ADS8681: Reading conversion data from ADC

Hi Hiep,

The pseudo-code looks right.  Can you send over a screen shot as well as let us know what range you have selected and how you have the reference configured (internal or external)?

Hi Tom,

     I'm writing RANGE_SEL_REG(0x14) = 0xb, so I'm using "1.25 × VREF" and "ADC internal reference"=Enabled.  Hence, the AIN range = 0 to 5.12V

Also,
1) Is it required to send 4 bytes NOP command (0x0 0x0 0x0 0x0) to get the Conversion data?  
2) Do we need to pull CONVST/CS pin high before the ADC chip send out the converted data to SDO line?

Thanks

Hi Hiep,

Can you provide the timing screen shot I asked for?

This screenshot only has 2 bytes NOP (0x0, 0x0) to read from SPI after CONVST/CS signal pulled low but the 16-bit conversation data is still 0xFFFF.  

Also, I verified when I set just RANGE_INCL bit[8] in DATAOUT_CTL_REG Register(10h), I do see my "ADC input range" in bits[15:12] of the output data so this confirms that 0xFFFF are the actual conversion data (bits 31:16) from ADC chip.

0xFF 0xFF 0xB0 0x00

What is the yellow trace above?

That's the CONVST/CS signal for ADC.

OK - thank you.  So what is the clock and data while CONVST/CS is high in that capture?  Here is an example of a command that includes data on SDO_0:

And here is another just reading the conversion results with a 16-bit NOP command:

Can you get some separation in the traces from your capture and possibly zoom in a bit?  The captures above are with single SDO, range set to 0.625*VREF (bipolar) and a 1V DC input on AIN.

hi Tom,

    That clock and data is for another SPI device that's on another Chip Select pin (nothing to do with ADS8681).  I don't have access to the traces anymore.

    From your 2nd plot, the 0xB295 ADC conversion data appears to be from 12-bit ADS7038.
    Do you or any TI support SME have access to 16-bit ADS8681?

Thanks.

Hi Hiep,

Don't be distracted by the file name in the screen shot.  That is 16-bit data from the ADS8681.

Ok thanks for clarifying.

a) On your first capture, you're sending 0xD00E0000 (4 bytes command).  Is that needed before reading conversion result?

b) Could you capture the conversion data with range set to "1011b = 1.25 × VREF" ?

For a) No, that command is not needed.  For b) sure thing!  Let me know how you have the rest of your registers set as well.

These 2 registers are the only ones I set.  Everything else is default.

RANGE_SEL_REG Register (address = 14h)     = 0xB  (to set 1.25 × VREF )
DATAOUT_CTL_REG Register (address = 10h) = 0x10 (so I can see 4-bit input range setting (RANGE_SEL[3:0]) in SDO-x output stream)

Here you are:

I included SDO_2 with this capture so that you can see I'm getting 0xFFFF there, like you saw earlier.  The actual conversion results of 0x32A8 are on SDO_1.  This is with a 1V DC source applied to AIN.

Hi Tom,

     We only have single SDO (which is the MISO line) and no changes to SDO_CTL_REG register (0xC).  How are you getting data 0xFFFF on one SDO and 0x32AB on another SDO output?  Are you suggesting I need to set SDO_CTL_REG register (we didn't have to touch this register in our previous PCB)?

Thanks.

Sorry for the confusion, I thought this was both SDOs from the ADS8681 - I looked again and it seems like MISO and MOSI:

So with single SDO default setting, are you getting ADC conversion data similar to 0x32A8 for 1V DC source?  Thanks.

Yes, that is correct.

hi Tom, after pulling CS pin low, how long do we have to wait before sending NOP command to read conversion data?

That would be the tSU_CSCK parameter, see section 6.8 on page 9 of the datasheet.

Hi Tom,

     Could you post the sample code that reads ADC conversion result by using just 2-byte NOP command (0x0000) that you had working for ADS8681?

     It's only working for me when I first 0xD00C0000 in frame F then 0x0000 in frame (F+1)

Thank you.

I will try.

Thank you.

Hi Hiep,

For the captures with 16 bit (2 bytes) NOP, there really is not much to share in the way of code.  I used a Digilent 'Digital Discovery' puck to generate the /CS, SDI and SCLK.  Outside of the context of the tool, I don't think it is going to be of much help to you.

Hi Tom, I thought you were testing using ADS8681-EVM and had some test code to read the Conversion data.

Hi Heip,

I'm using the EVM for both actually, it's just difficult to get you code based on our GUI since it's essentially a pattern generator running on an FPGA so there isn't any 'C' source to provide.  The GUI for the ADS8681EVM also does not let me customize any commands to try and, for example, write the same thing you do to the device.  With the Digital Discovery, I have more control over the communications, I just can't run the part anywhere near full speed.