Some applications require calibration of the input offset voltage (VOS) of the current-sense amplifier to improve current measurement accuracy. However, calibration is not a simple process due to the amplifier’s minimum output voltage (VOL) and input VOS. This application note discusses a simple method to “tune” the input VOS of a unidirectional current-sense amplifier. Using this method to measure the input VOS is not limited by VOL, and can improve the overall current measurement accuracy. This technique is discussed using the MAX4080 current-sense amplifier as an example.
Current-sense amplifiers are mature ICs that are widely used in Electronic equipment to monitor load current in real time. The system controller performs power management operations according to the load information to change the characteristics of the load current itself, and can provide a flexible overcurrent protection scheme.
Current-sense amplifiers can reject input common-mode voltages while amplifying weak differential voltages. This function is similar to traditional differential amplifiers, but with one key difference: For current-sense amplifiers, the allowable input common-mode voltage range The supply voltage (VCC) can be exceeded. For example, the MAX4080 current-sense amplifier can withstand an input common-mode voltage of 76V when operating at VCC = 5V. With a stand-alone amplifier architecture, the current-sense amplifier is immune to common-mode rejection (CMRR) caused by resistor mismatches. The MAX4080 has a DC CMRR of 100dB (min), whereas traditional op amp-based differential amplifiers are CMRR limited, and their effective input VOS is amplified through the signal chain.
Figure 1. MAX4080 Precision Unidirectional Current-Sense Amplifier
Improve accuracy with calibration
The MAX4080 current-sense amplifier features a precision input offset voltage (VOS) of ±0.6mV max at 25°C and ±1.2mV max over the -40°C to +125°C temperature range. However, many applications require higher current measurement accuracy and therefore require further calibration of the input VOS. This calibration is achieved by measuring VOS during production and storing the results in firmware. Using the stored data, the VOS can be adjusted in the digital domain when the device is put into actual use in the field.
For ease of production, the preferred option for calibration is to measure VOS at zero load current (zero input differential voltage). The output VOS can be measured and subtracted from subsequent measurements. Unfortunately there is a disadvantage to this approach, the output voltage may not accurately reflect the input VOS due to the interaction between VOL (minimum output voltage) and the input VOS. All single-supply amplifiers have this problem.
Taking the MAX4080T with a gain of 20 as an example, and assuming that the input VOS is zero, the measured value of the amplifier output should be zero. The reality is that even at zero input differential voltage, the amplifier cannot guarantee an output voltage below 15mV (10µA sink). If the measured output voltage is used directly for VOS calibration, the amplifier’s input VOS is 0.75mV (15mV/20 = 0.75mV).
Likewise, if the MAX4080T has VOL = 0, a positive voltage input VOS should produce a positive output VOS. The negative voltage input VOS is not “reflected” to the output because the amplifier cannot produce an output voltage below ground. Thus, at zero input differential voltage, the input VOS cannot be calibrated by “directly” measuring the output voltage.
During production, there are two ways to calibrate the VOS:
Bidirectional current-sense amplifiers have an internal reference. For example, the MAX4081 has a 1.5V reference, which can offset the output measurement voltage at 1.5V, so that when the input differential voltage is zero, the output is 1.5V ±VOS, introducing errors. The 1.5V voltage is higher than the VOL of the amplifier and will not affect the error analysis. The VOS error can be calculated by measuring the difference between the output voltage and the theoretical voltage of 1.5V. However, this approach has a disadvantage: reduced dynamic range. For ADC devices with a 0 to 5V input dynamic range, the dynamic range is reduced by 30% and the output range is 1.5V to 5V. In addition, this method requires the use of expensive bidirectional current-sense amplifiers for unidirectional measurements. Finally, using a low-drift 1.5V reference or an additional channel just to measure the 1.5V reference is difficult for designers to accept.
The two-point measurement method applies two known differential input voltages (load currents) to the current-sense amplifier. First, based on the measured voltage, a linear approximation method is used to extrapolate the input VOS corresponding to the zero-sense voltage on the graph. Then, calibrate using the voltage measurements. The disadvantage of this method is that it requires two “known” precise current values, which are difficult to obtain in production, and also increases test time. Finally, a caveat: For differential input voltages close to zero, it is difficult to get an accurate measurement because the VOL limit will introduce errors at very small sense voltages.
Adjust input VOS with input resistance
This application note describes a third method for measuring the input VOS of a current-sense amplifier. Taking the MAX4080 as an example, applying a zero-input differential voltage can cancel the interaction between VOL and VOS—it can be easily used for production line testing.
All current-sense amplifiers have input bias currents, and input resistors must be used carefully (for example, as part of an input filter) because resistors introduce uncertain gain and offset errors. Application Note 3888: “Performance of Current Sense Amplifiers with Input Series Resistors” discusses these issues. This article uses a similar technique, but deliberately chooses mismatched input resistors to introduce additional output VOS. The MAX4080’s bias current is temperature compensated, with 5µA (typ) and 12µA (max) bias current over the entire operating range. A 2kΩ resistor in series with RS- (Figure 2) produces an input VOS of 10mV typical and 24mV maximum, respectively. The introduced VOS yields a corresponding output offset range of 200mV (typ) and 480mV (max), sufficient to overcome the MAX4080 VOL and VOS limitations. The error VOS introduced by the input resistance is temperature dependent and depends on the temperature drift characteristics of the input resistance (typically 100ppm) and the bias current (negligible).
Figure 2. MAX4080 configuration with an external 2kΩ resistor in series with RS-
A +100ppm resistance temperature drift characteristic will produce a +1% resistance change (ie +20Ω) over a 100°C temperature change. Thus, the additional input VOS drift due to the input resistance is about +0.1mV typical and +0.24mV maximum (over the full bias current range). And this temperature drift value is only 20% of the input VOS bidirectional error (±0.6mV) without calibration, which is an acceptable result without calibration.
Further reducing the series input resistance reduces drift error. Assuming a VOL of 15mV and an input VOS of ±1.2mV over temperature, the minimum value of the additional input VOS must be 1.2mV + 15mV/20 = 1.95mV ≈ 2mV. Table 1 lists the test results over the entire temperature range. The MAX4080 ignores the VOS temperature drift of the test amplifier itself. The measured VOS temperature drift is generated by the input resistance and its ppm temperature drift.
Table 1. Temperature Test Results with and Without Input Resistance
|No Input Resistors||-0.015mV||0mV||-0.005mV||-0.01mV|
|2kΩ in Series with RS-||9.69mV||9.73mV||9.76mV||9.80mV|
This application note describes a method for calibrating input offset by introducing a known input VOS into the MAX4080 current-sense amplifier by appropriately adjusting the input resistance. Device manufacturers can use this method to calibrate VOS at zero input current during production, improving real-time measurement accuracy.