Simple five and a half digit voltmeter with
three gears: 1.8V
18V
180V
First look at the power supply part:
it is powered by a 530mA small battery, and RT9193 low voltage difference LDO is used to convert 3.3V to power the MCU. Its quiescent current is only 70uA, and the maximum output current can reach 300mA. .
Use 0.5% accuracy TL431 and ICL7660 to form a positive and negative 2.5V voltage to power the ADC. Because the negative voltage is to be measured, an ICL7660 is needed. The TL431 is connected like this to solve the problem of insufficient output current. If there is no ICL7660, you can also use the beggar version of the negative voltage. Voltage generation circuit, mentioned in "Teacher Tang Talks about Electric Competition", link: https://www.bilibili.com/video/BV1VW4y1z7pn/?share_source=copy_web&vd_source=21af5b42fe5e2354ac6021260250b2f1
Next, let’s look at the voltage attenuation circuit:
using 4052 dual four-way Analog switches and relay shifting are cheap and easy to use. They support 3.3V power supply, but their analog on-resistance is slightly larger, but it is negligible compared with Mohm.
Using a 9M resistor and a 100K resistor to divide the voltage at 180V, the input resistance is 9.1M, which is attenuated 90 times; using a 9M resistor and a 1M resistor to divide the voltage at 18V, the voltage is attenuated 10 times, and the input resistance is 10M. Why is the input resistance not uniform? As for 10M, because it is a hassle, you have to find the corresponding resistor, or you can use a standard voltage divider resistor network. I use a 0603 package with a 0.1% precision 25ppm resistor which is not bad. Each 0603 resistor can withstand a voltage of 75V. This This is also the reason why a 9M resistor is not used.
BAV199 is used for clamp protection in the basic gear, and its reverse leakage current is only 5nA, which is a relatively excellent parameter. Many of them are uA level, and their equivalent resistance is not that large, so they will inevitably form a voltage divider with the previous 30K resistor, causing a voltage drop on the 30K resistor, which is a very bad thing.
The latter stage directly uses a high-precision operational amplifier followed by low-pass and then sent to the ADS1118. Nowadays, domestic operational amplifiers are becoming more and more powerful. The parameters of Runshi Technology's RS8551 are very powerful. Its Vos is only 1uV (maximum 5uV), which is very suitable for high-precision instruments.
ADS:
The 16-bit ADS1118 is used, and its typical error is only 0.25LSB. This parameter is very strong. If the PGA uses 2.048V, the LSB is 62.5uV, which is the minimum resolution, but in fact it is only 15bit because the other half is used to measure the negative voltage. , because each input channel has a protection diode, the negative voltage will be clamped when measuring negative voltage, so the ADS1118 uses dual power supplies. When it is turned on, channel one is calibrated to see how much the voltage of DGND is compared to VDD-VEE ( Theoretically it is half), this voltage is regarded as "0"V, so if it is higher than DGND, it is measured as positive voltage; if it is lower than DGND, it is measured as negative voltage.
The main control uses STM32F103C8T6, which is easy to adjust and writing code is more convenient.
Next, look at the test data:
Because our DC power supply can only output a minimum voltage of 47mV, it can only test voltages above 50mV.
100mV test:
1V test:
20V test:
30V test:
40V test:
50V test:
60V voltage test:
Finally Thanks to Jialichuang for sending me the components, which allowed me to successfully complete this small project.