A while ago, I wrote about zero resistance ammeter. The more common name for it is transimpedance amplifier. It is a device that converts current to voltage. So an ammeter really. But unlike the ordinary ammeter, transimpedance amplifier (TIA) has zero voltage drop across its terminals. To achieve that, it has to be an active device. The most simple form comprises of an op-amp in inverting configuration with a single resistor, RF, that determines the transimpedance gain. In reality, there has to be an additional capacitor, CF, to compensate for the op-amps input capacitance [Figure 1]. The output voltage is then V = -RF·I.
An excellent overview on this topic from Robert A. Pease can be found here: http://electronicdesign.com/analog/whats-all-transimpedance-amplifier-stuff-anyhow-part-1.
|Figure 1: The principle of transimpedance amplifier.|
The big advantage of TIA is that is has zero voltage drop on its terminals so it does not affect the circuit being measured – unlike a regular multimeter. Also, regular multimeters tend to be much less precise on current measurements than on voltage measurements. Using a precision current to voltage converter, you can increase the precision of measurement and not influence the measured circuit at the same time!
When one range is not enough
If you need more gains (ranges), then you can place a switch in series with the resistors. It is a good idea to have a make-before-break (shorting) switch so you don't interrupt the flow of current during switching. The switch, however, adds uncompensated capacitance and series resistance. This limits the use of semiconductor switches because their capacitance is typically higher than the capacitance of mechanical and electromechanical (relay) switches. Switch resistance will produce a measurement error which increases when the value of gain resistor decreases. It is possible to use a double throw switch and use the second part to sense the voltage drop before the current enters the first part of the switch.
Powering from a single supply
To have a bipolar ammeter on single supply, you will need another op-amp for biasing. This op-amp needs to be as fast and as powerful as the first one. Remember, the op-amps will be sourcing/sinking the same current you are measuring.
The design idea
My design [Figure 2, 3] uses two OPA567 power op-amps in non-inverting configuration supplied from a single cell li-ion battery (RCR123A). One of the op-amps is held at bias voltage of 1.25 V (from a series voltage reference). The second is controlled by OPA320 in such a way that its voltage creates zero potential difference between the sense terminals. The reason for using OPA320 is that its input offset voltage is very low. Both OPA567s are used only for their high current capabilities. This configuration also enables easy remote (4-wire) sensing which elliminates the input cable resistance [Figure 4].
The current passes through precision (0.1%) resistors, the switch and two protecting PTCs. The voltage drop across the precision resistors is multiplied by two using an instrumentation amplifier INA333 referenced to the 1.25 V node. This amplifier isolates the voltage output so any loading of the output terminals does not affect the TIA's operation. The output voltage range is ±1 V so the voltage drop across the precision resistors is 0.5 V maximum. The remaining 0.75 V can be (although not entirely because the OPA567 cannot swing to exactly zero) sacrificed on other resistances, particularly the two PTC protection fuses and cables that will connect the TIA to its source.
|Figure 2: Complete schematic of my transimpedance amplifier.|
|Figure 3: Finished prototype. The range is switched to 1 V = 1 mA. Green LED on the left indicates battery OK. Red LED on the right would indicated an overload.|
|Figure 4: Testing the input offset voltage. At 102.9 mA (range set at 1 V = 100 mA), the input terminals were 30 µV apart.|
Safety and monitoring
Output is ESD protected by high-speed TPD2E001 diode array and 100R series resistors. Inputs are protected by series resistors (sensing) and PTCs (power). Battery input is reverse-polarity protected and also has a PTC and a TVS.
A microprocessor (ATTINY24A) monitors the battery voltage, output voltage of the instrumentation amplifier and output voltages of both power op-amps. Firmware controls whether the battery is OK and the output voltages stay in the device limits so there will be no clipping of the signal. Two LEDs are controlled by the micro: One shows the battery charge status and the other indicates overvoltage. The overload LED control is latching so it will be visible even when the overload has low duty cycle.
If battery voltage drops below 3 V, the LED will shine red, if it drops further to 2.7 V, the LED will blink and the microprocessor will shut down the power amplifiers because they will no longer be guaranteed to deliver enough power. At this voltage, the device effectively shuts down. Going below 1.8 V will disable the microprocessor as the brown-out detection will trigger. That means the LED won't blink any more and the device will be "dead". The battery monitoring system has some hysteresis built into it so the LED does not change its color wildly.
Thermal protection is built in the OPA567s and the microprocessor also monitors this and will indicate overheat using the red overload LED.
A mechanical switch physically disconnects both battery and one of the inputs so no current can flow when the device is unpowered.
The bandwidth is approximately 10 kHz, above that the THD starts to rise. The maximum rated current is 100 mA. With active cooling and different switch, up to 1 A will be reasonable. Minimum current range is 1 µA because the switch has 6 poles and I am using decadic ranges. Lower current ranges will be also reasonable but the input bias current of INA333 is 70 pA so 100 nA would be the lowest high-precision range. Remote (4-wire or Kelvin) sensing is enabled by a mechanical switch and is there to compensate for cable resistance. The quiescent current is 16.2 mA so the maximum runtime on 650 mAh cell is 40 hours. But it will depend on how high the measured current is. The device must supply the same amount of current as is being measured so at the maximum (100 mA), the runtime on single charge would be about 5.5 hours.
If you are interested in the design files, all resources can be found on my Google Drive.