aamar
Community Member
Offline
Posts: 57
Germany
|
I do it in another way, inorder to plot gm,gds vs. Vds, I use the parameteric analysis too but in the following way.
you set your schematic to get the current tansfer c/cs. once against vds and once against vgs (I mean here the sweep parameter), as you can sweep just one of the too at once then you can set the other parameter to a constant value or use it as a parameter for your parameteric analysis.
regarding the gds it should be easy, after plotting your current curves one can use the derv function in the calculator to get the d(Ids)/d(Vds) which is your gds in this case for a defined Vgs value, and as I said you can set the Vgs as a variable and repeat the simulation for different values of Vgs (normally Vgs is predetermined in your circuit because it depends on your biasing) but what can be done is varying the length of the transistor as an example while keeping the W/L constant.
inorder to make the same for gm , we use the other curve Ids vs. Vgs and derive it to get gm but it is in this case gm vs. vgs , so inorder to translate it into gm vs. vds you have to run the simulation for Ids vs. Vgs for several different vds values say 10 as an example, so the deriviation will result in a family of curves all are gm vs. vgs but for different vds values, then you use the function value from the calculator to caluculate for you the gm values at a definit value of Vgs and plot this expression it will plot for you the gm vs. vds.
So in short you have 3 expressions which you can save in your analog enviroment and run it every time after running the parameteric analysis.
after getting your required gm,gds or gm/gds you can go back and check for the current required to achieve these values (Ibias).
This solution doesn't need editing in the netlist.
best regards,
aamar
|