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.

DAC8814: Looking for multi-channel precision current DAC

Part Number: DAC8814
Other Parts Discussed in Thread: DAC8803, DAC80508, OPA388

Hi everyone,

I was looking for a solution to test multiple DC regulated rails (from 0.8V to 4V) using a current output DAC.  The current output would essentially serve as a variable current load to these regulated rails.

Are there any programmable 12-bit+ multi-channel DACs that can sink / source currents from 1uA  - 1 mA? Looking for a settling time (when switching between currents) to be around or less than 1us.

Potential DACs I found were DAC8814 and DAC8803 but datasheet mentioned Iout low of 1.6mA which would not work.

Thanks for your help!

Best,

Nik 

  • Nik,

    These two DACs are multiplying DACs, or MDACs. While in the parametric search tool they are listed as current output DACs, which is accurate, it's not necessarily a proper current source or sink DAC as you would normally think of. Essentially the output is a current because the R-2R MDAC ladder is connected to the voltage reference, but this current isn't buffered in any way - so the exact current through the ladder is still influenced by load conditions (i.e. think of this as more of a programmable resistor than a constant current source or sink).

    In a normal application the MDAC acts as a binary weighted current output, driven by the reference voltage. It is followed by a transimpedance amplifier at the output to convert the current into a voltage - often with a ratiometrically matched internal resistor to get this done more precisely as the R-2R imepdance itself could vary 10-20%, depending on the device.

    For your purposes do you actually need both source and sink capability or just a current sink?
  • Hi Kevin,

    Thanks for your support! This means that If I connect the Iout to Vdc (say 4V), then my total current would be I = Vdc - Vref / R correct? If so, unfortunately I'd need something that I can sink directly to ground.

    I'm essentially looking for something that can mimic an electronic load, a current source to ground, to the rails of a DC regulator. My goal is to characterize the DC regulator with different load conditions (Iload) and I was wondering if any DAC solution can mimic the load to ground, on the regulated output rails.

    Maybe below picture will help!

  • Nik,

    Page 16 of the DAC8814 datasheet illustrates a simplified diagram of the internal R-2R ladder. If you were to connect either your 4V or 3.3V sources to the VREFX pin of the DAC and GND to IOUTX you would effectively have your voltage source, a variable resistance, and GND in series. As you change the codes of the DAC, you would change the equivalent input impedance seen by your voltage sources.

    So, the exact current value would be dependent on both the voltage source and the equivalent impedance of the DAC ladder, which I do not think is what you want. You need a constant current sink so you can characterize the voltage response of your sources.

    We have a circuit cookbook for a programmable low-side current sink which should get you where you want to be. You can view that here: www.ti.com/.../slaa868.pdf

  • Hi Kevin,

    Thanks for the note! 

    I was looking a pairing a DAC80508 with a OPA388 and NTA7002N to accomplish this. I'm attaching a simulation of this below.

    I have multiple rails on the regulator however the rail I'm simulating in the attached schematic is a 4V rail referenced to 3V. The Vin is simulating the DAC input with using a gain of x2 5V fullscale output.

    Would you be able to advise if you see any issues with this schematic? Do I need Rfb, Riso and Cf for this?

    Thanks!

    NikSimulation_4.0-3.0VV.TSC

  • Nik,

    Generally speaking we'd like to see between 45 and 90 degrees of phase margin at 0dB gain to suggest circuit stability. In my simulation this design has a bit of a peak in the phase plot at 0dB with 99.25 degrees of phase margin - so it is sort of marginal. Running a small signal step response simulation you'll see a little bit of ringing which agrees with this marginal stability.

    I would suggest you place those components at least in your layout so you have the hooks to tweak the circuit should you find your results undesirable.