ADS1247: Problem with getting the Ref voltage output

I have fixed the problem and I am now reading the following on the ADC:



It is not the 2.048V I was expecting but close. The output voltage on the pin is 1.8V (Bit low)

I have done more test.

Using a 3.2V PSU to power the ADC, and with register 0x02 set as 0x30 and 0x03 as 0x00 I obtained the following:

Since the above where not optimum but acceptable to do further test, I tried to put another PSU as an input using a voltage divider. Two input circuit were tested with the following results:

I then decided on test 2 to remove PSU and short the input but did not get the expected 0V. Can you advise on those results?

Thank you

Thanks Joseph.

I have tried the same divider with VREFOUT on the resistor and I am getting

Hi,

Can you help clarifying the below:

1. Is it possible to get the ADS1247 to work on unipolar 0-3.3V with internal clock like I am trying to do?

2. Do I need to balance the inputs like suggested in the datasheet unsing 1M pull-up, 1M pull down?

Test conducted today with the below circuit: (with and without the 1M on the input)

With the 1M, I obtained 0x7FFFFF when reading conversion and nothing was plugged in while when input were shorted, I obtained 0x46B4 but expected 0x3FFFFF as it would be mid-range. Same test was done without the 1M, and when input were shorted, I obtained a value constantly changing between 0x4F1 and 0xFFFF58.

I also tried to put 1M between the inputs when having pull up and pull down and obtain 1.1V as expected. When replicating the test with 10ohms, I obtained 3.5mV instead of 16.5uV (if this makes sense). I am not sure why the device cannot work on small voltage

I am using a pic16LF1788 to communicate with the device.

Thanks for your help

Fabien

Fabien,


First, I would like you to read out all of the registers and report them back. I'd asked that earlier, but you only gave me two of the registers. The point is that maybe something else is set other than what you expect. I would also like you to include the calibration registers just to make sure the values close to expected. It would also be better if you read the calibration register instead of telling me what you've written, just in case there was a mistake in your setup.

The ADS1247 should work with a unipolar 3.3V supply with an internal clock. It is a common configuration. However, I thought you were using a 5V supply. In your previous posts, they show a 5V source for the input and I had thought that you had used the same for the analog supply. I don't think this would normally be a problem, but if you accidentally over-voltaged the input pin, then you might damage the device.

The 1M pull-up/pull-down resistors are typically used for thermocouple biasing. I don't remember that being mentioned in the datasheet, but they give thermocouples a DC operating point at mid supply. If you're not measuring a thermocouple, I'd leave these resistors out. If you have the these resistors in, and the inputs are floating, these pull-up/pull-down resistors will pull apart the inputs and give a reading of 7FFFFFh. If the resists are in, and the inputs are shorted, the input reading should basically be 0 (not including the offset and noise). If you are getting between 4F1h and FFFF58h this translates to +1265 to -168. This appears to be a lot of noise, but it depends on your data rate (I assume that your PGA is 1).

If you're having trouble measuring the 1M/10Ω divider. I'd suggest a different measurment to start. Replace those resistors with a 1k/1k divider and make sure that value is correct. Right now you're just testing the device in a gain of 1, and you should start with larger voltages. With smaller voltages, the errors that you might see start to add up fast. Leakage and input biasing currents have a large effect in the measurement, especially if the voltages are very small.

I'd note that the input measurements of 351Eh and 37BBh translate to about 3mV, which is something you should be able to see with a precision multimeter. Note that making measurements with the multimeter and taking measurments with the ADC at the same time may cause noisier ADC readings. The ADC works with periodic capacitive sampling which can be disrupted by the multimeter's measurements.


Joseph Wu

Register read (obtained using 0x20 command)

I use a 3.3V supply with the last schematic shown. AVCC and AVDD are both 3.3V and last time I used an external PSU at 5V for a voltage divider purpose and have the ability to read small voltage.

The intent is to be able to measure voltage differential and surface temperature using a thermocouple which based on the datasheet is the reason for the 1M pull up/pull down addition.

The code used is such as:

void main (void)

{  

   //Init SPI

   //Reset Device

   ResetSPIDevice();

   InitialiseSPIDevice();

   SetContinuousConversion();

   while (1) {

       if (!DRDY) {

           CheckLastConvertedValue(Data);

       }

   }

}

void ResetSPIDevice(void)

{

   //IReset SPI Device

   SPI_CSN = 1; // CSN high

   SPI_SCK = 0; // SCK low

   SPI_START = 0;

   IOCIE = 1; //interrupt on change for DRDY

   WaitForDRDY ();  

   SPI_CSN = 0; // CSN low

   Wait(80); //Wait 80ms

   SPI_START = 1;

   WaitForSDI = GetTimer1CounterNow();

   //Delay waiting for device to be ready

   Wait(10); //Wait 10ms

   //Reset Device

   WriteSPI(0x06); //Write CMD With Register Nbre

   Wait(10); //Wait 10ms

   SPI_START = 0;

   Wait(10); //Wait 10ms

   SPI_CSN = 1; //CSN high*/

}

void InitialiseSPIDevice(void)

{

   uint8_t ReadValues[16];

   uint8_t i;//flag when reading configuration

   ResetDRDY();

   SPI_CSN = 0; // CSN low

   Wait(10); //Wait 10ms

   SPI_START = 1;

   Wait(10); //Wait 10ms

   //Configure the Device

   WriteSPI(0x40); //Write CMD With Register Nbre

   WriteSPI(0x03); //Number of register to write 4

   WriteSPI(0x01); //MUX Control Register AN0 POS and AN1 NEG

   WriteSPI(0x00); //BIAS No BIAS set

   WriteSPI(ControlRegister); //MULTIPLEXER Control Internal Oscillator, Internal Ref ON during Conversion, Internal Ref selected, Monitor Control

   WriteSPI(MuxCalRegister); //Control PGA=1 DR =1000SPS

   //Check by Reading

   WriteSPI(0x20); //Write CMD With Register Nbre

   WriteSPI(0x0E); //Write CMD With Register Nbre

   for(i=0;i<15;i++)

       ReadValues[0] = ReadSPI(0xFF);

   //Delay waiting for device to be ready

   Wait(80); //Wait 80ms

   //Send SYNC

   WriteSPI(0x04);

   WriteSPI(0x04);

   Wait(10); //Wait 10ms

   SPI_START = 0;

   Wait(60); //Wait 10ms

   SPI_CSN = 1; // CSN high

   Wait(80); //Wait 80ms

}

/** @brief SetContinuousConversion */

void SetContinuousConversion(void)

{

   SPI_CSN = 0; // CSN low

   Wait(10); //Wait 10ms

   SPI_START = 1;

   Wait(10); //Wait 10ms

   //Send Enable constant conversion

   WriteSPI(0x14); //Write CMD With Register Nbre

   Wait(10); //Wait 10ms

}

/** @brief CheckLastConvertedValue */

void CheckLastConvertedValue(uint8_t *dataStorage)

{

   SPI_CSN = 0; // CSN low

   Wait(10); //Wait 10ms

   WriteSPI(0x12); //Write CMD With Register Nbre

   dataStorage[2] = ReadSPI(0xFF);

   dataStorage[1] = ReadSPI(0xFF);

   dataStorage[0] = ReadSPI(0xFF);  

   Wait(10); //Wait 10ms

   //CS LOW, START High

   SPI_CSN = 1; // CSN high

}

I also done the test with the 1K resistors as mentioned above and got the following ( 3 x 1K with the positive and negative inputs used to measure the voltage drop across the middle resistor and  analogue supply (3.16V used)

Experiment

With an analog supply of 3.16V, and a gain of 1 using the above setup, the value expected was 1.05V

Samples read in 0.2s intervals (5SPS)

Fabien,

I'd forgotten the attachment....

Joseph Wu

Test1.xlsx

Thanks for the feedback.

If I do internal checks like Vref/4 using the 0x3D settings in register 0x02 and have value not as stable as you. What would be the source of the problem?

When I checked that register today, the value returned as not as stable as the one you obtained.

Fabien

Hi,

I have checked my supply and had to add capacitor on both analog and digital supply. My Vref is now more stable and is as shown below:

 I was however expecting value nearer to 2.048V. Is that going to cause problem when measuring thermocouple.

Also I am supplying the device using 2x alkaline battery on which I see ripple oscillating within 200uV. This surely would affect my reading when looking at >10uV.