2nd Generation (TC3xx) 32-bit TriCore™ AURIX™ microcontrollers

About this document

Scope and purpose

The Infineon TC3xx microcontroller family provides EDSADC (Enhanced Delta-Sigma ADC) as well as EVADC (Enhanced Versatile ADC).

The delta-sigma technology is one of the most widely used analog to digital conversion methods. It samples the analog input signal at a very high frequency, which is normally hundreds of times faster than the digital result output rate. Each digital output is the average value of several samples. In this document we introduce the basic features of EDSADC and its typical configuration in automotive applications.

This document describes the EDSADC electric parameters, components, digital connection with other peripherals, calibration methods, and result handling.

Intended audience

This document is intended for hardware and software engineers who are designing EDSADC applications.

EDSADC features

The TC3xx family of products has different numbers of EDSADC modules (up to 14x).

Unlike EVADC, each EDSADC module is equipped with 1 pair of differential input channels (single ended connection is also possible). Some EDSADC modules, but not all, have a multiplexer to allow switching between different pins.

As with other peripherals, the EDSADC module is connected to the cores via the SPB bus. EDSADC in general is able to generate a higher Effective Number of Bits (ENoB).

The main EDSADC features are:

  • Switched-capacitance input structure

    • Input voltage range 0…5 V
    • Common mode voltage: VAREF / x (x = 2, 4, 8), connection selectable for each pin (input pin, VAREF, VAGND or VREFX)
    • Programmable gain (1:1 / 1:2 / 1:4)
    • DC offset <5 mVDC, gain error < ±0.2% (calibrated, on product level)
    • Equivalent input impedance 500 kΩ typ. (for gain = 1, fMOD = 26.67 MHz)
  • On-chip modulator, single-bit second-order feed-forward modulator, sample frequency 16 … 40 MHz

  • Options to connect external standard DS modulators

    • Selectable data stream inputs
    • Selectable DS clock input or output
  • Demodulator (concatenated hardware filter stages)

    • Configurable CIC filter with decimation rates of 4…1024 (with pre-CIC enabled)
    • Overshoot compensation filter, optional
    • FIR filter with 8 coefficients (10-bit) with decimation rate 2 (FIR0)
    • FIR filter with 28 coefficients (10-bit) with decimation rate 1 or 2 (FIR1)

Pass Band:

Table 1. EDSADC pass band

Passbandcondition

fd/3

FIR1 decimation rate = 2:1

fd/6

FIR1 decimation rate = 1:1

Pass band ripple dfPB < ±1%

Stop band attenuation:

Table 2. EDSADC stop band attenuation

BandwidthAttenuation

0.5…1 × fd

> 40 dB

1…1.5 × fd

> 45 dB

1.5…2 × fd

> 50 dB

2…2.5 × fd

> 55 dB

2.5…OSR/2 × fd

> 60 dB

  • Optional high-pass filter for DC compensation (f-3dB = 10-5 × fd), configurable

  • Hardware offset/gain calibration and compensation

  • Support for resolver applications

    • Carrier signal generator (differential sine output). An additional external analog circuit is needed to transform the digital output from carrier signal generator into a sine wave for example.
    • Signal evaluation including rectification, delay compensation and carrier elimination.
  • Automatic limit checking

    • Two-level boundary comparator.
    • Separate indication signals (hardware signal to notify comparison result: above the upper limit, within the defined band or below the lower limit).

In case TJ ≤ 150°C and fPB = 30 kHz, the SNR (Signal to Noise Ratio) is at least 80 dB, corresponding to an ENoB of approximately 13 bits.

Modulator and filter chain

The modulator and demodulator (also called the filter chain) are the kernel parts of the EDSADC. The on-chip filter stages generate digital results from the selected modulator signal.

Figure 1. EDSADC block diagram


../figures/image5.png

If necessary an external modulator can also be selected and the digital input will directly connect to the main filter chain. A typical use case for an external modulator can be for the purpose of galvanic de-coupling.

Figure 2. Main filter chain (demodulator)


../figures/image6.png

Filter chain

The filter chain configuration determines the attenuation and group delay properties of the filter. The configurable CIC (Cascaded Integrator Comb) filter, also known as a SINC filter, provides the basic filtering and decimation with a selectable decimation rate.

There are 2 FIR filters:

  • FIR0 with a fixed decimation rate of 2

  • FIR1 with a configurable decimation rate (1 or 2)

The filters allow effective signal shaping by attenuating the upper frequencies of the signal spectrum. The digital signal will go through the main CIC filter and 2 consecutive FIR filters.

The high-pass filter provides offset compensation by removing the DC component of the input signal. The Pre-CIC is provided mainly for backwards compatibility (on TC2xx, the decimation rate of FIR1 is fixed to 2, and this Pre-CIC can be optionally enabled to have the same total decimation rate).

The on-chip carrier signal generator produces a PWM signal that induces a sine signal in the excitation coil of the resolver.

Synchronization of each input signal to the carrier signal ensures correct integration of the resolver input signals. The integrator integrates the result values generated during the defined integration window by adding a configurable number of values to build the final result value. Integration can be used to measure currents through shunt resistors at defined positions in the signal waveform. It can also remove the carrier signal component in the resolver application.

Table 3. ENoB defined in data sheet (differential mode)

Bandwidth (kHz)Min SNR (dB)ENoBCondition

100

74

12

TJ ≤ 150°C

50

78

12.7

TJ ≤ 150°C

30

80

13

TJ ≤ 150°C

In case EDSADC is in single ended mode, around 6 dB (corresponds to 1 ENoB) will be lost for each case.

The following table shows the SNR as well as the input current under more cases with Gain =

  1. For other gains, the input leakage is proportional to the gain factor (linear). fPER on TC3xx is fixed to 160 MHz.

Table 4. Properties of Operating Ranges

Modulator frequency fMOD

Input current IRMS1fPB < 10 kHz2fPB < 30 kHzfPB < 50 kHz3fPB < 100 kHz3

SNR

OSR

SNR4

OSR

SNR4

OSR

SNR4

OSR

16 MHz (fPER/10)

6 µA

≥ 78 dB

≥ 267

≥78 dB

≥ 178

≥74 dB

≥ 107

≥65 dB

≥ 54

20 MHz (fPER/8)

7.5 µA

≥ 80 dB

≥ 334

≥78 dB

≥ 223

≥74 dB

≥ 134

≥68 dB5

≥ 67

26.67 MHz (fPER/6)6

10 µA

≥ 80 dB

≥ 445

≥80 dB

≥ 297

≥78 dB

≥ 178

≥74 dB

≥ 89

40 MHz (fPER/4)

15 µA

≥ 80 dB

≥ 667

≥80 dB

≥ 445

≥78 dB

≥ 267

≥74 dB

≥ 134

Note: Related to hardware characteristic for passband frequency ≤ 10 kHz the FIR1 has to be used with a decimation rate of 1:1.
1

These typical values refer to an input voltage of 5 V and the typical value of the switched capacitor:

IRMS = 5 V × fMOD × 2 × CSW. The equivalent input impedance can be calculated by 5V / IRMS.

2 Passband = fd / 6 (the decimation rate of FIR1 is set to 1 instead of 2).
3 Passband = fd / 3 (the decimation rate of FIR1 is set to 2).
4 The reachable signal-noise-ratio is limited by thermal noise within the modulator.
5 With this configuration, the signal-noise-ratio is limited by quantization noise.
6 These OSRs reference the Datasheet values.

Hardware triggers

EDSADC generates a continuous output stream once the modulator is enabled.

Hardware triggers are used for the following features:

Integration trigger

The external signal defines the integration window; i.e. the timespan during which result values are integrated. A typical usage for the integrator is to cancel the carrier signal in resolver application.

Timestamp trigger

The external signal requests the actualization of the timestamp register. This feature is necessary for software interpolation, because the EDSADC outputs a digital result at a fixed rate of fd.

Input multiplexer trigger

The external signal requests the switching of the analog input multiplexer to the next lower input, or to the defined start value, respectively.

Modulator run gate

The on-chip modulator can be controlled by an external gate signal while it is enabled. In this case it can be disabled temporarily to save power. Please note that the modulator needs some wakeup time when the gate signal changes from inactive to active. The wakeup time is documented in the user manual.

Service request gate

Service requests for the filter chain can be restricted to the high or low times of the selected trigger signal.

Note: The digital connection (hardware triggers between EDSADC and other peripherals) is referenced in the user manual.

Calibration and Result interpretation

Offset calibration and gain calibration are done by hardware. The calibration values are determined automatically by a hardware algorithm. This can be triggered initially after a reset or repeatedly during operation, for example when TJ changes more than 20°C.

The time needed for calibration depends on modulator frequency, decimation rate and gain. Please see the user manual for the calibration duration.

The digital result of EDSADC is stored in 2’s complement format with 16 digits (1 sign bit and 15 value bits).

The digital result in stored in result register RESMx.RESULT.

The result is normalized to ±25000 after reset:

  • -25000 represents -5V

  • 25000 represents 5V

  • values in between are linear

Note: Please note that this is only an indication. The exact mapping can only be determined after calibration.

If full scale is expected, to for example minimize quantization errors, it is possible to use the full range from -32767 to 32766. This is achieved by configuring GAINCORR.

As the digital result register is 32-bit wide, it is possible to put 2 results to the higher and lower halves of the result register respectively and read them out together (this is also convenient for interpolation calculations).

The digital result output rate (fd) is relatively high and therefore a hardware FIFO feature is provided to allow for a flexible read out. The maximum FIFO depth is 4.

Result FIFO

EDSADC results can be placed in FIFO and later read out (enabled by bit field DRM). Together with time-stamp feature (enabled by bit TSM), the following modes are supported:

Table 5. DRM = 1 (FIFO disabled)

Read mode

FIFO disabled

This mode is similar to TC2xx, which has no FIFO.

Content in result register EDSADC_RESMx

TSM = 0

DRM = 1

No FIFO

[31...16] bits

0

[15...0] bits

Result

TSM = 1

DRM = 1

No FIFO

[31...16] bits

Timstamp

[15...0] bits

Result

DRM = 0 or 2 means the FIFO is enabled (DRM = 3 is reserved and shall not be used). Please note the FIFO structure is transparent to the user and the user must read the EDSADC_RESMx result register several times to fetch contents stored in FIFO. The FIFO filled level is indicated by the bit field EDSADC_RFCx.FILL.

FIFO overflow and underflow errors are indicated by the bit fields EDSADC_RFC.WRERR and RDERR respectively.

Table 6. DRM = 0 (lower 16 bits in result register utilized)

Read mode

FIFO enabled

DRM = 0

Content in result register EDSADC_RESMx

TSM = 0

DRM = 0

4th read

[31...16] bits

Sign extention

[15...0] bits

Result 3

3rd read

[31...16] bits

Sign extention

[15...0] bits

Result 2

2nd read

[31...16] bits

Sign extention

[15...0] bits

Result 1

1st read

[31...16] bits

Sign extention

[15...0] bits

Result 0

TSM = 1

DRM = 0

4th read

[31...16] bits

Sign extention

[15...0] bits

2nd result after timestamp

3rd read

[31...16] bits

Sign extention

[15...0] bits

1st result after timestamp

2nd read

[31...16] bits

Sign extention

[15...0] bits

Result before or at timestamp

1st read

[31...16] bits

Shall be ignored by user

[15...0] bits

Timstamp

Table 7. DRM = 2 (both lower and higher 16 bits in result register utilized)

Read mode

FIFO enabled

DRM = 2

Content in result register EDSADC_RESMx

TSM = 0

DRM = 2

4th read

[31...16] bits

Result 7

[15...0] bits

Result 6

3rd read

[31...16] bits

Result 5

[15...0] bits

Result 4

2nd read

[31...16] bits

Result 3

[15...0] bits

Result 2

1st read

[31...16] bits

Result 1

[15...0] bits

Result 0

TSM = 1

DRM = 2

4th read

[31...16] bits

6th result after timestamp

[15...0] bits

5th result after timestamp

3rd read

[31...16] bits

4th result after timestamp

[15...0] bits

3rd result after timestamp

2nd read

[31...16] bits

2nd result after timestamp

[15...0] bits

1st result after timestamp

1st read

[31...16] bits

Result before or at timestamp

[15...0] bits

Timstamp

Revision History

Document revision

Date

Description of change

All

First version.

V1.1

Added chapter Result FIFO. Updated Table 4 according to User Manual V2.0.

V1.2

2024-04-17

Template update; no content update.

1

These typical values refer to an input voltage of 5 V and the typical value of the switched capacitor:

IRMS = 5 V × fMOD × 2 × CSW. The equivalent input impedance can be calculated by 5V / IRMS.

2 Passband = fd / 6 (the decimation rate of FIR1 is set to 1 instead of 2).
3 Passband = fd / 3 (the decimation rate of FIR1 is set to 2).
4 The reachable signal-noise-ratio is limited by thermal noise within the modulator.
5 With this configuration, the signal-noise-ratio is limited by quantization noise.
6 These OSRs reference the Datasheet values.