Okay I have decided to use circuitlab.com to simulate a 2N3904 npn. You can repeat this by going to circuitlab.com, grounding E, fixing Vbe=0.7V and sweeping Vce.
With Vbe=0.7V, I plotted Ib (into base), Ic(into collector), -Ie (out of emitter)
The plot is as follows:
When Vce=0V, according to circuitlab sim, there is actually current flowing out of both collector and emitter! (negative sign of Ic)
Vce needs to rise above 5mV before Ic current reverses to be INTO the collector!
This is what I thought should happen and sims show it is true.
Vce=0.3V Ib=32.6uA (At this point, Ib is fairly steady and we are well in active mode)
Vce=0.2V Ib=32.8uA (only a few percent higher)
But as we go below Vce=0.1V, Ib in fact increases exponentially.
So far this all makes sense. But that still leads me to one last question.
By fixing Vbe and lowering Vce, I see that both Ie and Ic go down.
Is it the case that as Vbc gets forward bias, it adds a reverse diffusion current that cancels out the Vbe forward bias diffusion current?
I've never heard of saturation described this way but it seems like this is what is happening.
Rather I've heard it described confusingly as base current is being "saturated" with carriers and can't hold any more. But clearly Ib is increasing so this isn't quite true.
But looking at it as two opposing diffusion currents (one from collector electrons flowing to base & the other from emitter flowing to base) makes a lot of sense. Is that what is happening or does it just look that way and it is the wrong way to think about saturation?