Conditional Stability
Understanding Conditional Stability
In microwave amplifier design, an amplifier is essentially an oscillator that hasn't started oscillating yet. Conditional Stability (also known as potential instability) describes a dangerous operating state where a transistor amplifies signals perfectly when connected to ideal laboratory 50-ohm test equipment, but will violently erupt into spontaneous oscillation if it encounters specific mismatched (non-50-ohm) impedances at its input or output ports.
Because real-world antennas are never a perfect 50 ohms across all frequencies—and often change impedance wildly due to environmental factors like ice, physical damage, or proximity to metal—deploying a conditionally stable amplifier directly to an antenna is highly risky. If the antenna's return loss happens to fall into the unstable impedance region of the Smith Chart, the reflected RF energy will feed back through the transistor's internal parasitic capacitance (S12), initiating a runaway positive feedback loop that will instantly melt the semiconductor.
Rollett's Stability Factor (K)
Engineers evaluate stability using S-parameters. By calculating the Rollett Stability Factor (K) and the delta matrix (Δ), the designer can mathematically prove stability. If K > 1 and |Δ| < 1, the amplifier is Unconditionally Stable, meaning no passive impedance on earth can make it oscillate. If K < 1, it is Conditionally Stable. To fix a conditionally stable transistor, the designer must aggressively add resistive loading (sacrificing gain and noise figure) or use negative feedback to force K > 1 before matching the network.
|Δ| = |S11 S22 - S12 S21|
Stability Factor (K):
K = (1 - |S11|2 - |S22|2 + |Δ|2) / (2 × |S12 S21|)
If K > 1 and |Δ| < 1 → Unconditionally Stable.
If K < 1 → Conditionally Stable (Oscillation risk).
Comparison
| Stability State | K Factor | Smith Chart Implications | Required Action |
|---|---|---|---|
| Unconditionally Stable | K > 1 | Stability circles lie entirely outside the Smith Chart. | Design matching networks for maximum gain. |
| Conditionally Stable | 0 < K < 1 | Stability circles intersect the Smith Chart. | Must map forbidden zones; add resistive loading. |
| Negative Resistance | K < 0 | S11 or S22 is > 1 (Amplifier reflects more than it receives). | Do not use as an amplifier; Use to build an oscillator. |
Frequently Asked Questions
Why don't transistor manufacturers just make them all unconditionally stable?
To make a transistor unconditionally stable at all frequencies, you have to kill its high-frequency gain and completely neutralize its internal feedback (S12). At extremely high frequencies (e.g., millimeter-wave), transistors naturally run out of gain. If you add resistive stabilization to force K > 1, the transistor might have no gain left to actually amplify the signal. Therefore, high-frequency transistors are almost always conditionally stable by default.
What is a Stability Circle on a Smith Chart?
A stability circle is a boundary drawn on the Smith Chart. Any impedance (point) chosen inside the 'forbidden' region of the circle will cause the amplifier's input or output reflection coefficient to exceed 1.0, triggering oscillation. The engineer must design the matching network so that the impedance stays strictly in the 'safe' region.
Can I use a conditionally stable amplifier if I put an isolator on the output?
Yes. An RF isolator is a ferrite component that acts like a one-way valve. It passes the RF signal to the antenna but absorbs any reflections bouncing back. This guarantees the amplifier always 'sees' a perfect 50-ohm load regardless of what the antenna does, keeping the conditionally stable amplifier perfectly safe.