[cadence virtuoso practice record (3)_gm/id simulation method]

1 Scan NMOS curve

1) Schematic diagram
Set w = 1u , l as a variable, I = L ,

2) Create a new .scs file to save all the working points of the pipe.

Type the following command:

save NM0:all

After the creation is complete, click ADE/setup/simulation files...

Add the created file to definition files and click ok

3) Run the simulation and view all DC operating points in the Results Browser.

All saved simulation information can be seen in the Results Browser.

####################
No new save.scs, check the dc option in the Results Browser, only a few selected points, not all working points.

#############

4. Output setting
   1° transconductance efficiency gmid
   as follows:
   ①
   
   ②
   
   ③
   
   ④
   
   
   ⑤
   

2° cut-off frequency $f_T$

According to the previous calculation formula, $f_T$ = $\frac{gm}{2\pi Cgg}$
Added gm and Cgg to calculator. #Note that the denominator should be multiplied by 2π
①

②

3° Intrinsic gain gmro
①

②

4° current density $\frac{I_D}{W}$
①

② Select w in instance/NM0, find w, right click and import it into the calculator

③

Here, the output is all set

to check the output curve↓

5)
1°
Note that the abscissa is VGS, what we want to get is the abscissa is gmid, ie $f_T$Curves about gmid.
The operation is as follows:
① First, only check the first two in the output, and draw the curve.

②Draw the curve as shown in the figure

③Modify the abscissa, click on the abscissa, right click, select YvsY, select gmid, and click OK.


get $f_T$Curves about gmid.

2° Same as above, you can draw $g_m$$r_o$About the curve of gmid

3°
Draw the curve of IdoverW about gmid

6) Use the ocean script to scan L to obtain a set of curves. The above single curves are the values ​​scanned under a fixed channel length, that is, to obtain a set of curves by scanning L under the
initial setting of L= 200n . ① ② Set L 0.18u - 1u to scan 10 points ③ Do not click to run, save this setting as an ocean script. ④ Modify the file name, click Ok, the file will be saved in the working directory, and add script commands later.

⑤ Delete all the original plot lines

and replace them with the following command:

newWindow()
ocnYvsYplot(?wavex gmid ?wavey ft)
addSubwindow
ocnYvsYplot(?wavex gmid ?wavey gmro)
addSubwindow
ocnYvsYplot(?wavex gmid ?wavey IdoverW)

⑥Save the file and run it in the CIW window
Enter the command:

load("NMOS_gmid.ocn")

Running result:
g,
only one line comes out,

and an error is reported after checking

#############
Find the problem ing
#############
Save it as an ocean script and report an error. It was possible to use ic51 before, and the prompt is that there is a problem with spectre, and no solution has been found.
The direct scanning method is now used to simulate the gmid curve, which is more concise.
The process is as follows↓↓↓

1) The circuit is shown in the figure

2) ADE simulation settings

Pay attention to the emulation settings of dc

Scan L is set as follows

3) Mainly record the output settings:

1° gmro output curve on gmoverid

Click ADE/outputs/setup, open calculator

Select waveVswave

Select os, pop up the select instance, and then click the pipe, and various parameters of the pipe will be displayed in the list drop-down menu.

Click gmoverid in the list, it will appear in the calculator, and then copy it to the select x Trace of wave Vswave.

Then select gm/gds and copy it to the select y Trace of wave Vswave. Click ok.

get expression, gmro's output about gmoverid is set

Here you can also directly select self gain in the list, the result is the same.

The result of the scan is the same

2° ft about the output curve of gmoverid

According to the formula, $f_T$ = $\frac{gm}{2\pi Cgg}$
Added gm and Cgg to calculator. #Note that the denominator should be multiplied by 2π

3° IdoverW About the output curve of gmoverid
Here w is a variable,

click os, select id

and click var, select w

and then directly click / on the left small keyboard
to get the expression of IdoverW, copy it to the y-axis below, and click ok.

4° Draw another lambda curve about gmoverid

$r_o$ = $\frac{1}{g_{ds}}$ = $\frac{1}{i{\_}_d}$

Therefore, $\lambda$ = $\frac{g_{ds}}{i_d}$

So far, the curves of nmos tubes have been scanned

2. Scan the curve of the PMOS tube.

Same step as above

1) PMOS tube circuit diagram

ADE setting

2) Output setting
1° ft

2° gain

3° IdoverW

4° $l$

The gmid simulation is completed,
mainly the output setting, and then try to build the circuit with gmid