Verilog-A: How to trigger event, if input parameter changes in DC sweep

Can you please share the Verilog-A code and explain how you're passing the sweep variable to this parameter? There isn't a way to create an event to do this, primarily because there should be no need to do so - it should just work.

Andrew

Hi Andrew, 

please find attached a simplified Verilog-A code which can demonstrate my issue. It seems that the @timer paragraphs breaks the DC simulation result. With this paragraph, the variable Vout_val is constantly kept at the starting value of the DC sweep. Without this paragraph the DC sweep simulation works fine. I need this timer for a transient analysis and would not have assume this to influence the DC analysis, but somehow it does!? 

Also, I have attached two screenshots, how I pass the variable from the DC sweep in Maestro to the Verilog-A block.

Do you understand where this is coming from and have an proposal how to fix this? Thanks a lot!

Best regards
Niklas

`include "constants.vams"
`include "disciplines.vams"

module VA_sweep_example(Vout, VSS);

	output Vout;	
	input VSS;
	electrical Vout;
	electrical VSS;
	
	parameter integer sweep_param = 0 from[0:31];
	
	integer Vout_val;
	real delay1 = 1u;

analog begin
	@(initial_step) begin
		Vout_val = sweep_param;	
	end

	@(timer(delay1)) begin
		Vout_val = Vout_val +1;
	end

	V(Vout, VSS) <+ Vout_val*1e-3;
end
endmodule

The initial_step is going to only get fired for the first point in the DC sweep, so that's why it doesn't work (it's nothing to do with the @timer). So Vout_val only gets set once.

You could add this:

if(analysis("dc")) Vout_val = sweep_param;

and that would fix it. I normally don't like using analysis() in Verilog-A code because generally analysis-specific behaviour is incorrect, but it's not that obvious how you'd resolve this otherwise.

Regards,

Andrew

Hi Andrew, 

by using analysis() I could resolve my issue. Thank a lot for the hint. 

The @timer function has some influence on my DC sweep simulation. If I have it included in the Verilog-A code, only the starting value of the DC sweep is used for all sweep values. If I remove it from the code, Vout_val is updated with all the "maestro_var" settings of the sweep. See behavior of Vout in the attached plot for both cases. Maybe this is a bug?

Regards
Niklas

Hi Niklas,

I think this is because with any complex behaviour (actually anything within the code that alters Vout_val) the simulator determines that the changing parameter doesn't affect the operating point. Without the @(timer) then it recomputes the operating point which triggers the initial_step again. I'm not 100% sure, but I think that's what's going on. Might be worth contacting customer support to investigate further to be sure.

Andrew