AMC3302-Q1: High Side BMS Shunt Current Measurement

Lukas Geiger

Part Number: AMC3302-Q1
Other Parts Discussed in Thread: ADS131B02-Q1, ADS131B04-Q1, ADS131B26-Q1

We are trying to design an isolated current measurement design based on a shunt measurement principle for high battery charge and discharge currents. The design should feature digital signal transmission to a master MCU, preferably via CAN. The AMC3302-Q1 could be a starting point for the design, but i wondered wether there is a more comprehensive solution to this problem.

With our power loss requirements (see below) the input voltage range of the AMC3302-Q1 is not favorable. We would not be using the full range when using a small shunt resistance (25µR to 35µR), therefore sacrificing resolution.

We would like to meet the following specifications for our measurement:

Parameter	Value	Unit
Nominal Measurement Range	±500	A
Power Loss (Shunt)	<9	W
Total Accuracy	±0.4	%rdg
Offset	75	mA
Linearity	0.01	% of range
Resolution	27	mA

Additionally we want to integrate various temperature measurements. Specifications on this still need to be worked out, but it may be relevant to an answer for the current sensing architecture.

over 3 years ago

+1 Alexander Smith over 3 years ago

TI__Guru 56592 points

Hi Lukas,

27mA of resolution for a full-scale range of 1000A is very high accuracy requirement, 27ppm of error.

Are the values listed typical or maximum allowable errors?

What is your sampling rate requirement?

We may need to explore considering a higher accuracy delta sigma or SAR ADC then using a digital isolator to communicate across the barrier.

https://www.ti.com/data-converters/adc-circuit/precision-adcs/products.html#p84=24;32&sort=p1130;desc

https://www.ti.com/isolation/digital-isolators/products.html

0 Lukas Geiger over 3 years ago in reply to Alexander Smith

Prodigy 40 points

Hi Alexander,

thanks for your quick reply.

27mA is typical, all other listed values are maximum.

Sampling rate is quite low (minimum 50 S/sec), as this measurement is not needed for any regulation task, only for monitoring/functionaly safety.

I will explore your suggestion, and maybe get back to you, if anything comes up.

+1 Joachim Wuerker over 3 years ago in reply to Lukas Geiger

TI__Mastermind 24285 points

Hi Lukas,

thanks a lot for reaching out with your request. Alex just made me aware of your inquiry.
I have a lot of focus on current shunt sensing in BMS. Your requirements seem pretty typical for this application.

We usually solve this problem by using a high resolution, high accuracy, 24-bit delta sigma ADC. The ADS131B02-Q1 and ADS131B04-Q1 are great options for it. Unfortunately those ADC do not integrate the data isolation and isolated DCDC-converter. Means you would need to add them externally. But TI has many different options for the digital isolator and isolated DCDC-converter.

Following some rough estimations using the ADS131B04-Q1 with a 35uOhm shunt:

Max voltage across shunt: ±500A x 35uOhm = ±17.5mV
FS range of ADS131B0x-Q1 @ Gain = 32: ±37.5mV
Input-referred noise of ADS131B0x-Q1 @ Gain = 32, Data Rate = 250SPS: 0.42uV_rms
Resolution in mA: 0.42uVrms/35uOhm = 12mA_rms
Offset @ 25°C, Global chop mode enabled: 2uV (max)
Offset in mA: 2uV/35uOhm = 57mA
Gain error needs to be calibrated at room temperature at least. Only the gain and VREF drift are important then.
Gain error due to temperature drift (e.g. Ta = -20°C to 60°C): 25ppm/°C x (25°C - (-45°C)) = 1125ppm = 0.11%

Overall it looks like the device would meet the requirements pretty well.

You can leverage some of the other available ADC channels in ADS131B0x-Q1 to measure battery voltage and temperature.

We also provide some basic functional safety documentation (Base FIT rate, FMD, Pin FMA) on the product folder for those ADCs.

Regards,
Joachim Wuerker

0 Lukas Geiger over 3 years ago in reply to Joachim Wuerker

Prodigy 40 points

Thank you Joachim,

i think we're gonna sacrifice some resolution and choose Gain 8 to extend measurement range. This was very helpful!

Regards

Lukas Geiger

0 Joachim Wuerker over 3 years ago in reply to Lukas Geiger

TI__Mastermind 24285 points

Hello Lukas,

gain = 8 is actually what is commonly used, at least in automotive BMS to also detect overcurrents of multiple kA.
If you look at the noise table in the ADS131B04-Q1 datasheet you will actually realize that the input-referred noise doesn't really degrade much between gain = 8 and 32.

Regards,
Joachim Wuerker

0 Joachim Wuerker over 3 years ago in reply to Joachim Wuerker

TI__Mastermind 24285 points

Hello Lukas,

I couldn't mention it two weeks ago, but I can now finally tell you that we put our new automotive BMS analog front-end for current shunt, voltage and temperature measurements in preview on ti.com. The device is called ADS131B26-Q1 and is developed according to our internal development process for functional safe hardware.

Product folder: https://www.ti.com/product/ADS131B26-Q1

Regards,
Joachim Wuerker

0 Lukas Geiger over 3 years ago in reply to Joachim Wuerker

Prodigy 40 points

Hi Joachim,

thanks. If this isn't good timing

Kind regards,

Lukas Geiger

Amplifiers

Amplifiers forum

AMC3302-Q1: High Side BMS Shunt Current Measurement