Insane Voltage Regulator

This is pretty cool, and in fact has perfect stability (zero dependence on input) with ideal transistors. Unfortunately, transistors are considerably unideal, and in this case the Early effect ruins things. The Early effect is the dependence of gain upon collector-emitter voltage with greater voltages leading to greater gain. The strength of the Early effect is described by the Early voltage which is the voltage at which the gain doubles, assuming the rate of change of gain is constant and equal to the rate of change at low collector-emitter voltages. However, this is just an idealization and the actual gain at the Early voltage isn’t necessarily double, assuming the transistor doesn’t just break down before that voltage. Because of the Early effect, there will be a considerable dependence upon input voltage when using real transistors. However, I do not have any experience with real transistors so I cannot say how big the effect would be.

Hey glad to hear from you, it’s been a while hasn’t it :) Interesting info on that Early Effect, never heard of it before, but that makes a lot of since considering that a lot of RF transistors usually specify better gain with higher Vce. I’ll for sure have to look into this further. Also, I recently came up with a flawless new design for a current source that’s I think even much better than Widlar or Wilson current source topologies. But unlike the Widlar/Wilson current sources, It also has zero dependence on input voltage like the circuit above, and is achieved with still a very simple layout just like in Widlar/Wilson requiring only 4 transistors and 1 resistor. Let me know what you think if you have the time http://everycircuit.com/circuit/4931394111012864

That’s quite an interesting design! Can’t say I’ve thought of that before, or seen it anywhere either. It does look much stabler than the traditional design when the supply voltage is expected to vary.

Unfortunately I just breadboarded the one from the link and it couldn’t even regulate current through a little LED when using 5-20 volts, it varied enough to make the LED glow noticeably brighter at 20v, not what I was hoping for.

Have you breadboarded this circuit? Does it work? I ask because I can't figure out how the MOSFET gate gets a carrying voltage drop (depending on load) without any BJTs actually changing current. Perhaps the voltage drop is created by a current change too small to see here??? It just seems like I should be able to see where the voltage drop is coming from, but I don't see it...

Not a 100% sure what you mean, but I will say that this circuit may require a high value start-up resistor (1-10M ohms should work) to activate the base of one of the bjt transistors so that everything gets initially biased. -This is because the feedback loop doesn’t have any pull up/pull down resistors or biasing without assuming the mosfet is already on. Does that help? I haven’t breadboarded this one just yet, but I probably can sometime today and let you know how it does. Only reason I haven’t yet is cause I recently discovered when testing this current-mirror feedback approach, just how dependent the circuit becomes upon inequalities in practical matching of the transistors, so I’ve been going back to the drawing board a lot… and pretty much just assumed that this circuit in the schematic was gonna be just as unstable as my “Out-Performing the 555 Timer” Post that I breadboarded come to find out huge sensitivity problems with temperature. Even touching the transistor casing with your finger is enough of an imbalance to trigger thermal runaway (even if it starts out cool to the touch, which it does in my testing). It would work great otherwise. Here’s the link for it if your curious…

http://everycircuit.com/circuit/6046032613605376 if it redirects you to the website while on mobile, just paste the numbers into the community search bar