Post Go back to editing

AD790 propagation delay with sin wave input


I understand that there is a propagation associated with input voltage overdrive. Is there any way I can relate this number to sin wave of different frequency and amplitude? Say 1kHz, 1V amplitude sin wave referenced at 0V.

I tried used a step function formula:

Vo == Ao*(1-exp(-t/Tc))*Vi

Vo = ~2 ((VOH-VOL/)2)

Vi = overdrive amplitude

Ao = DC gain

Tc = 1/wc, first pole frequency

This gave me sensible result if I had Vi a constant voltage but if I had Vi*sin(2*pi*f*t), it gives me unreasonably long result. I am guessing the hysteresis feedback comes into play?


  • Hi,

    I contacted the product owner of this part. I will keep you posted.



  • Hi Goz,

    I am wondering if there is any update?


  • Hi ttse007,

    Can I know what software are you using during your simulation? can you please share your schematic in case you have one, and can I also have a copy of the result that you get when you use a constant input and a varying input?



  • Hi Jvillanu,

    I don't have any simulation that I am happy with, which is why I am seeking help here. This part doesn't have any model and in either case, I may be running into numerical issues.

    In Matlab script (attached), I implemented to above equation in my first post, assuming a a step input and output to try and determine the Ao and pole frequency.

    In LTSpice, I modeled the amplifier stage of the comparator as a single pole model in SPICE (attached). To see its response to sin or square wave, change VinSin and VinSquare respectively. Changing the GBW and Slew rate limit of the stage has a large impact on the propagation delay as expect.

    I also have a circuit which consists of a AD9838 DDS at the input. The DC voltage removed a C-R divider and fed to the AD790 comparator. The delay is measured by comparing the square wave output of the DDS to the comparator's output. Taking the imepdances into account, I still notice a change in delay respect to amplitude and frequency. Of course, there are a practical issues (for example, offset voltage, purity of my signal generator, accuracy of the passive components), it is really hard to separate all of them.

    I am interested in a relationship that predicts the delay with amplitude and frequency. Perhaps you can suggest another part that may perform more predictably? I do need bipolar input with a reasonably fast propagation.


  • This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin