There is none, and it is actually not a collector except in the regard I mentioned in #17. In general, it is just a diode array.
This is the actual circuit of (a quarter of) the 7400 logic IC. The details are in no way in question any respect! 
We are just discussing how it actually operates. 
Even if you add the third (base-collector) diode you will get a much worse behavior IMHO. The inputs need to sink about 1 mA to get LOW. IIRC diode forward drop decrease about by 0.1 V with 10 times decrease of forward current. That means (assuming equal diodes etc.) the "base-collector diode" would conduct 1 uA when Q2 Base is min(A,B) + 0.3 V. Also the inputs are unable to drain any leakage current (perhaps collector-base of Q2 leakage).
Another reason is speed - the stored charge removal you mentioned is very important IMHO. The only reason bipolar logic survived so long against CMOS was its speed. When you replace Q1 with diodes you will get super-slow IC I think.
(And a wild guess - a single multiemitter transistor takes less die space than equivalent amount of diodes).
Alright, my transistor minded friends, if I wanted to make my own for breadboard use, and I wanted to make them from diodes specifically for 5v logic switching and not amplification, which diodes would be best for making homemade transistors?
That way I can experiment with transistors with more than 3 pins and maybe learn something.
You can't make a transistor from diodes(*). You can make wired-AND and wired-OR gates using them, which is what is being discussed (way off topic really!).
(*) transistor action relies on thermal diffusion of charge carriers across the ultra-narrow base region into the electric field of the base-collector junction (depletion zone). It can also behave like two diodes back to back, but there's no way to mimic a transistor with diodes.
Ah, very good
Hi,
I think your idea of a transistor needs some research.
Although a BJT tests in a DC way as two diodes connected either cathode to cathode (PNP) or anode to anode (NPN), because of the atomic structure and bonding between the junctions when used in a circuit perform differently.
https://www.electronics-tutorials.ws/transistor/tran_1.html
Tom... 
That was exactly the point of the question
Perhaps I don't know enough to ask the question correctly.
If I wanted to simulate the logic switching capabilities, that are one of the many uses of a transistor, could I do so with diodes? And if so what would be a good choice of diodes for that application in the 5V range?
Hi,
If you Google.
diode logic
There are quite a number of hits on the subject.
Tom...
Thanks, that's what I needed.
Edit: aaaand now I get why the transistor took over
So a transistor amplifies a signal, only from the signal's point of view, by acting as a variable gate for an energy source that represents the sum total of the overall machine's capabilities? That way it obeys the laws of physics in that the energy is provided, not actually amplified?
Yes. All sorts of electronic "amplifiers" merely control a voltage/ current being supplied from a power supply, in proportion to a smaller voltage/ current.
The term "transistor" is derived from the concept of a "transfer resistor", that is its resistance is controlled by the base current. More-or-less. 
One step further, thank you.
This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.