This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC8734 Quad 16bit DAC Bipolar Output voltage swing limitation from power supply voltages

Other Parts Discussed in Thread: DAC8734, DAC9881

I would like to use the DAC8734 to generate a +- 5V bipolar output signal into a 50Kohm load. I would like to use +- 6VDC supplies for AVDD and AVSS. Will the DAC8734 deliver +-5V into the 50K load with +- 6V power supplies. What is the required power supply headroom for a light circuit load like this?

I plan to use Vref = 2.5V and x4 Gain for a bipolar output swing of +- 5V. AVDD = +6V, AVSS = -6V.

End application is generation of an ECG signal of 1 mVp-p AC magnitude, riding on a DC bias voltage of up to +-400mVDC. Load impedance will be 50 Kohms. I plan to use the DAC to generate output signals of up to +- 5 VDC, and attenuate the output level down to +- 450mVDC with an external 11:1 attenuator.

If the DAC8734 cannot support this output voltage level, what happens if I scale the output to reach this level, and simply allow the DAC output stage to clip the waveform? What maximum output voltage can be generated into the 50K load with +-6V power supplies?

Thanks for your help!

  • Howdy Ronald,

    I believe the maximum reference value that you can use for an AVDD value of 6V is a reference value of 3V.  The "absolute maximum ratings" gives us more insight into this parameter as "REF-x to REFGND-x,  it states that the reference has an abs max rating of "-0.3 to min(AVDD/2, -AVDD/2)".  Therefore, the reference value can go up to 3V for an AVDD value of 6V.  You will need to supply a min value of 10V on AVDD for a reference value of 5V.  If the requirements presented in your previous post are concrete and cannot relax to conform to this requirement, we can definitely search other devices that may work.

    Please let me know.

    Best Regards,

    Matt

  • Hi Ronald,

    After further investigation, I believe you can achieve an output of -+5V if you follow a similar approach to the following example:

    AVDD:6V, AVSS:-6V, DVDD = 5V (abs max is 6V), IOVDD (abs max is DVDD)

    Set REF to 2.5V, GAIN to 2x

    The output range should then be -5V to 5V. Note that the output voltage must not be greater than (AVDD -1.0V) or less than (AVSS + 1.0V); otherwise, the output may be saturated. This is cutting it close with an AVDD/AVSS supply of +-6V, therefore additional margin may be helpful.

    Best Regards,
    Matt
  • Hi Matt,

    Thanks for your very quick and thorough replies. It caused me to read the data sheet more carefully (!), which disclosed the output voltage limitations with respect to the AVDD and AVSS supplies, basically the output swings to within 1V of either supply. 

    Your suggested solution looks like just what I need. I will use a 2.5V reference and +-6V supplies. However, my interpretation of the data sheet is that the total output swing, peak to peak is Vref * Gain. Since I desire an output voltage swing of +- 5V (or 10V p-p), this leads me to think that I will need to use the GAIN setting of x4, ant no x2. Am I correct in this interpretation?

    I will use an external precision voltage divider to reduce the +-5V output levels to +- 500mV output levels, to cover the full ECG + DC Offset Voltage requirement. This will then set the LSB at about 15 uV / LSB. This is a little on the high side. Can you suggest any other solution which could provide very tight Gain matching between channels (0.1% gain mismatch max.) with more resolution? I looked at the 24 bit Audio DAC possibility, which has better AC dynamic range, but their matching is not adequate for the 9 channels which I need. I think that the DAC8734 is my best option. Any comments?

    Thanks again for your great help! Ron

  • Howdy Ron,

    You are correct, you will require a GAIN setting of x4, for a -5V to 5V span.  We have a single channel 18-bit DAC in our portfolio, the DAC9881. It is a unipolar DAC, but you can definitely create a bipolar output by using an op-amp in a difference amplifier configuration.  This information is shown in page 33 of the DS.  The part features a max gain error of 32LSB.

    One additional thing to note is that the tolerance value of the external passives will also impact gain error.  Therefore, it is important to pick passives with low (%) tolerance while creating the bipolar output circuit, or the attenuator circuit. Additionally, the op-amp chosen must exhibit low leakage, input bias current, and offset error, as this will contribute to the overall error of the system.

    Best Regards,

    Matt

  • Hi Matt,

    Thanks for your further followup correcting the Gain needed for a +-5V output to x4 with a 2.5V reference. I will use this configuration. Thanks also for the suggestion of the DAC9881, which is a Very good part indeed. However, I will stick with the DAC8734 for its higher level of integration and Bipolar output capability. I will be using a precision output divider network which has 0.05% matching to attenuate the DAC output down to the required ECG level signals. The Quantization noise of the DAC, with LSB = 150 uV will be down to about 7 uVrms over the ECG 100 Hz BW, assuming a sampling rate of 8KSPS. After further attenuation to match the ECG signal levels, this quantization noise will drop to less than 1 uVrms over 100 Hz. The required noise performance is less than 5 uVrms, so this will be an acceptable solution (actually a very Good solution!). 

    Thanks again for your excellent assitance in discussing this issue.

    cheers, Ron

  • Howdy Ron,

    That's great! I'm glad I was able to help. 

    Best Regards,

    Matt