Preparing for a lab defense, a student was asked how the emitter resistor reins in a rising collector current — and answered that it can't, because VE is pinned at VB − 0.7 V and never moves, so VBE never changes and nothing throttles the current. That answer is wrong, and it's wrong at the very first step.
Here's the fix. VE is not a fixed voltage handed down by the base. VE is the voltage across the emitter resistor RE, and that's whatever the current through RE makes it: VE = IE·RE, with IE ≈ IC. So when IC rises, more current flows through RE and VE rises right along with it. VB stays put — it's held by the resistor divider — so a rising VE eats into VBE = VB − VE, which throttles the base current and pulls IC back down. That throttling is the whole point of the emitter resistor.
If you've been stuck on the feedback story, this is almost always the broken link: the belief that VE can't move.
Why "VE is fixed" feels right
You learn the relationship VE = VB − 0.7 early, and it does something quietly misleading. It makes VE look derived from VB — and since VB is a fixed level set by a rock-solid voltage divider, your brain files VE in the same drawer. Fixed input, fixed output. VE becomes a clamped node: a little shelf 0.7 V below the base, sitting there no matter what.
Under that picture the feedback chain genuinely doesn't parse. If VE never moves, then VBE = VB − VE never moves either, so nothing throttles IB, and there's no mechanism to rein in a runaway IC. The story falls apart at the first link. That's not sloppy thinking; it's a correct deduction from a wrong premise.
The wrong premise is direction. VE = VB − 0.7 isn't a law that sets VE. It describes where things land when the transistor sits at a normal operating point with VBE near 0.7 V. It's the resting position, not a clamp. Push the current and the resting position moves.
What VE actually is
Look at where VE physically lives: the top of RE, with the bottom of RE tied to ground. RE is just a resistor to ground, and the voltage at its top node is fixed by exactly one thing — Ohm's law on the current passing through it.
VE = IE · RE
That's it. RE doesn't know or care what VB is doing. It develops whatever voltage its current demands. VB and the 0.7 V drop determine how much current the transistor tries to push, but once that current is flowing, VE is the resistor's honest report of it.
Think of RE as a pressure gauge plumbed into the emitter line. Harder current push, higher reading. The needle isn't pinned; it tracks the flow. "VE = VB − 0.7" is just where the needle rests when everything is calm. This is also exactly why adding RE makes IC nearly independent of β — the same feedback that lets VE move is what swamps out the transistor's own gain.
The numbers
Take a concrete circuit. RE = 2 kΩ, and the divider sets VB at about 2.7 V.
At the design Q-point, IC = 1 mA. Since IE ≈ IC:
VE = IE · RE = 1 mA · 2 kΩ = 2 V
So VBE = VB − VE = 2.7 − 2 = 0.7 V sits across the base-emitter region — exactly the normal silicon junction drop. Stable.
Now a temperature spike hits. BJTs conduct more eagerly when they heat up, so IC drifts upward — say toward 1.5 mA. Watch VE:
VE = 1.5 mA · 2 kΩ = 3 V
Not 1 V. Not unchanged at 2 V. Three volts. One and a half times the current through the same resistor is one and a half times the drop — the most ordinary thing a resistor can do. The "VE is clamped" model has no way to produce this number; the Ohm's-law model produces it without thinking.
Here's where the feedback bites. VB is still 2.7 V — the divider didn't move. So now:
VBE = VB − VE = 2.7 − 3 = −0.3 V
VBE just collapsed from 0.7 V toward −0.3 V — a brutal squeeze on the base-emitter junction. A BJT's base current depends exponentially on VBE, so crushing VBE chokes IB hard, which chokes IC (IC = β·IB). The current that tried to run away to 1.5 mA gets reined back in. The circuit never actually reaches that −0.3 V — long before VE climbs to 3 V, the shrinking VBE has already throttled the current. The circuit settles at a new point where VE has risen just enough to restore a sane VBE, somewhere between the two extremes, not all the way out to 1.5 mA.
Run it the other direction and the same loop catches you: if IC tried to fall, VE would drop, VBE would grow, IB would rise, and IC would climb back up. The emitter resistor pushes back on motion either way. That bidirectional pushback is what "negative feedback" means here.
The one sentence to keep
VE moves because it is a measurement, not a setting — it reads the current through RE. The loop exists at all only because VE can rise. If it were truly clamped to VB − 0.7, RE would stabilize nothing and you'd be back to the fragile fixed-bias circuit the emitter resistor was added to fix.
The order of causation, stated cleanly: current rises → VE rises (Ohm's law on RE) → VBE shrinks (VB fixed) → IB falls (exponential junction) → IC falls. Step 1 is the one people drop, and dropping it breaks the other four. If you want to see where the new equilibrium actually lands, it's the same machinery behind how the Q-point shifts in a voltage-divider bias circuit.
Check yourself
In an emitter-degenerated BJT bias circuit, RE = 2 kΩ and the divider holds VB fixed at 2.7 V. The collector current settles at IC = 1 mA. A temperature rise drives IC up toward 1.5 mA. What happens to VE and VBE, and what does it do to the transistor?
A) VE stays at 2 V (clamped by VB); VBE stays at 0.7 V; nothing pushes back. B) VE rises from 2 V to 3 V; VBE shrinks from 0.7 V to −0.3 V exactly, cutting the transistor fully off. C) VE rises from 2 V toward 3 V; VBE shrinks, throttling IB and pulling IC back down before VE ever reaches 3 V. D) VE falls because higher IC means lower resistance; VBE grows, increasing IC further.
Correct answer: C.
At IC = 1 mA, VE = 1 mA · 2 kΩ = 2 V. If IC actually reached 1.5 mA, VE would be 1.5 mA · 2 kΩ = 3 V — but it won't get there, because as VE climbs toward 3 V, VBE = VB − VE = 2.7 − VE shrinks, IB drops, and IC is reined in. The circuit lands at a new equilibrium with VE a bit above 2 V, not the full 3 V the unchecked current would have demanded.
A is the clamped-VE misconception — it denies step 1 and so denies the whole loop. B does the arithmetic but treats it as a static snap to full cutoff rather than a self-limiting process that settles partway. D inverts the physics: RE is a fixed resistor, and the feedback is negative, not regenerative.
Close the gap
The shaky link was a single arrow of causation: does VB set VE, or does the current through RE set VE? The whole feedback loop hangs on getting that one arrow pointing the right way. When a student flags exactly that link as the part that doesn't make sense, the fix isn't to re-explain the whole circuit — it's to find the one premise doing the damage and replace it. That targeted, mid-explanation correction is what Gradual Learning is built to do.