AC BJT analysis

[lordvader88]

If u use the Ebers-Moll model and replace the Vbe with a voltage source of Vbe, then if u had an emitter resistor, and then the (Vbe) V.S., can u do a source transformation on that into a current source if u want ? And at AC, do we treat the (Vbe)V.S. as any other DC source, so a short at AC when drawing a BJT AC model?

I've watched videos on this already, but so far none explicitly said that's the case. I've only just been learning how to treat V.S. and C.S. in mesh/nodal analysis, I should have done that last year. So I'm only getting into EM BJT models and other models, and AC analysis and actually using the equations for such now....and actually remembering them. But I love it.

[T3sl4co1l]

Vbe is in series with r_pi, the dynamic base resistance.  This is simply dVbe/dIb = Vth / Ib, plus Rbb' if you're counting.

Then it can be transformed into a Norton equivalent if you like, yes.

Note that due to feedback, if there is positive emitter impedance, the base input impedance is bootstrapped even higher, to roughly h_fe * Z_e.

Look up the hybrid pi model -- this is the formalism usually used with BJTs.  You can apply arbitrary small-signal parameters, or derive them from device properties (N, Is, etc.) and use Ebers-Moll, or Gummel-Poon or whatever you like.

Tim

[lordvader88]

Well I think what I was rambling about late last night was....say if I make a 5 BJT circuit with resistors, even if it looks messy, I should be able to find all the voltages and currents using that type of model, and it's just another mesh/node problem. And indeed, at that point I'm allowed to split up the transistor just like anything else in the circuit

Right now I have 5BJT amp on paper and LTspice, and I'm working backwarks through what the voltages and currents will be for the output I want.

But so far I still only had to use KCL/KVL. I should try it with node/mesh analysis too for the DC, I haven't gotten near the AC yet