Current limiting that includes foldback can be understood in principle by looking at the very simple current limited voltage regulator circuit shown in this MultiSim SPICE simulation:
Q1, Q3, R3, R5 and R6 comprise a very rudimentary voltage regulator feeding our variable load, R7, and although this one is a really loose voltage regulator indeed, its simplicity will help us understand what comes next.
Q2 and R1 comprise a current limiter, but the current limiting action of having Q2 pull base drive away from the base of Q1 as Q2 responds to the voltage drop across R1 is partially bucked out by the voltage divider of R2 versus R2+R4.
Therefore, as heavier load currents pull the output voltage down, the voltage that appears across R2 also diminishes and becomes less effective in opposing the R1 voltage drop that is presented to the base to emitter junction of Q2. The Vbe of Q2 is the difference of the R1 voltage drop less the R2 voltage drop.
The effect is to reduce the current limit value as the output voltage gets lowered. This is the foldback mechanism at its simplest.
If we upgrade the voltage regulator, we can get something like this:
The added parts Q4, Q5, D1 etal that have replaced the 100O R3 of the earlier circuit comprise an active current source that I was afraid might have obscured the foldback principle in the prior example, but everything we said before about the foldback principle there applies here as well.
For all of that, while we're at it, it can't hurt to just take a quick look at that active current source.
“Um … the first schematic is the same as the second. So, it is hard to reproduce the first graph.”
“Steve fixed it. I hope things make sense now.nnJohn”
“The diference between first and second diagram is the current source, the second diagram has a current source, so when input voltage changed, the output voltage dosn't change, is very strong and much better”