Ballast in a gas discharge
A Socratic walk-through of ballast in a gas discharge — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a lit neon tube draw ever more current as its voltage falls, so it destroys itself unless something limits it?
Wire a resistor across a battery and it settles at a current and stays there. Wire a neon tube across the same battery and it either does nothing or destroys itself. Every discharge lamp ever built carries a separate component whose only job is to hold the current back. A resistor needs no such chaperone. What does a glowing gas do that a wire does not?
Reasoning it through
REASONING #Ask first what a resistor's obedience rests on. Current is carriers times charge times drift speed times area. In a metal the number of carriers is fixed by the material — roughly one loose electron per atom, whatever you do to it. Raise the voltage and only the drift speed responds, so current rises in proportion. Ohm's law is less a law of nature than a consequence of the carrier count being a constant.
Now a gas. At rest it has essentially no free carriers; it is an excellent insulator, which is why a tube needs a high voltage to strike. So every carrier in a running discharge was made by the discharge: an electron accelerated by the field hits a neutral atom hard enough to knock an electron off it, and now there are two. The steady state balances that manufacture against the losses — recombination, and drift to the walls and electrodes.
The consequence: the carrier count is no longer a constant handed to us by the material but a strong function of the current itself. So when we ask what field is needed to sustain a given current, the answer can go down as the current goes up, because a denser, hotter plasma is easier to keep ionised. The tube's voltage falls as its current rises, and a curve that slopes downward is what people mean by negative differential resistance.
Why is that fatal on a battery? Write the loop, including the small unavoidable series inductance L of any real circuit, a supply Vs, a series resistance R, and a tube characteristic Vt(I):
L dI/dt = Vs - I R - Vt(I).
An operating point is where the right-hand side vanishes. Nudge the current to I + x and expand: L dx/dt = -(R + dVt/dI) x. The disturbance dies away only if R + dVt/dI is positive — that is,
R > -dVt/dI.
For a resistor, dVt/dI is its own positive resistance, so the condition holds with R = 0 and no ballast is needed. For a discharge, dVt/dI is negative, and the condition becomes a real requirement: the series impedance must exceed the magnitude of the tube's downward slope. Fail it and x grows exponentially — more current, less tube voltage, more of the supply left over to push current, more current still — until an electrode melts. Note what that criterion is: a statement about slopes, not about sizes. The ballast need not "use up" the surplus voltage; it needs to be steeper in the right direction than the tube is in the wrong one.
The analogy
THE ANALOGY #Imagine a tap whose valve is worn so that the flow through it drags the valve further open. A trickle opens it a little, which lets more through, which opens it further. Nothing about the tap stops this. The cure is not to fix the tap but to put a fixed narrow orifice in the pipe upstream: once most of the pressure drop is taken across something that genuinely resists more when more flows, the pair is stable.
The worn valve reaches a hard stop when fully open, so its runaway ends by itself; the discharge has no such stop, and the only limits are what the supply delivers and what melts first. The analogy also has no counterpart for the tube's other oddity — that striking it takes far more voltage than running it.
Clarifying the model
THE MODEL #The tidy sentence "more current makes more carriers, so resistance drops" is directionally right and mechanistically sloppy, and it is worth saying where it misleads, because the real characteristic is not one smooth slide. A low-pressure discharge passes through distinct regimes. In the normal glow the discharge covers only part of the cathode, and extra current is accommodated by spreading over more cathode area rather than running any part harder — so the voltage is remarkably flat, not falling. Push past full coverage into the abnormal glow and voltage actually rises with current. The steep negative slope belongs to the transition into an arc, where two thermal feedbacks take over: the gas heats, so at fixed pressure its density drops and the field per molecule rises, and the cathode heats until it emits thermionically, which costs far less voltage than knocking electrons out by ion bombardment. The falling characteristic is real, but it is a thermal story more than a carrier-counting one — and a lamp can sit on a flat stretch and still be unstable, because flat means dVt/dI is about zero and any R above zero suffices, barely.
Put illustrative numbers to the criterion. Suppose a small indicator lamp strikes near 90 V and settles near 60 V once lit — approximate figures, recalled rather than measured — and we want 1 mA from a 120 V supply. Then R = (120 - 60)/0.001 = 60 kilohms. If the tube's voltage droops by 5 V for each extra 0.1 mA, its slope has magnitude 50 kilohms, and 60 exceeds 50, so the point is stable. Run the same lamp from 70 V and R falls to 10 kilohms, well under 50, and it runs away: supply voltage, which looks like a safety matter, is really a stability matter.
A picture of it
THE PICTURE #How to readStart at the rounded terminal at the top, the supply, and follow the current down through the ballast (the slanted box, the component being sized) into the gas column. The two boxes below are the mechanism: the current manufactures its own carriers, which is why the store shape is a population rather than a constant. Everything then turns on the single diamond, and both exits are labelled. The "yes" branch reaches a rounded terminal and the circuit settles. The "no" branch runs back up into the gas column — that back-edge is the runaway, leading nowhere but a dead electrode.
What became clearer
WHAT CLEARED #A resistor is well behaved because its carriers are given to it. A discharge makes its own, so its voltage can fall as its current rises, and a downward-sloping characteristic on a stiff voltage source is a positive feedback loop with nothing to stop it. A ballast supplies not voltage headroom but slope: enough positive differential impedance to outweigh the tube's negative one.
The load-bearing claim is that only the slope matters, and it is directly testable. Drive the tube from a current source instead of a voltage source and the runaway should vanish, because the loop has no leftover voltage to convert into current — the lamp will sit at whatever current you dial and let you plot its falling curve at leisure. If a discharge run from a true current source still ran away, the slope criterion would be wrong. A second test follows: since only differential impedance counts, a purely reactive element with no dissipation should work as well as a resistor. It does, which is why mains fluorescent lamps have historically been ballasted by a choke rather than a resistor that would waste as much power as the lamp.
Where to go next
ONWARD #- Why a tube needs a much higher voltage to strike than to run, and what the striking voltage depends on.
Key terms
TERMS #| Term | What it means |
|---|---|
| Negative differential resistance | a region where voltage across a device falls as current through it rises, so its characteristic slopes downward even though the voltage is positive. |
| Ballast | any series element, resistive or reactive, supplying enough positive differential impedance to stabilise a discharge. |
| Thermionic emission | release of electrons from a hot surface, which in an arc replaces the far more voltage-hungry ion bombardment of a cold cathode. |
Every term the collection defines is gathered in the glossary.