TI blog post: Battery test equipment signal chain

qwqwqw2088

TI blog post: Battery test equipment signal chain [Copy link]

With the rise of the lithium battery industry, the market for battery testing equipment has also become huge, and it is mainly used in the capacity division of 3C batteries and power batteries. 3C batteries have a small number of strings, and the actual use does not require high consistency for each string of batteries. However, since power batteries have hundreds of strings and are used in a relatively extreme environment, in order to ensure a longer service life, they require much higher consistency than 3C batteries. Therefore, the battery requires a higher current accuracy in capacity division. At present, according to market requirements, maintaining the requirement of 0.02% is a design challenge faced by battery testing equipment manufacturers. In order to gain a higher market share, the pursuit of accuracy, efficiency, power density and other performance has never stopped. It should be known that in battery equipment, it is mainly divided into three parts, namely bidirectional AC-DC power conversion, data processing unit, and battery testing unit. This article mainly analyzes the signal chain part of the battery testing unit that is closely related to the key points of realizing battery capacity division technology.

Signal chain
Since battery test equipment requires high output voltage and current accuracy, especially power battery test systems, we need to understand each level of signal conditioning. The typical block diagram is shown in Figure 1. Since the first-stage signal amplification factor is in the range of 50-100, the shunt resistor voltage drop is small, and a voltage change at the microvolt level will cause an error of ten thousandths.

Figure 1 Voltage loop and current loop

The DC error caused by the input bias voltage of the first-stage signal amplifier
can be eliminated in the final calibration process of the equipment, but the error that varies with temperature and input and output conditions is difficult to eliminate through linear calibration. The main influencing factors of the first stage are:

1. Input voltage offset drift of amplifier

Generally, the appropriate value range is selected according to the temperature rise value of the equipment. The common application scenarios are shown in Table 1:

Table 1: Typical application environment

Current sensing using instrumentation amplifier INA821: temperature drift 0.4 V/°C

It can be seen that at the maximum current, the shunt resistor voltage drop is 60mV, and the INA821 output drift caused by temperature drift is 0.4*50=20 V. At this time, the error is 0.0333%. The actual circuit board temperature rise is less than 50℃, so INA821 definitely has a good advantage in actual use. At the same time, you can also choose a zero temperature drift device such as INA188.

2. Common-mode rejection ratio (CMRR) of the amplifier

In high-precision battery test equipment, a high-side current detection method with good noise environment and high reliability is usually used. Due to the high common-mode voltage, an amplifier with high common-mode rejection ratio is required. First, the common-mode rejection ratio can be expressed as

Ad is the common mode gain, Acm is the differential mode gain, and the error caused by the common mode rejection ratio can be expressed as

Vin_cm is the input common mode voltage, Vin_d is the input differential mode voltage, and the common mode error seems to be an error that can be calibrated. When the common mode voltage remains unchanged, it can indeed be offset by software calibration. However, since the actual voltage of the divided capacity battery increases from 0V to 4.2V when fully charged, the common mode voltage changes with the charge and discharge time, so the common mode error will become an error that cannot be calibrated. At this time, a device with a higher CMRR needs to be selected. At a gain of 100 times, the errors caused by the CMRR of several different devices are given according to equations (1) and (2):

3. Other factors

When selecting other passive components such as shunts, there are also temperature compensation methods that can reduce the errors caused by temperature drift, which will not be discussed here.

Of course, there are some manufacturers who try to reduce the nonlinear error during calibration by implementing multi-segment fitting. However, due to the consistency problem during mass production, this requires a lot of work to verify the batch data and find a universal temperature drift multi-segment calibration curve. However, due to the consistency problem, it is easy to cause overfitting error.

The gain of the operational amplifier in the second stage compensator is
within 10 times, and the output voltage of the compensator is above 1V. Usually, the noise and temperature drift of the operational amplifier are at the microvolt level, and the error caused is only a difference of one hundred thousandth of a percent. Since the output dynamic response required by the battery test equipment is not high, the design of the compensator parameters only needs to ensure good steady-state characteristics, that is, sufficient phase margin and large DC gain of the compensator.

Current command setting and data acquisition
Small current battery test equipment only needs one or two ADCs and DACs to solve the transmission of current commands and information collection of the entire machine. Using the structure shown in Figure 2, the multi-MUX solution can achieve a 1:128 or 1:256 usage of the main control board ADC or DAC and the test channel.

Figure 2 MUX & ADC sampling circuit

Due to the system software calibration technology mentioned above, the error mainly comes from ADC nonlinear error INL, temperature drift, and considering that in low-current battery test equipment, the time to read the voltage and current values of all channels in the system can be in the order of seconds, so the required sampling rate does not need to be very fast. However, in order to meet the current accuracy of one thousandth, a cost-sensitive ADC with a bit position of 12 bits or more is required, such as:

As for high-current battery detection equipment, the accuracy of newly manufactured equipment on the market can reach 0.02%. Therefore, the ADC needs to be more accurate and the sampling rate of each channel should be greater than 1kHz. This will increase the voltage and current refresh rate of the system, allow the ADC with bipolar differential input to provide a wider current variation range, and ensure that the reference of all signal chains from the instrumentation amplifier to the ADC detection is ground. When the sampling rate is lower than 100kHz, delta-sigma ADC is more commonly used: ADS131M08 is recommended.

天天1

Why is there a diode in series at the output end of TLV07? Wouldn't the positive output signal be cut off?