Hartley Oscillator

Part of Vacuum Tubes

The Hartley oscillator uses a tapped inductor to provide the feedback needed for sustained oscillation — one of the simplest and most robust LC oscillator designs for radio frequencies.

Why This Matters

Ralph Hartley invented this oscillator circuit in 1915, and it remains one of the most widely used LC oscillator topologies for good reason: it works reliably over a wide frequency range, starts easily, and requires only a single wound component (the tapped inductor) plus one capacitor to determine the frequency. For community radio transmitters and receivers, the Hartley oscillator provides a simple, buildable VFO (variable frequency oscillator) with good stability.

Compared to the Colpitts oscillator, the Hartley circuit is generally easier to wind — you need one coil with a tap rather than two separate capacitors. The feedback fraction is determined by the position of the tap on the inductor, which can be adjusted during construction by moving a wire connection, giving you physical control over a parameter that is otherwise fixed by component values.

In salvage and field-building scenarios, the Hartley circuit often makes better use of available components. If you have a single good variable capacitor and can wind a coil, you have everything needed for the Hartley VFO. Knowing both Hartley and Colpitts designs lets you choose based on what components are available rather than preference.

Circuit Description

The Hartley oscillator uses a triode in common cathode or grounded-grid configuration with the tank circuit’s inductor tapped. The complete circuit:

The main tuning capacitor (C) connects across the entire inductor (L). This capacitor, together with the full inductance, sets the resonant frequency. The inductor has a tap somewhere between its ends — typically 10-30% of the way from the cathode end.

The tube’s plate connects to one end of the inductor through a DC blocking arrangement (either the B+ supply connects at the top of the inductor, or a choke connects plate to B+ and the plate connects to the inductor end through a capacitor). The cathode connects to the tap. The grid connects to the other end of the inductor through a coupling capacitor and grid-leak resistor arrangement.

The feedback mechanism: the oscillating current in the tank circuit produces a voltage across each section of the tapped inductor. The voltage across the smaller section (cathode to tap) is fed to the cathode. The voltage across the larger section (tap to grid end) appears at the grid. These two voltages are in opposite phase (they are across opposite sections of the same coil), providing the 180° phase shift needed to complement the 180° phase shift of the common cathode amplifier and complete the positive feedback loop.

Component Values and Frequency

The resonant frequency depends on the total inductance and the capacitance:

f = 1 / (2π × √(L × C))

For a broadcast band oscillator (1 MHz): L total = 100 µH, C = 250 pF f = 1 / (2π × √(100µH × 250pF)) = 1 MHz

For a 40m amateur band oscillator (7 MHz): L total = 4 µH, C = 125 pF f = 1 / (2π × √(4µH × 125pF)) = 7.1 MHz

Coil construction for 100 µH: approximately 100 turns of 0.3mm wire on a 25mm diameter form, close-wound, about 50mm long. The tap at 10-15 turns from one end.

Coil construction for 4 µH: approximately 20 turns of 0.5mm wire on a 25mm diameter form, slightly spaced. The tap at 3-4 turns from one end.

The tap position determines the feedback fraction. Too little feedback (tap too close to the cathode end, small section too small) and the oscillator may not start reliably, especially with cold components or at lower supply voltages. Too much feedback (tap too close to the middle) causes the oscillator to produce a clipped output with high harmonic content. A practical starting point: tap at approximately 15-25% of the total turns.

Variable Frequency Operation

A variable capacitor across the complete inductor tunes the frequency. As the capacitor increases, the resonant frequency decreases (longer period, lower frequency). The variable capacitor range determines the tuning range of the oscillator.

For a 40m VFO covering 7.0-7.2 MHz (200 kHz range): At 7.0 MHz with 4 µH: C = 128 pF At 7.2 MHz with 4 µH: C = 121 pF

Required capacitance change: 7 pF over the tuning range. A variable capacitor with 15-30 pF maximum, padded with a fixed 100 pF capacitor, gives this range. The 100 pF padding capacitor reduces the sensitivity of frequency to capacitor position, making fine tuning more controllable.

Dial calibration: after building the VFO, calibrate the dial against a known frequency standard. A WWV time signal receiver (if available) or a calibrated receiver can provide reference frequencies for calibration points. Mark at least three or four frequencies on the dial and interpolate between marks.

Stability and Drift

All LC oscillators drift in frequency as temperature changes because both the inductance (wire length changes with temperature) and capacitance (physical dimensions change) vary. For a VFO to be useful in a transmitter, frequency drift must be small enough that the signal stays within the receiver’s passband.

Causes of drift and their remedies:

Thermal expansion of the coil: wind the coil on a form with a low expansion coefficient. Polystyrene, Teflon, and ceramic coil forms have much lower expansion coefficients than wood or cardboard. A polystyrene form reduces coil drift by a factor of 5-10 compared to cardboard.

Capacitor temperature coefficient: NP0/C0G ceramic capacitors have near-zero temperature coefficients. Silvered mica capacitors are also excellent. Disc ceramic (Z5U, Y5V) capacitors have large, unstable temperature coefficients and must not be used in oscillator circuits. Use NPO capacitors for all capacitances in the tank circuit.

Heater warmup: during the first 10-30 minutes of operation, the tube’s heater warms the surrounding components, causing them to expand and the frequency to drift. Allow a warmup period before using the transmitter for communication requiring precise frequency control. After warmup, drift typically becomes much smaller.

B+ voltage variation: the tube’s transconductance and capacitances change with supply voltage, shifting the oscillation frequency. A regulated power supply for the oscillator stage, or at minimum a decoupled supply with good filtering, reduces frequency pushing from load variations.

Comparison with Colpitts

The Hartley oscillator’s tapped inductor versus the Colpitts oscillator’s capacitive divider makes each better suited to certain situations.

Hartley advantages: easier to realize variable frequency operation (one tuning capacitor instead of two capacitors that must track together or be carefully proportioned). The tap position is adjustable after winding. The circuit starts more reliably because the inductive feedback has low impedance at resonance.

Colpitts advantages: capacitors are more stable and reproducible than inductors with taps. The feedback fraction is determined by capacitor ratio, which is more precise than a coil tap position. For crystal-controlled oscillators (where stability is paramount), Colpitts is preferred.

Practical choice: for a VFO that must tune over a range, use Hartley. For a crystal-controlled fixed-frequency oscillator, use Colpitts (or its close relative, the Pierce circuit). Both circuits work with the same tubes and similar component values.