
Special Notes on Connectivity
7
SPRUI76A – March 2017 – Revised January 2019
Copyright © 2017–2019, Texas Instruments Incorporated
Delfino™ TMS320F28379D controlCARD R1.3
4.3
Evaluation of the Analog-to-Digital Converters (ADCs)
When using the F28379D on-chip ADCs, there are some useful guidelines to help realize the performance
numbers listed in the
TMS320F2837xD Dual-Core Delfino™ Microcontrollers
data sheet. This is especially
true for the AC parameters such as: SNR, THD, and SINAD. Furthermore, it can also be shown that there
is a direct correlation between the SNR of the ADC results and the spread of ADC codes seen for a DC
input. These tips will improve the range and standard deviation of a DC input as well. Finally, while topics
addressed will be with respect to the controlCARD, they are also applicable to other implementations
using the F28379D MCU.
On-board resistors and capacitors:
By default (
), all inline resistors to the ADC pins are a
simple 0-
Ω
shunt and all capacitors to the ground plane are not populated. While this circuit can be used
to supply the ADC inputs with a voltage, likely both the resistor (R) and capacitor (C) will need to be
populated based on the voltage source's characteristics. Referring to the
, the ADC input
has its own RC network made up of the internal sample and hold capacitor, switch resistance, and
parasitic capacitance. By changing the inline resistance and parallel capacitor, the input circuit can be
optimized to assist with settling time and/or filtering the input signal. Finally, it is recommended in general
to use either Negative-Positive 0 PPM/°C (NPO) or Ceramic On Glass (COG) as these have better
stability over temperature and across input frequencies than other types of capacitors.
Figure 4. Partial Schematic Showing Default R and C values
Voltage source and drive circuitry:
While the on-chip ADCs are configurable 12-bit/16-bit architectures
(4096/65536 distinct output codes when converting an analog signal to the digital domain), the translation
will only be as precise as the input provided to the ADC. The typical rule of thumb when defining the
source resolution to realize the full specification of an ADC is to have a 1-bit better source than the
converter. In this case, that would mean that ideally the analog input should be accurate to 13 bits for a
12-bit mode and 17 bits for a 16-bit mode.
Typically, voltage supplies or regulators are not designed to be precise, but rather accommodate a wide
range of current loads within a certain tolerance, and for this reason, are not ideal to show the
performance of a higher bit ADC, like the one on the F28379D. This does not take into account that many
times the supply in question is providing the main voltage to power the MCU itself, which also introduces
noise and other artifacts into the signal.