Claude,
MrAl certainly deserves a thank you, but you neglected to mention Jony130 for his contribution. And do I not also deserve an honorable mention for not walking away, and agreeing with something I did not quite believe just to settle the matter? Unfortunately my unfamiliarity with Ltspice and a bum transistor chart did not help to shorten the discussion either.
MrAl,
I said that I agreed with you about Ic controlling the transistor, especially during saturation. That is certainly true, but I was wondering why you made that point. It was never in doubt, was it? Much of the discussion was whether Ic increased enough to bring the transistor out of saturation. And while in saturation, Ib doesn't have much effect on Vce(sat), but it does on what value Ic brings the transistor out of saturation. So it can be said that Ib also controls the transistor during saturation at the same time. I don't want to start another round of discussion on this, because I believe we each see the other's point of view.
Ratch
By thanking Mr. Al, I did not mean to exclude you or Jony130. It's just that Mr. Al has been on this thread since the beginning & has spent much time & energy doing simulation, & presented his findings in an organized manner. That is why I thanked him, because his methods are how science advances. Brownout has also brought good info to the table which we all benefitted from. But now that you mention it, I thank you for agreeing w/ Mr. Al.
One salient point I'd like to make is that earlier on, you mentioned that the I-V curve of a bjt cease to be exponential when the device goes into saturation. Although that is true if we view only Ic vs. Vbe, keep in mind that there are 2 things going on. Ebers & Moll pointed this out in their 1954 paper. In saturation, the b-c jcn is forward biased. The base current has 2 components, base to emitter, & base to collector. In the base region these 2 components add together. But the collector current Ic, is made up of the component due to the b-e jcn forward bias, & the other component, due to the b-c jcn forward bias. The 2nd component subtracts from the 1st. If we visualize an upside down bjt, the collector becomes the emitter & vice-versa.
The relation between Ic & Vbe is still logarithmic, even in saturation, but Ic/Ie consists of 2 parts in opposition.
Ic = alpha_n*Ies*(exp(Vbe/Vt)-1) - Ics*(exp(Vbc/Vt)-1),
Ie = Ies*(exp(Vbe/Vt)-1) - alpha_i*Ics*(exp(Vbc/Vt)-1),
where alpha_n/alpha_i are the normal/inverse common base current gains resp., & Ies/Ics are the reverse saturation scale currents for the b-e/b-c jcns resp.
In the active region, the 2nd term involving Ics/Vbc is small since the b-c jcn is reverse biased, i.e. leakage current. Hence 1 term is all that is needed, & it exhibits exponential I-V. In saturation, 2 terms are involved, & the exponential nature of the junctions is not immediately apparent.