Dextrine said:
I completely understand what they're doing up until the point where one would need a full transfer function. They start talking about the frequency response characteristics, which all make sense to me, but the gain, how to calculate the gain, is specifically where I am stuck.
In that paper they use assumed
open loop gain of 80db throughout. See page 8, and every Bode plot.
Section 7 is a remarkably clear nuts-and-bolts level explanation of stability .
In a voltage regulator, loop gain directly determines regulation
so
the more gain the better will be its voltage accuracy under changing load . You'd like infinity if you could achieve it.
Do you see that fig 4 has for input to error amp = (k X Vout) - Vref ,
where k = (attenuation of divider R1 & R2) = R2 / (R1 + R2) and is a not very small number typically between 0.1 and 1 ?
So gain of the loop = whatever is ( gain of error amp ) X (gain of the two transistors ) X 1/k ?
They've assumed 80 db which is a good sized number. I get 10,000 what do you get ?
So, when opamps became available the approach became grab all the gain you can muster , tune the loop for stability over frequency range of your amplifier.
Dextrine said:
I would like to make a series pass linear regulator using discrete components.
For a first design with discrete components , just grab a nice zener diode and tack an emitter follower behind it.
Here's a far simpler "How To " from TI :
https://e2e.ti.com/blogs_/b/powerho...ies-linear-regulator-with-discrete-components
i hope above helps.
@Dextrine That TI application note
you found is wonderful, i'd not seen it before.