DAC8568: Bipolar operation with DAC8568C/D at AVdd +3.3V and internal 2.5V reference?

Hi Karl,

I would use a higher AVDD if able to, as using the full VOUT range will be more accurate.

Based on Equation 4 in the datasheet, on page 51, if you are using grades C/D, you will get a asymmetrical bipolar output, if only using one amplifier.

Since grades C/D have a gain of 2, the positive range will always be further than the negative range.

You could in turn use a DC offset op amp circuit to correct these, but this would complicate your circuit further.

Thanks,
Lucas

Indeed, and this was essentially also why I considered the application example incomplete as is. What would be your take on the three-resistor plus op-amp variant outlined in https://www.ti.com/lit/an/slaa869/slaa869.pdf though, given the constraints and needs outlined above? Thanks :)

Hi Karl,

That document is useful, but it has the same issue as the previous equation.

Since VDAC is x2 the value of VREF, the positive scale will always be larger than the negative scale.
These circuits were designed with a VREF = DAC Fullscale.

Since a RFB/RG1 ratio of 4 would be needed to get -10 V zero code, it would make the minimum full scale value attainable as 15 V.
(1 + 4 + 0) Vdac - 4 Vref

Using a 5 V external reference, dividing the DAC output, or gaining the VREF from the DAC x2 (depending on desired accuracy), could be a possible solution.

Thanks,
Lucas

Thanks Lucas - as there is seemingly no practical/easy way around the various aspects of the 2x gain on C/D in this case, I will consider changing to AVdd and external Vref at +5V (nb. also for the ADC then) instead, assuming the digital logic remaining at 3.3V levels will still be OK, cf. the datasheet page 7 stating a minimum high of 0.625 × AVdd for 4.5V ≤ AVdd ≤ 5.5V (i.e. minimum 5*0.625 = 3.125V for the high state, which is arguably a bit tight but presumably a scenario considered in the design of the DAC856x series?).

BR//Karl