(solved) Arduino Due insists on CPOL=0, CPHA=0

[system]

I need it at SPI Mode 2, which is CPOL=1, CPHA=0. As I understand it, it means SCLK is high initially, and every time it dips down, that's when the device should capture data. Just to reassure everyone that my code sets the SPI to MODE2, here's my code:

#include <SPI.h>
#define ADC_CS 4

void setup() {
  // put your setup code here, to run once:

  SPI.setDataMode(SPI_MODE2, ADC_CS); // page 16
  SPI.setBitOrder(MSBFIRST);
  SPI.setClockDivider(8.4); //ADC: 20MHz max, DAC: 50MHz write;8MHz read

  SPI.begin(ADC_CS);
  SPI.transfer(4, 0x10, SPI_CONTINUE);
  SPI.transfer(4, 0x11);
  
  SPI.setDataMode(SPI_MODE0, 10);
  SPI.begin(10);
  SPI.transfer(10,0x00);
}

void loop() {
  // put your main code here, to run repeatedly:

}

But see the screenshot. 0x10, 0x11 only show up when I tell the logic analyzer to interpret it as CPOL=0, CPHA=0. It doesn't seem to matter whether I put SPI.begin() in front of the settings or after.

AS a bonus, I set setDataMode to SPI_MODE0, just in case MODE0 meant "SCLK high" to the Arduino and the behaviour, as you can see, didn't change.

EDIT: I'm running Arduino 1.5.5, but I find it hard to believe that such an obvious flaw would be fixed in 1.5.6 if it hasn't been found by now...

[system]

#include <SPI.h>
#define ADC_CS 4

void setup() {
  // put your setup code here, to run once:
  SPI.begin(ADC_CS);
  SPI.setDataMode(SPI_MODE2, ADC_CS); // page 16
  SPI.transfer(4, 0x10, SPI_CONTINUE);
  SPI.transfer(4, 0x11);
  

  SPI.begin(10);
  SPI.setDataMode(SPI_MODE0, 10);
  SPI.transfer(10,0x00);
}

void loop() {
  // put your main code here, to run repeatedly:

}

Just to check, I moved the SPI settings line to after SPI.begin(), and the result is still exactly the same.

[SurferTim]

I see you monitoring only one slave select (enable). Where is the other slave select output on the logic analyzer? Put one on D4 and the other on D10.

edit: Disable both SPI outputs before starting any SPI device.

pinMode(4, OUTPUT);
digitalWrite(4, HIGH);

pinMode(10, OUTPUT);
digitalWrite(10, HIGH);

I'm not certain about the Due SPI. The SPI mode (in you case the clock polarity) should be changed before the slave select is activated (taken LOW), and I don't see that on your logic analyzer. Maybe it is a bug.

[system]

OK, I put your code before any SPI initialization happens.

#include <SPI.h>
#define ADC_CS 4

void setup() {
  // put your setup code here, to run once:
  pinMode(4, OUTPUT);
  digitalWrite(4, HIGH);

  pinMode(10, OUTPUT);
  digitalWrite(10, HIGH);
  
  SPI.begin(ADC_CS);
  SPI.setDataMode(SPI_MODE2, ADC_CS); // page 16
  SPI.transfer(4, 0x10, SPI_CONTINUE);
  SPI.transfer(4, 0x11);


  SPI.begin(10);
  SPI.setDataMode(SPI_MODE0, 10);
  SPI.transfer(10, 0x00);
}

void loop() {
  // put your main code here, to run repeatedly:

}

The "device" on pin 10 is just there so that I can see if SCLK behaves differently as I change from SPI_MODE2 to SPI_MODE0. It doesn't. Here's another screenshot....

[SurferTim]

I will repeat what I posted above as an edit.

I'm not certain about the Due SPI. The SPI mode (in you case the clock polarity) should be changed before the slave select is activated (taken LOW), and I don't see that on your logic analyzer. Maybe it is a bug.

edit: I do not see the slave select going LOW. Is that channel 0?

[system]

My fault....

  SPI.setDataMode(SPI_MODE2, ADC_CS); // page 16

The correct syntax is:

  SPI.setDataMode(ADC_CS, SPI_MODE2); // page 16

It works perfectly now this is the logic analyzer set to SPI_MODE3, and my code going through all the different SPI_MODEs.

[SurferTim]

I just saw that myself. slave select is the first parameter.