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.

Looking for the "fastest" DAC

Other Parts Discussed in Thread: DAC8760, MUX36S16, DAC8830

Hello,

In the scope of a project, I have to design a µC - DAC - Amplifier chain with 30 outputs on my DAC block.

My most critical specification is that the timing between the µC calculation and the observable voltage at one output must be as fast as possible (~2 µs max).

So far, I've been looking for 16-outputs DAC to use on 2 differents SPI bus (i only have 2 SPI bus available) . However, SPI communication is quite slow, and is being quite impactful on the final timing.

In one of TI's docs (Multi-Channel Analog Output Module With Multiplexed Single-Channel DAC for PLCs Reference Design), I've seen that it was possible to use a single output DAC and a demultiplexer to achieve this result, but i haven't been able to find a 16 or 32 outputs demultiplexer with buffered outputs (my voltage level must stay on while my DEMUX is switching)

So here is my question : what would be the wisest solution in this case? and no matter the solution, would you have any component to advice me on using? (please note that the DAC should be 12 bits min, FSR doesn't matter)

Best regards,

Alexandre Lecomte

  • Hi,

    Since you have 2 SPI signal interface, I suggest the following approach.

    Use 2 DAC8760 or if you want voltage output only, choose any MDAC from our portfolio and 16 * 1 demultiplexer, Capacitor and Buffer 

    SIngle 16 channel looks something like this.

    uC --> DAC8760 ( or MDAC) ---> MUX36S16 ---> Hold Cap --> Buffer --- >

    Regards,

    AK

  • Hello,

    Thank you for your fast answer.

    I've looked at the DAC, and i fear that it may have a settling time way too long for my application, but since i'm only using voltage output, i could use DAC8830 for instance. For the demux, it seems like a suitable component for the application.

    The next part seems quite blurry for me :I don't know about hold capacitor, but won't it cause a problem when commutating? let's say I'm switching from 0 to 5V, I=C* du/dt that would mean a current spike of a few Amperes right? I know that in most application, you can put an amplifier between them to avoid this problem, should i do it here?

    Thanks,

    Alexandre

  • Hi,

    Hold capacitor will be in the order of Nano farads and assuming your commutation is in the order of micro seconds, this leads to milli ampere currents, not amperes. Pease read the reference design you mentioned above to get an indicative value for your design. 

    Also all these depends on the accuracy, settling time and how much leakage voltage drop you can accept in your design.

    Do a rough calculation to arrive at these numbers and we can discuss.

    Regards,

    AK