Testing

Part of Semiconductors

Instruments, methods, and test circuits for characterizing semiconductor devices and verifying circuit function.

Why This Matters

Without measurement, semiconductor fabrication is blind. You cannot tell if a transistor has acceptable gain, if a diode leaks too much, or if your purification produced the right resistivity. Testing provides the feedback that drives process improvement and the verification that enables reliable circuit design.

For a rebuilding civilization, test equipment is a foundational investment. A basic multimeter (voltmeter + ammeter + ohmmeter) enables component verification and circuit debugging. An oscilloscope reveals signal behavior invisible to DC instruments. A curve tracer gives complete device characterization. Each instrument built extends what is knowable about the circuits and devices you are producing.

The test instruments themselves are electronic circuits — they are built from the same components they measure. This circularity requires careful bootstrap: start with the simplest instruments built from known-good components (resistors, capacitors, simple tubes if available), use these to characterize semiconductors, then use characterized semiconductors to build better instruments.

Essential Test Instruments

Voltmeter: The most fundamental instrument. A sensitive galvanometer (moving-coil meter) with series resistors creates a multi-range voltmeter. Galvanometers with full-scale deflection of 50-100 µA are suitable. Series resistor for 1V range with 100 µA movement: R = (1V - 0.1V) / 100 µA = 9 kΩ. For 10V range: 99 kΩ. For 100V: 999 kΩ. Use precision resistors (wound wire resistors for stability) and calibrate against a known reference.

Input impedance matters: the voltmeter draws current from the circuit being measured, disturbing it. A 100 µA movement with 10 kΩ/V sensitivity (10 kΩ in 1V range) loads circuits with 100 kΩ on the 10V range — this disturbs high-impedance circuits. For transistor measurements, use a high-impedance voltmeter (vacuum tube voltmeter or FET input operational amplifier when available).

Ammeter: A galvanometer with shunt resistors. Shunt for 1 mA range with 100 µA movement: shunt = (V_full_scale) / (1 mA - 100 µA) = 0.9V / 0.9 mA ≈ 1 Ω (where V_full_scale is the meter’s internal resistance × 100 µA, typically 0.01-0.1V). Include protection diodes across the galvanometer to prevent burnout if overloaded.

Ohmmeter: Battery, series current-limiting resistor, galvanometer, and scale calibrated in ohms. The meter reads full-scale (0 Ω) with probes shorted; reads zero (infinite Ω) with probes open. Non-linear scale (ohms is reciprocal of current). Best range is around mid-scale (typically 100-1000 Ω for most circuits). Use for quick component checks and continuity testing.

Oscilloscope: Cathode ray tube (CRT) with horizontal sweep and vertical deflection amplifiers. The horizontal sweep generates a sawtooth wave that moves the electron beam across the screen at adjustable speed; the vertical deflection is proportional to the input signal. Displays signal waveforms in real time.

Building a CRT oscilloscope requires a vacuum tube CRT plus a high-voltage supply (~1-5 kV), sweep generator, and vertical amplifier. This is an advanced project — build it once basic semiconductor fabrication is established. The oscilloscope then enables measurement of signal quality, frequency response, and timing in all subsequent circuit development.

Transistor Parameter Measurement

hFE measurement jig (described fully in quality-sorting): fixed base current source, collector current meter, calculated ratio. Build a dedicated test socket and resistor-switch network so any transistor can be measured in 5-10 seconds.

VBE measurement: Connect transistor as a forward-biased diode (base-emitter junction). Apply 1 mA through a 1 kΩ series resistor from 5V. Measure voltage across base-emitter with voltmeter. This is VBE at 1 mA.

ICEO measurement: Base terminal open (no connection). Apply 9V collector-to-emitter through 10 kΩ series resistor. Measure voltage across resistor; ICEO = V/10000. Values below 1 µA for silicon and below 10 µA for germanium are acceptable.

BVCEO measurement: Collector-emitter with base open. Start with collector at 0V relative to emitter (NPN). Slowly increase until collector current suddenly increases. The voltage at this transition is BVCEO. Do not exceed this voltage or the device may avalanche destructively — limit current with a 10 kΩ series resistor. As soon as current increases noticeably, record the voltage and stop.

Leakage temperature coefficient test: Measure ICEO at room temperature, then hold the device between thumb and finger for 30 seconds (heating to ~35°C) and remeasure. Germanium: leakage should roughly double per 10°C. If it increases 10× or more, the device has contamination-enhanced generation and will fail in warm environments.

Four-Point Probe for Material Characterization

Essential for measuring semiconductor material resistivity before and after purification and doping.

Construction: Four tungsten needle probes (from compass needles, fine sewing needles, or machined tungsten rod) mounted in a row with 1 mm spacing in an insulating block (Bakelite, hard rubber, or ceramic). Each probe spring-loaded to maintain contact pressure. Leads to terminals 1-4.

Measurement: Force current I (e.g., 1 mA) through outer probes 1 and 4. Measure voltage V between inner probes 2 and 3.

Resistivity calculation (for sample much larger than probe spacing):

• For bulk samples (thick): ρ = 2π × s × V/I, where s = probe spacing (1 mm = 0.1 cm). ρ = 2π × 0.1 × V/I ≈ 0.628 × V/I (Ω·cm).
• For thin slabs (thickness t << s): ρ = (π/ln2) × t × V/I ≈ 4.53 × t × V/I.

Calibration: Measure resistivity of a known-resistivity sample (carbon composition resistors of known value can serve as a rough check on the probe geometry correction factor).

For germanium: acceptable range for transistor fabrication is 0.1-10 Ω·cm. For silicon: 0.1-100 Ω·cm depending on intended doping.

Basic Circuit Debugging Procedure

When a circuit fails to work as designed, systematic debugging finds the problem faster than random probing:

1. Power supply check: Measure supply voltage with no circuit connected. Verify it is correct value and stable. Measure with circuit connected — if voltage drops significantly, the circuit is drawing excessive current (short circuit, or wrong component value).

2. DC operating point check: For each transistor stage, measure VB, VE, VC relative to ground. Compare to calculated values. Transistor cut off (VC ≈ Vcc): check base bias network, check transistor for continuity. Transistor saturated (VC ≈ ground): check collector resistor, check base resistor, check for excessive base current.

3. Signal trace: Inject a test signal (from an audio oscillator or a square wave from a simple multivibrator) at the input. Trace with oscilloscope or headphones through capacitive probe (10 nF capacitor in series with headphone) from input to output, stage by stage. The stage where signal disappears is the problem.

4. Component substitution: When a stage is suspect, substitute known-good components one at a time. Replacing the transistor most efficiently diagnoses transistor failure.

5. Isolation: Connect stages one at a time. Each stage that is verified working is not the problem. Narrow the problem to a single stage, then to the specific component.

Keep a fault log: for each circuit worked on, record what the symptom was, what the cause turned out to be, and how it was fixed. This builds diagnostic intuition faster than any textbook.