Monte Carlo Voltage Source

This is because you have used mismatch in your models. For mismatch to work, you need to be varying a parameter within a subckt (see Recommended Spectre Monte Carlo modeling methodology). If you'd used process in the statistics block, it would have worked.

If you genuinely need mismatch variation (which seems slightly odd for a single voltage source), you'd need to place the vdc within a sub-block. Create a simple schematic just containing the vdc (as you have it), create a symbol for it, and then place that in your test bench and proceed as you have right now. You'll then see the variation.

Regards,

Andrew

Hi Andrew,

Thank you for your quick reply! I have tried this, but unfortunately I am still having no luck.

I modified the test bench so it now looks like this: 

Where the block with terminals A and B just contains this circuit:

I also made some very minor changes to the ".scs" file, seen here:

"simulator lang=spectre
statistics {
    mismatch {
         vary MYVDD dist=gauss std=1
    }
}"

I included this file again by going to Setup > Model Libraries. I had to define MYVDD as a Design variable and set it to zero (as I understand it, this defines the mean of the distribution). Unfortunately, I still get a dc voltage of 0 across every monet carlo run. I think I am messing up the inclusion of the model file, because when I look at the netlist being simulated by going to Simulation > Netlist > Display I see this:

" simulator lang=spectre
global 0
include "$SPECTRE_MODEL_PATH/design_wrapper.lib.scs" section=mainlib
include "$SPECTRE_MODEL_PATH/design_wrapper.lib.scs" section=localMC
parameters MYVDD
include "netlist" "

If I go to the file called netlist I see:

" simulator lang=spectre

// Library name: libName
// Cell name: cellName
// View name: schematic
subckt jco1625_vdc A B
    VMC (A B) vsource dc=MYVDD type=dc
ends jco1625_vdc
// End of subcircuit definition.

// Library name: libName
// Cell name: testCellName
// View name: schematic
I4 (vout 0) jco1625_vdc
R0 (vout 0) resistor r=1K" "

But nowhere is there an include statement or reference to my custom model file. Thank you again for getting to this so quickly!

Do you have corners set up in ADE? If so, then the model files won't be being included if you then use Setup->Model Libraries. If that's the case, you could either add this file to your corners setup, or perhaps just use Setup->Simulation Files and then enter the file as a "Definition File" (Definition Files are included from the test even if using corners in ADE).

Note too that you should also set a value for MYVDD in the Variables in ADE as otherwise there's no centre value for the distribution.

Andrew

Hi Andrew,

The corners setup was the missing piece. I added the model file to my Monte Carlo corner and now I see variation that is defined by the model file and the mean I set in ADE.

Thank you so much!