I'm trying to implement a method used by Mr. Jeff Miller in a Matlab training session entitled "Fixed Point Made Easy," and had a question regarding his use of look-up tables

My question is regarding the splitting of the sfix18 angle value into two numbers, one composed of the upper 9 bits the other composed of the lower 9 bits. Isn't the LUT_HIGH for angles =>pi, and the LUT_LOW for angles <pi so that only one LUT is being used at a time? I was also wondering what the workings of the HDL_Complex_Multiplier would look like. Thank you for your attention to my question.

Sincerely, Michael

Answer 1

[Kiran Kintali]

Can you share the training material and models you are referring to?

Looking at the picture you attached the two LUTs are not two sets of angles but upper and lower bits of the angle input.

You may find attached models helpful.

[Michael Kussmaul]

Here's the link to Mr. Miller's presentation: Fixed-Point Made Easy for FPGA Programming - YouTube. Here's the link to the materials in the presentation: https://bit.ly/378BEy3. You're right about the angles. I was thinking that the lower LUT would have sine values relating to angles of 0 to pi and the upper LUT would have values relating to angles of >pi to 2*pi. The angle though only has a single value at a time and would seem to only use one LUT for each clock cycle. Thanks very much for your response to my inquiry.