Recently, I built an R2R ladder resistor DAC circuit using resistors with a precision of one thousandth. Through testing, the circuit achieved a linear accuracy of 12 bits. Now, I ponder a question: if I reduce the order of R2R, for example to 12 bits, what would its linear accuracy be? Below, I will attempt to answer this question through experiments or corresponding data.
▲ Figure 1.1.1 Test Circuit Schematic
▲ Figure 1.1.2 Test Circuit PCB2. Test Results
By programming, I measured the DAC channels from 1 to 15, corresponding to the output voltage for each DAC output. Using the DM3068, I recorded up to 128 measurement data points for each channel. Although this process was completed automatically by the program, it still took me nearly half an hour of measurement time this morning. For each set of data, I calculated the linear error curve. The results show that as the number of output channels increases, the linear error curve gradually decreases. When the output channels reach 15, which is the longest data set in the chart, the corresponding error is the smallest.
▲ Figure 1.2.1 Different Channels and Error CurveHowever, when plotting the maximum error and standard deviation from the measurements of different DAC channels, the situation changed. The maximum value of linear error is represented by the cyan curve, while the root mean square of linear error is shown below. It can be seen that the linear error reaches its peak when the number of DAC channels is 11, and then decreases as the number of channels increases. The root mean square of the error presents a stable trend, meaning that regardless of the number of output channels, the root mean square of the DAC linear error remains around 0.5mV. This result was indeed unexpected.
▲ Figure 1.2.2 Maximum Linear Error and Standard Deviation for Different Channel Counts※ Conclusion ※
This morning, I measured the linear error of the 15-bit resistor R2R DAC made a few days ago. Although all the resistors are of one-thousandth precision, the root mean square of the corresponding linear errors varies little with different output channel counts. This is the result of experimental measurement; I will conduct a theoretical analysis when I have time later. Today is the day before the Qingming Festival, and I still have classes this morning, so I will head to class shortly.
15-bit R2R Resistor Ladder DAC Performance Test: https://zhuoqing.blog.csdn.net/article/details/137032684