TAC5212: Incremental mode

Hi Timothy,

Further documentation on this feature is still in progress, please wait as I contact our validation and design team on more information to assist you. I will update this thread next week.

Thanks!  If one-shot mode is not needed, and instead a DC-coupled input is to be converted continuously at the rate set by BCLK, etc., is incremental mode needed at all, or is setting DC coupling and disabling low-pass filtering all that is required?

Hi,

I'm not following. You don't want to use IADC to measure DC voltage of INx?

You just want to send a AC signal, in dc coupled configuration, and don't want to lose the input values at low frequencies? If so, yes that shouldn't be a problem. Just ensure that if negative voltages are in the input, the INx pin common mode voltage is biased correctly.

Right, just trying to understand the benefits of using incremental mode rather than "normal" mode for DC measurements.  Is the voltage scaling different for incremental mode?  Here is some data comparing TAC5212 ADC output at 12 kHz sampling rate with similar measurements from test equipment.  With register 0x50/0x55 = 0x48, ADC full scale is 1.414 V, and I get good agreement with test equipment by scaling by 2.828/2^24 rather than 0.925*2.828/2^22:

With 0x50/0x55 = 0x4A, ADC full scale is 2.828 V, and scaling by 2.828/2^23 works:

Looks like the input voltage level can be accurately read from these measurements, so what capabilities does incremental mode add to the picture?

Hi, apologies for the delay I will update thread soon

Hi Timothy,

Can you share the method or script used to measure your input signal alternative to the IADC?

HI Daveon,

I'm using a script written in CircuitPython which you can find at https://github.com/timchinowsky/tac5.  The configure method in tac5.py is where the I2C programming happens.  Here's the key part:

self.write_reg(0x01, 1, address=a) # Reset all registers to defaults
time.sleep(0.1)
self.write_reg(0x02, 9, address=a) # No sleep, DREG and VREF enabled
if self.width == 32:
    self.write_reg(0x1A, 0x30, address=a) # TDM, 32 bit
elif self.width ==24:
    self.write_reg(0x1A, 0x20, address=a) # TDM, 24 bit
elif self.width == 20:
    self.write_reg(0x1A, 0x10, address=a) # TDM, 20 bit
else:
    self.write_reg(0x1A, 0x00, address=a) # TDM, 16 bit
self.write_reg(0x78, 0xEE, address=a) # Power up all enabled ADC and DAC channels
# self.write_reg(0x72, 0x0A, address=a) # disable ADC HPF
self.write_reg(0x72, 0x8A, address=a) # disable ADC HPF, ultra-low latency decimation filter
self.write_reg(0x73, 0x0A, address=a) # disable DAC HPF
self.write_reg(0x1B, 0x40, address=a) # transmit hi-Z for unused cycles

self.write_reg(0x50, 0x4A, address=a)
self.write_reg(0x55, 0x4A, address=a)

self.write_reg(0x55, 0x4A, address=a)

Hi Timothy,

Thanks! So you configured the device as 'normal' but logged voltage digital output and mapped to excel graph? No IADC or diagnostic registers used?

Yes, that's right.  So when the datasheet says "TAC5212 supports incremental mode of ADC where analog input channels can be used for DC measurements" but 'normal' mode can also be used for DC measurements, I'm wondering what are the unique features of incremental mode.

Hi Timothy, 

Thanks for the explanation, I've forwarded this thread to our design engineer for this feature to get his feedback.