How to draw the internal circuits of an operational amplifier?

Question

I'm trying to do the next schematics about the 741 opamp:

but I don't know where to anchor the first transistor or even which should be done first, and how to do the transistor with double emitter. I'm looking for a start.

\documentclass[border=10pt]{standalone}
\usepackage[siunitx, RPvoltages]{circuitikz}

\begin{document}
\begin{tikzpicture}[american voltages, european resistors]
    \draw (-0.5,4)
     node[pmos](Ap){}
     (Ap.D) to [short] ++(0,-1) node[nmos, anchor = D](An){}
     (Ap.G) to (An.G)
     ($(Ap.G)!0.5!(An.G)$) to [short,-o] ++(-1,0) node[anchor=east]{$V_i$}
     ($(Ap.D)!0.5!(An.D)$) to [short,-o] ++(1,0) node[anchor=west]{$V_p$};
 
\end{tikzpicture}\end{document}  

Thanks.

UPDATE

Thanks for the great ideas and superb support! I'll post the advance of the circuits, here.

Answer 1

Consider locating all multi-outlet components as nodes; Where possible using relative locations, not absolute as in my code, I did it to nest them in a for loop, to do so you can use help lines, then you can start creating the connections with the corresponding bipoles using the coordinates of each node created, Also add the code to be able to place 2 lines in the names of the components. For the rest such as the use of endings, changing the size of each family of components consult the manual of course. I also recommend that you consult the visual manual of tikz so that you can see how to use relative locations, in midpoints and intersections.

RESULT:

MWE:

\documentclass[tikz,border=1cm]{standalone}
\usepackage[american currents]{circuitikz}
% This code is from Claudio Fiandrino https://tex.stackexchange.com/a/65792/154390
% Ads new label styles to allow aditional labels like two line descriptions.
\makeatletter
\ctikzset{lx/.code args={#1 and #2}{ 
        \pgfkeys{/tikz/circuitikz/bipole/label/name=\parbox{1cm}{\centering #1  \\ #2}}
        \ctikzsetvalof{bipole/label/unit}{}
        \ifpgf@circ@siunitx 
        \pgf@circ@handleSI{#2}
        \ifpgf@circ@siunitx@res 
        \edef\pgf@temp{\pgf@circ@handleSI@val}
        \pgfkeyslet{/tikz/circuitikz/bipole/label/name}{\pgf@temp}
        \edef\pgf@temp{\pgf@circ@handleSI@unit}
        \pgfkeyslet{/tikz/circuitikz/bipole/label/unit}{\pgf@temp}
        \else
        \fi
        \else
        \fi
}}

\ctikzset{lx^/.style args={#1 and #2}{ 
        lx=#2 and #1,
        \circuitikzbasekey/bipole/label/position=90 } 
}

\ctikzset{lx_/.style args={#1 and #2}{ 
        lx=#1 and #2,
        \circuitikzbasekey/bipole/label/position=-90 } 
}
\makeatother


\begin{document}
    \begin{tikzpicture}[]
    %Size adjust
    \ctikzset{nodes width/.initial=0.05}
    \ctikzset{transistors/arrow pos=end}
    \ctikzset{bipoles/thickness=0.75}
    \ctikzset{sources/scale=0.75}
    \ctikzset{diodes/scale=0.5}
    % Canvas help lines.
    \def\wcanvas{15}
    \def\hcanvas{12}
    \draw[red!5,step=0.25] (0,0) grid (\wcanvas,-\hcanvas);
    \draw[cyan!40,step=1,line width=1] (0,0) grid (\wcanvas,-\hcanvas);
    \foreach \x in {0, ..., \wcanvas} {%
        \node[anchor=-90] at (\x,0) {\tiny\x};
        \node[anchor=90] at (\x,-\hcanvas) {\tiny\x};
    }
    \foreach \y in {0, ..., \hcanvas} {%
        \node[anchor=0] at (0,-\y) {\tiny-\y};
        \node[anchor=180] at (\wcanvas,-\y) {\tiny-\y};
    }

    %Create each transistor by type, orientation, ID,name,xposition,yposition.
    %
    \foreach \nodetype/\xscal/\alias/\name/\px/\py in {
        npn/1/Q1/Q1/2/-3,
        npn/-1/Q2/\ctikzflipx{Q2}/5/-3,
        pnp/-1/Q3/\ctikzflipx{Q3}/2/-4.25,
        pnp/1/Q4/Q4/5/-4.25,
        npn/-1/Q5/\ctikzflipx{Q5}/2/-9,
        npn/1/Q6/Q6/5/-9,
        npn/1/Q16/Q16/7/-8.25,
        npn/1/Q17/Q17/9/-9, 
        npn/1/Q22/Q22/11/-8.25, 
        npn/1/Q14/Q14/13/-4,
        npn/1/Q20/Q20/13/-7%%       
    }{\node[\nodetype,xscale=\xscal](\alias) at (\px,\py) {\name};}
    \draw[color=black!40!red]
    (6,-1.5) 
        node[anchor=-90,inner sep=7pt]{$V_{CC}$}
        to [short,o-*] ++(0,-0.5) coordinate (lineVcc)
        -| (Q1.C)
    (Q1.B)
        to [short,-o] ++ (-0.25,0)
        node[anchor=0,inner sep=7pt]{$V_{P}$}
    (Q3.B)
        -- (Q4.B)
    ($(Q3.B)!0.5!(Q4.B)$)
        to [I,*-o,lx_=$Q_{10}$ and $I_1$] ++ (0,-1.5) 
        node[anchor=90,inner sep=7pt]{$V_{EE}$}
    (Q3.C)
        -- (Q5.C)
        to [short,*-] ++ (1.5,0) coordinate (temp)
        to [short,-*] (Q5.B-|temp)
    (Q5.B)
        -- (Q6.B)
    (Q5.E)
        |- (6,-10) coordinate (lineVee)
        to [short,*-o] ++ (0,-0.5)
        node[anchor=90,inner sep=7pt]{$V_{EE}$};
    
    \draw[color=black!40!green]
    (Q2.C)
        to [short,-*] (Q2.C|-lineVcc)
    (Q2.B)
        to [short,-o] ++ (0.25,0)
        node[anchor=180,inner sep=7pt]{$v_{N}$}
    (Q4.C)
        -- (Q6.C)
        to [short,*-](Q16.B)
        (Q4.C)
    (Q6.E)
        to [short,-*](Q6.E|-lineVee)
    (Q16.C)
        to [short,-o]++(0,0.25)
        node[anchor=-90,inner sep=7pt]{$V_{CC}$}
    (Q16.E)
        -- (Q17.B)
    (Q17.C|-lineVcc)
        to [I,*-,lx_=$Q_{13A}$ and $I_2$] (Q17.C|-Q14.B)
        to [short,-*] (Q17.C)
        to [short] (Q22.B)
    (Q17.E)
        to [short,-*](Q17.E|-lineVee);
        
    \draw[color=black!40!blue]
    (lineVcc)
        -| (Q14.C)
    (Q14.E)
        -- (Q20.C)
    ($(Q14.E)!0.5!(Q20.C)$) coordinate (mid1)
        to [short,*-o]++(0.5,0)
        node[anchor=180,inner sep=7pt]{$v_{O}$}
    (Q20.E)
        |- (lineVee)
    (Q22.C|-lineVcc)
        to [I,*-*,lx_=$Q_{13B}$ and $I_3$] (Q22.C|-Q14.B) coordinate (inter)
        to [short] (Q14.B)
    (inter)
        to [D,l_=$Q_{18}$](inter|-mid1)
        to [D,l_=$Q_{19}$,-*](inter|-Q20.B)
    (Q20.B)
        -| (Q22.C)
    (Q22.E)
        to [short,-*](Q22.E|-lineVee);
    
\end{tikzpicture}
\end{document}

Answer 2

For the multi-emitter (and multi-collector) transistors you can look at the section "multi-terminal bipolar transistors" in the manual: https://texdoc.org/serve/circuitikz/0#subsubsection.4.14.2.

There is an example later in the section; in that case I started by positioning Q1 at the origin and then I started building everything around using only relative ana perpendicular coordinates. That way you can move the circuit around freely:

\documentclass[border=10pt]{standalone}
\usepackage[siunitx, RPvoltages]{circuitikz}
\begin{document}
\begin{circuitikz}
    \ctikzset{transistors/arrow pos=end}
    \draw (0,0) node[bjtpnp, xscale=-1](Q1){%
    \scalebox{-1}[1]{Q1}};
    \draw (Q1.B) node[bjtpnp, anchor=B, collectors=2]
    (Q2){Q2} (Q1.B) node[circ]{};
    \draw (Q1.E) node[circ]{} node[vcc]{} (Q2.E)
    node[vcc]{} (Q1.E) -| (Q1.B);
    \draw (Q1.C) to[R, l_=$R_0$, f=$I_0$] ++(0,-3.5)
    node[ground](GND){};
    \draw (Q2.C) -- ++(0,-0.5) coordinate(a);
    \draw (Q2.C1) -- ++(1,0) coordinate(b) -- (b|-a);
    \draw (a) ++(0,-0.1) node[flowarrow, rotate=-90,
    anchor=west]{\rotatebox{90}{$I_0$}};
    \draw (b|-a) ++(0,-0.1) node[flowarrow, rotate=-90,
    anchor=west]{\rotatebox{90}{$I_0$}};
    \path (b) ++(0.5,0); % bounding box adjust
\end{circuitikz}
\end{document}

Answer 3

To know in which direction to take the diagram, in fact we can take several which will ultimately give the right diagram but sometimes with more complex paths.

To help you:

• name the transistors as in the diagram. For example : \draw (0,0) node[below]{$v_p$} to[short, o-] (0.5,0) node [npn, anchor=B](Q1){Q1};. So to connect a component to the emitter of transistor Q1, you can use as coordinate (Q1.E)
• to flip a component, use the command xscale=-1
• do not hesitate to abuse the name of nodes. An example: for the node located between two transistors Q3 and Q4, you can name it q3q4 with the command coordinate(q3q4)