirement is below:
1. Min current = 10nA
2. Max current = 1.5A
We do not need to measure DYNAMIC current, just want to measure AVERAGE current at a specified working condition such as Active, Sleep, Shutdown
Our design proposal is in attached file, that we will use a sensing resistor in series to the load. Because the current range is quire wide, so, we split the measure range into two big ranges:
1. LOW range: 10nA to 200mA
2. HIGH range: from 200mA to 1.5A#
In each range we will use several smaller ranges with the current sense resistor value selected so as to voltage drop on it in each range from 0.25mV to 25mV.
We want using Analog device's Instrumentation amplifier AD8220 as the sense amplifier. To switch between ranges, we use GPIOs from a MCU to control ON/OFF Power MOSFETs which in series to sense resistors to select appropriate range.
I have some questions:
1. Do you have any concern on AD8220 to perform this measurement? If you have a better solution, please let us know your proposal.
2. Is there any concern on connecting AD8220's inputs pin? I am thinking of two possible connections as below:
a). Include MOSFET's & Rsense resistor (as in the drawing) - This way we can use one Amplifier for several range, but don't know this connection cause any side affect that affect to the sense voltage, specially at nano Amprere current range?
b). Connect Amplifier's two input pins to two ends of Rsense - This case need one dedicated Amplifier for one range, that increase design's BOM cost and complicateness.
3. We plan to select sense resistors so that the voltage drop on the current sense resistor from 0.25mV to 25mV on each range. Is 0.25mV input to AD8220 too small?
4. Any recommendation to select for Rsense resistor to reduce the thermal noise ?
Thanks and regards
What is your target bandwidth? This is important for calculating the sensitivity of your system.
Do you have any restrictions/limitations on the design? How about power supply, what is the available supply voltage for the inamp? Common problem in using the instrumentation amplifier is when running into the input common mode voltage vs output voltage swing limitation. Please refer to this FAQ for more detailed explanation.
What is your target accuracy and are you planning to calibrate the initial error of the design?
Thanks for your reply.
Regarding your questions on our design, I would like to provide you with further information below:
1. Target bandwidth: This design is for measuring average current consumed at a specific working condition, which quite stable, we do not need to capture the quick current transition. So, this is nearly DC amplifier, a bandwidth less than 100Hz is enough for us.
2. Our design powered from three series AA batteries, so that the power source voltage will varies and lower than 4.5V - Minimum requirement for AD8220's power supply. So, we use a boost converter to rise it up to 5V providing a stable power source for IN-AMP circuit.
3. Our target accuracy is 2% average, but at the lower range (nA range), we can relax our specs to 4% or so.
Hope that with these information, you can let us know your advice on items that I list at the fist thread.
Thanks for the additional information. Please see my answer below.
1. Based from the available supply voltage of 5V and common mode voltage of 3.6V, you are operating outside the input voltage range of the AD8220 as specified on page 4 of its datasheets. I would recommend to use the AD8422 as replacement of the AD8220.
2. Having the MOSFET inside the AD8421 inputs means that any voltage across the drain and source (VDS) would reflect as error on the measurements. Make this error source as low as possible or much better if you could calibrate this out.
3. The AD8422 should be able to detect your input range. Please use this tool to check if you have valid setup conditions (Supply, Vcm, Input range, Output range, Ref voltage. etc) that will make the inamp to work properly. Also, you might need to calibrate out the initial error of the system (offset voltage, bias current x Rsource, CMRR) to achieve the targeted performance.
4. Thermal noise is dependent on the value of your Rsense. You can get the calculation here. Just make sure that the noise contribution of the Rsense would not be more than the smallest signal that you want to detect (250uV). I will also pick an Rsense with good TempCo because the drift could directly affect the accuracy of your measurements.
I hope this helps.
Thanks Emman for the details answer.