Understanding Crystal Oscillator Circuits

Basic solid state crystal oscillator circuit configurations are today more developed, almost all circuits being modifications of the widely recognized vacuum tube systems like the Pierce, Hartley, Clapp and Butler oscillator and work with both bipolar and FET devices.

Although all these circuits fundamentally meet their designed objective, there are plenty of applications that call for something completely different or where functionality requires to be accurately described.

Listed below are a range of circuits, for a variety of applications right from LF through the VHF range, which are not typically seen in existing amateur usage or books.

Basic solid state crystal oscillator circuit techniques are by now well established, most circuits being adaptations of the well-known vacuum tube technology such as the Pierce, Hartley, Clapp and Butler oscillator and use both bipolar and FET devices.

Whilst these circuits basically fulfill their intended purpose, there are many applications which require something different or where performance needs to be reliably characterized.

Presented here are a variety of circuits, for a range of applications from LF through the VHF range, that are not commonly found in current amateur use or literature.

MODES OF OPERATION

A point rarely valued, or simply overlooked, is the fact that quartz crystals can oscillate in a parallel resonant mode and a series resonant mode. The two frequencies are split up with a minor difference, usually 2-15 kHz over the frequency range.

The series resonant frequency is smaller in frequency compared to parallel.

A specific crystal designed for use in the parallel mode might be appropriately applied in a series resonant circuit should a capacitor equivalent in magnitude to its exact load capacitance (typically 20,30, 50 or 100 pF) is attached in series with the crystal.

Unfortunately, it isn't possible to invert the task for series resonant crystal in parallel mode circuits. The series mode crystal will probably oscillate beyond its calibrated frequency in its situation and might not be feasible to capacitively load it down enough.

Overtone crystals run in the series mode generally on the third, fifth or seventh overtone, and the manufacturer usually calibrates the crystal in the overtone frequency.

Running a crystal in the parallel mode and multiplying the frequency 3 or 5 times generates rather a new outcome by operating precisely the same crystal in the series mode upon its 3rd or 5th overtone.

While buying overtone crystals stay away from dilemma and identify the frequency you would like, instead of the apparent fundamental frequency.

Fundamental crystals within the range 500 kHz to 20 MHz are generally built for parallel mode functioning however series mode operation could be asked for.

For low frequency crystals up to 1 MHz, either mode could be chosen. Overtone crystals normally cover the range 15 MHz to 150 MHz.

WIDE RANGE or APERIODIC OSCILLATORS

Oscillators that never make use of tuned circuits are often very useful, whether as ‘crystal checkers’ or any different reason. Especially for LF crystals, tuned circuits could be rather huge.

On the other hand, they usually are not without their own traps. A few crystals are susceptible to oscillation on undesirable modes, specially the DT and CT cut crystals intended for LF quartz oscillators.

It truly is a good idea to make sure that the output is on the proper frequency and no "mode instability" is apparent. Minimizing feedback at the higher frequencies commonly solves this.

In special cases, the above theory can be forgotten and an oscillator possessing a tuned circuit applied as an alternative, (LF crystal oscillators are reviewed afterwards).

Crystal Circuits

The first circuit below is an emitter-coupled oscillator, a variation of the Butler circuit. The output of the circuit in Fig. 1 is basically sine wave; decreasing the emitter resistor of Q2 boosts the harmonic output.

As a result, a 100 kHz crystal generates excellent harmonics via 30 MHz. It is a series mode circuit.

A range of transistors can be employed. For crystals above 3 MHz, transistors having a high gain-bandwidth product are advised. For crystals within the 50 kHz to 500 kHz assortment, transistors with high LF gain, like the 2N3565 are preferred.

Additionally, for crystals within this selection, allowable dissipation is normally lower than 100 microwatts and amplitude constraining might be essential.

Reduced supply voltage, in step with efficient starting up, is suggested. Altering the circuit through the inclusion of diodes as shown in Fig. 3 is a more beneficial technique, and starting efficiency is enhanced.

The circuit is going to oscillate at as high as 10 MHz using suitable transistors and emitter resistor values. An emitter follower or source follower buffer is usually recommended.

Identical comments to the above connect with Fig. 2. An emitter follower buffer is incorporated within this circuit.

The two circuits are somewhat sensitive to frequency and to power voltage variations and load specs. A load of 1 k or higher is recommended.

TTL lC could be combined with crystal oscillator circuits although numerous published circuits possess terrible starting efficiency or experience non-repeatability due to vast parameters in lC's,.

The circuit in Fig. 4. has been experimented with by the author on the range 1 MHz to 18MHz and will be encouraged. This is a series mode oscillator and compliments AT-cut crystals.

The output is around 3 V peak to peak, square wave up to about 5 MHz above which this turns into more similar to half-sine pulses. Starting efficiency is superb, which appears to be mostly a critical factor with TTL oscillators.

LOW FREQUENCY CRYSTAL OSCILLATORS

Crystals within the range 50 kHz to 500 kHz demand distinctive factors not spotted in the more prevalent AT or BT cut HF crystals.

The similar series resistance is a lot bigger and their allowable dissipation is restricted to under 100 microwatts, ideally 50 microwatts or lower.

The circuit in Fig. 5 is a series mode oscillator. It offers the benefit of not needing a tuned circuit, and features a choice of sine or square wave output. For crystals within the spectrum of 50-150 kHz, 2N3565 transistors are advised even though publisher finds BC107's reasonable.

Both the variety may be adequate for crystals within the range 150 kHz to 500 kHz. If you think the crystal includes a big equivalent series resistance, then you can increase the value of R1 to 270 ohms and R2 to 3.3 k.

For square wave operations, C1 is 1 uF (or perhaps a magnitude alongside, or bigger than it). For sine wave output, C1 is not in circuit.

Amplitude control is needless. Sine wave output is approximately 1 V rms, square waive output around 4 V peak to peak.

The circuit in Fig. 6 is actually a revised type of the Colpitts oscillator, with the inclusion of resistor Rf to regulate feedback. Capacitors C1 and C2 must be minimized through calculated magnitudes as the frequency is increased.

At 500 kHz, values for C1 and C2 must be approximately 100 pF and 1500 pF correspondingly. The circuit as proven offers sine wave output using the second harmonic around 40 dB lower (or higher).

This is often minimized through mindful tweaking of Rf and C1. Remember that, at the decreased amount a feedback is essential to accomplish this, it requires around 20 seconds for the oscillator to attain full output.

Output is around 2 to 3 volts peak to peak. When you need an output loaded with harmonics, the easy inclusion of a 0.1 uF capacitor over the emitter resistor will accomplish that. Output subsequently increases to around 5 V peak to peak.

Power supply voltage could be decreased in such cases to reduce crystal dissipation. Other transistors may be used, although bias and feedback might have to be tweaked. For cantankerous crystals designed to oscillate in modes besides those you would like, the circuit of Fig.7 strongly suggested

Feedback is governed by a tap along the collector load of Q1. Amplitude confining is important to maintain the crystal dissipation inside boundaries. For 50 kHz crystals the coil needs to be 2 mH and its resonating capacitor 0.01 uF. Output is approximately 0.5 V rms, fundamentally a sine wave.

The utilization of an emitter follower or source follower buffer highly recommended.

In case a parallel mode crystal is utilized the 1000 pF capacitor indicated in series with the crystal must be changed to the crystal's selected load capacitance (typically 30, 50 to 100 pF for these types of crystals).

HF CRYSTAL OSCILLATOR CIRCUITS

Solid state designs for the well-known AT-cut HF crystals tend to be legion. But, results not necessarily what you might expect to have. The majority of essential crystals up to 20 MHZ are typically chosen for parallel mode functioning.

Nevertheless, this kind of crystals may be used in series mode oscillators by positioning the desired load capacitance in series with the crystal as stated earlier. The two types of circuit are discussed below.

A good oscillator for 3 to 10 MHz range that doesn't demand a tuned circuit is presented in Fig. 8 (a). It is naturally, the same circuit as Fig.6. The circuit works extremely well down to 1 MHz when C1 and C2 are higher than 470 pF and 820 pF respectively. It may be utilized to 15 MHz in the event C1 and C2 are decreased to 120 pF and 330 pF. respectively.

This circuit is advised for noncritical purposes in which large harmonic output is desired, or not an option. The inclusion of a tuned circuit as in 8b minimizes harmonic output significantly.

A tuned circuit having a substantial Q is usually recommended. In a 6 MHz oscillator, We have attained the below results. Having a coil Q of 50 the 2nd harmonic was 35 dB all the way down.

Having a Q of 160, it had been -50 dB! Resistor Rf could be altered (increase a bit) to enhance this. The output is additionally raised using a high Q coil.

As formerly observed, with decreased feedback it requires several tens of seconds to achieve 100 % output from turn on, even so, frequency stableness is fantastic.

Functioning at different frequencies can be achieved by adjusting the capacitors and coil effectively.

This circuit (Fig. 8) could also be changed into an extremely useful VXO. A tiny inductance is defined in series with the crystal and one of the capacitors within the feedback circuit is used as a variable type.

A common two-gang 10-415 pF transmitter tuning capacitor will perform the task perfectly. Each gangs are conected in parallel.

The tuning range is determined by the crystal, the inductance of L1 and the frequency. A larger range is generally accessible using the higher frequency crystals. Stability is extremely good, getting close to that of the crystal.

A VHF OSCILLATOR-MULTIPLIER

The circuit in Fig.10 is a modified version of the ‘impedance inverting‘ overtone oscillator. Typically, applying the impedance inverting circuit the collector is either untuned or grounded for RF.

The collector could be tuned to two times or 3 times the crystal frequency in order to minimize the output at the crystal frequency, a 2x tuned circuit is proposed.

YOU SHOULD NEVER tune the collector to the crystal frequency, or else the circuit may oscillate with a frequency which may be out of the control of the crystal. You need to maintain the collector lead very small and one on one as much as you can.

End results using this type of circuit were great. Just about all outputs besides the desired output had been at -60 dB or higher.

Noise production reaches at least 70 dB under the desired output. This creates an outstanding conversion oscillator for VHF/UHF converters.

Practically 2 V of RF can be obtained on the hot terminal of L3 (author's original at 30 MHz). A Zener regulated supply strongly recommended.

As pointed out within the diagram, various circuit values are essential for various transistors. Strays in specific structure could also require modifications. L1 may be used to move the crystal on frequency. Minor modifications in frequency (about 1 ppm) take place while adjusting L2 and L3 as well as using load variations. Having said that, in real testing, these things could be insignificant.