BQ27210 low current accuracy / sensivity

Don't think of this as a current measurement accuracy but really as voltage measurement accuracy. The coulomb counter input offset is specified to be +/-15uV in datasheet. Assuming that you are using a 20mOhm sense resistor, then 200uA would be 4uV across the coulomb counter input. You cannot count on that measurement to be accurate when it falls inside the offset range.

What is the time that this application spends at 200uA versus higher currents greater than 5mA?

Thanks for your answer MikeV.

First, I do not think that offset and sensivity/accuracy are identical. On many systems, once offset has been measured, the offset value can be substract and accuracy is very good. (I did not check if your device permits it).

This application spends between 95 to 99% of its time in low power (200uA), between 0.9 to 4.5% in medium power (4 to 20ma) and the rest of the time (0.1 to 1%) in high power (around 200mA).

Do you have any other solution to measured the battery Gas Gauge on low power monitoring system with GSM modem?

Best regards

Mich

What is the nominal cpacity of the battery you are monitoring?

For operation with currents this low you may want to use an Impedance Track type gauge which will not rely solely on coulomb counting but that can also make capacity reporting adjustments based on voltage measurements during battery relaxation states.

MikeV said:

What is the nominal cpacity of the battery you are monitoring?

=> the capacity is 2A but we can charge it at 200mA to permit the coulomb counting.

MikeV said:

For operation with currents this low you may want to use an Impedance Track type gauge which will not rely solely on coulomb counting but that can also make capacity reporting adjustments based on voltage measurements during battery relaxation states.

These low currents are the normal modes of the MSP430 (sleep, running) + radio (sleep, running) + GSM modem (sleep, idle, communication). This is a basic schematic mainly based on the MSP430.

=> how do you usually measure the current consumption of a MSP430? Do you have ny application note about the "impedance track type gauge"?

We don't try to report currents that low. As a matter of fact we have filters in place in our Impedance Track gauges so that currents below 3 mA are displayed as 0 mA. However the coulomb counter does take in account these small currents for charge accumulation counts but these are only noticeable after very long period of times.

For application notes on Impedance Track look in www.ti.com for application notes found on product web folders for devices such as bq27520-G3 and bq27541-G1.

MikeV said:

We don't try to report currents that low. As a matter of fact we have filters in place in our Impedance Track gauges so that currents below 3 mA are displayed as 0 mA. However the coulomb counter does take in account these small currents for charge accumulation counts but these are only noticeable after very long period of times.

I will check for accumulation. If it works fine, it will be good enough. Anyway I can change the Rsense to 50mohm (will permit 2A current max)

MikeV said:

For application notes on Impedance Track look in www.ti.com for application notes found on product web folders for devices such as bq27520-G3 and bq27541-G1.

These devices does not seems better (10uV of offset instead of 14uV) and are very slow (1sec for conversion time)

Hi Mich,

No coulomb counter will be perfect.  Even if you had a perfect coulomb counter, every battery from the production line will have a slightly different capacity.  Every cell will also have a different internal impedance profile and this impedance will increase unpredictably over time depending on temperatures, voltages, currents, time, etc.  Also, your discharge rate and temperature will always be varying, so the available capacity will always be changing and a fixed rate compensation won't be perfect, either.

Therefore the primary unique benefits of Impedance Track gauges are these:

1. gauging error (whether due to coulomb counting error accumulation or many other factors) can be corrected using an OCV (open-circuit voltage) update periodically
2. the true individual cell capacity can be learned with only partial discharges
3. the true individual cell impedance profile can be learned for accurate rate compensation, even with aged batteries
4. for system-side gauges, the gauge can have an accurate SOC immediately upon battery insertion using the OCV measurement.  It does not require charging all the way to full or all the way to empty to synchronize it.

In fact, the coulomb counter hardware on the IT gauges are probably superior to that of the bqJunior gauges, but even if it weren't it has other advantages that make the coulomb counting accuracy less important.

"However the coulomb counter does take in account these small currents for charge accumulation counts but these are only noticeable after very long period of times."

Can you clarify? What would the accumulator be set to and what is a very long time?

The gauge reports small currents (< DeadBand) as 0mA but they are still being used to calculate capacity parameters (RemCap, SOC, etc.). The smaller the current, the longer it takes to have a noticeable effect on these parameters (can be hours or days).