TMP126-Q1: Integration trouble, reading strange results with two separate devices.

Hi Ren- thanks for the quick answer. I apologize for needing quite a bit of handholding.

I had the 10k resistor inline with the MOSI line coming from the arduino as the diagram requests- but left out the capacitor between the voltage/ground and the resistor between alert/voltage as I didn't think I needed either of them for initial testing.

I've dropped down the frequency to 500kHz and changed the two write values to 0xFF as you recommended which all makes sense. It's reacting better to that now and not just feeding me garbage E0 values- but still giving me strange reading. So if I write 0xFF it'll read back FF on both bytes as expected when unplugged but when plugged in it'll read 7F and FF.

The SIO is quite tough to probe into given the location with my setup, but with an oscilloscope hooked up to the MOSI and MISO pins it produces the following graphs:

MOSI signal unplugged from thermistor:

MOSI signal plugged into thermistor:

MISO signal unplugged from thermistor:

MISO signal plugged into thermistor:

All looks healthy signal wise, just not getting the expected input.

Here the code in txt format if that's better for you:

#include <SPI.h>

unsigned char byte1;
unsigned char byte2;
//PINOUTS
#define CSpin 10 //ss
#define DATAout 11 //MOSI
#define DATAin 12 //MISO
#define SCLK 13 //SCLK
#define Alert 8 //Status 1

#define CSpin 10

void setup() {
// put your setup code here, to run once:
pinMode(CSpin, OUTPUT); //set CS/SS as output
SPI.begin(); //init SPI library
Serial.begin(9600);
}

void loop() {
// put your main code here, to run repeatedly:
SPI.beginTransaction(SPISettings(500000, MSBFIRST, SPI_MODE0)); //SPI settings (speedMaximum, dataOrder, and dataMode), Nano is 16MHz max
digitalWrite(CSpin, LOW); //set CS/SS LOW
SPI.transfer(0x01);
SPI.transfer(0x00); //address
byte1 = SPI.transfer(0xFF);
byte2 = SPI.transfer(0xFF);

digitalWrite(CSpin, HIGH);
SPI.endTransaction();

float t126 = (((int16_t) byte1 << 8 | byte2) >> 2) * 0.03125;

delay(1000);

Serial.print("byte1:");
Serial.print(byte1, HEX);
Serial.print(", byte2:");
Serial.print(byte2, HEX);
Serial.print(", temp:");
Serial.print(t126);
Serial.println("C");

}

Could you please confirm that you receive 0x0126 when sending 0x010C? This is the Device ID register, and should always report the same value regardless of configuration.

thanks,

ren

Err, nope, still just getting FF for both bytes even when changing it to 0x010C and verifying/uploading.

I can rewire everything, but I'm pretty sure everything is correct hardware-wise.

The 0xFF response implies TMP126 did not respond at all. Please use your oscilloscope to look for ringing. You will need to "zoom in" to a much smaller time scale to see the clock cycles. You will likely need to add series resistance (~50ohms) to stop the ringing caused by your wiring. Since TMP126 supports 10MHz, it will not filter high-frequency noise, and will be confused by all the activity it perceives. 

ren