Moving average offset in verilogA

The article Moving Average Function in OCEAN/calculator went through quite a bit of refinement to improve it with respect to getting the DC level correct (took a few attempts for me to get it right). I've never had the opportunity to try to do the same with the Verilog-A (it's quite hard because it uses a piecewise strategy of combining the results together in the calculator function to ensure that the startup behaviour doesn't break anything).

Unfortunately it's unlikely I'd have time to revisit the moving average Verilog-A to see if this can be addressed soon (the Verilog-A article was written by another colleague many years ago). This is one of those things you just need to sit in a quiet room and think carefully about - quiet rooms are in short supply at the moment!

Andrew

I found the verilogA and also encountered the offset issue. In my application I was able to work around it by first determining the initial offset at time=0 and used it to calculate the integral

parameter    real    gain=1.0;
parameter    real    time_window= 1u from (0:inf);

real x, k1;

real y0;

analog begin

    @(initial_step) begin

      k1 = gain/time_window;

     y0 = V(in);

    end

    x    =    k1*(V(in)- absdelay(V(in),time_window));
    V(out) <+  y0 + idt(x,0.0);
end
endmodule