SBAA791 June 2018 AMC1305M25
According to the specification, AMC1305M25 has an input range of ±312.5mV. When the input reaches 312.5mV, the output is nearly 100% “1” bit stream; when the input is -312.5mV, the output is nearly 100% “0” bit stream. Additionally, according to F28377S specification, SDFM module supports two data formats: The first is a 32-bit binary signed-complement code and the second is a signed 16-bit binary signed-complement code. In SINC3 filter mode, the final data needs to be shifted if 16 bits are used to represent the data due to the data bits being too wide. Incorrect shifts and conversions can result in a mismatch between the final digital result and the analog input amplitude. In this chapter, SINC3 filter is used as an example to calculate the input voltage from a 32-bit data format and a 16-bit data format, respectively.
Figure 3-5 SDFM output range for
different combinations of filter order and oversampling rate corresponds to
formula (8)
Figure 3-6 AMC1305 sampling front end
connects to 100mV voltageTo verify the input analog magnitude versus digital magnitude, a 100mV voltage source is connected at the input of the AMC1305M25 for reference. The actual input voltage magnitude is measured to be 100.12mV using a six-digit half-voltmeter.
Use 32 BIT binary signed-complement code (SDDPARMx.DR = 1):
In the sample code, SDFM filter output value can be obtained by reading the SDFM1_READ_FILTER1_DATA_32BIT register. When calculating the input voltage, first convert the data obtained by the register from the binary complement code to decimal, and then use the formulas (8) and Figure 3-5 to obtain corresponding analog input amplitude of AMC1305M25. The results calculated using formula (8) at different oversampling rates are compared in Table 3-1. The input amplitude is 100mV ±2% in all cases, consistent with the voltmeter test results.
R/OSR: oversampling ratio; OUTDecimal: Decimal expression of filtered data, binary signed bit complement code, MSB is a sign bit; O: SINC filter order; VClipping: Maximum analog front-end dynamic range
It can be known from AMC1305M25 specifications that VClipping = 312.5mV;
The following data is obtained by averaging 1024 samples of the digital signal using SINC3 filters with different oversampling rates:
| Filter Type | Convert binary complement code to decimal | Vin calculation | |
|---|---|---|---|
| SINC3 R/OSR64 | 84024 | 262144 | 100.16mV |
| SINC3 R/OSR100 | 319780 | 1000000 | 99.77mV |
| SINC3 R/OSR128 | 673092 | 2097152 | 100.14mV |
| SINC3 R/OSR256 | 5361067 | 16777215 | 99.86mV |
Use 16 BIT binary signed-complement code (SDDPARMx.DR = 0):
When 16 BIT data is used, the filter output data is stored in register SDFM1_READ_FILTER1_DATA_16BIT. Due to data overflow, F28377S automatically performs a right shift on the filtered results. Figure 3-7 gives the corresponding number of shifts for different SINC filters and oversampling rates.
Figure 3-7 Number of shifts when SINC
filter supports 16BIT output at different oversampling ratesCombining Figure 3-7 and the formula (8) yields the formula (9), which corresponds the filter output result to the input analog magnitude using the 16 BIT signed-bit configuration.Table 3-2 shows the correspondence between the digital output of F28377S and the analog input amplitude during actual test, with an input of 100mV.
S: The number of register shifts corresponding to different OSRs in Figure 3-8
| Filter setting | Convert binary complement code to decimal | 2 | Vin calculation | |
|---|---|---|---|---|
| SINC3 R/OSR 64 | 5244 | 16 | 262144 | 100.02mV |
| SINC3 R/OSR 100 | 10003 | 32 | 1000000 | 100.03mV |
| SINC3 R/OSR 128 | 5244 | 128 | 2097152 | 100.02mV |
| SINC3 R/OSR 256 | 10495 | 512 | 1677216 | 100.09mV |