Most of us have found, from experience with crystal oscillators, that the 6L6 does not always perform according to predictions based on upon the use of well-screened tubes. This article tells how to get the most satisfactory results with the 6L6 both at the fundamental and at the second harmonic.

Two of the most popular crystal-oscillator circuits in use these days are the Tri-tet and the "grid-plate" circuits. Their popularity is well deserved, because they have several desirable features. In the first place, they are frequency-multiplying oscillators; that is, they will deliver output at the harmonics of the crystal frequency as well as the fundamental. This not only a matter of importance in the simple oscillator transmitter, where it makes it possible to operate in more than one band with one crystal, but also in the multistage transmitter, where the number of required stages may be reduced. This is particularly desirable when there is no provision in the transmitter for cutting out doubler stages which must be well stabilized for fundamental operation.

Another feature of these circuits is that they exhibit characteristics similar to those of an oscillator-amplifier combination. In a true Tri-tet or grid-plate oscillator, the functioning of the oscillator circuit is, for most practical purposes, independent of loading and tuning of the plate circuit. The nuisance of critical adjustments of loading and tuning may be eliminated entirely. This is in contrast to the familiar behavior of the triode, tetrode or pentode "X"-grid, tuned-plate circuits where the plate circuit must be detuned to a rather critical point of resonance where less than maximum output is obtained for the sake of reliable starting and maintenance of oscillations.

As pointed out by James Lamb in his excellent treatment of crystal oscillators [Note 1], the more complete the screening between control grid and plate, the more independent become the loading and tuning of the plate circuit. Such tubes as the 802 and RK23-25 are recommended as the best tubes for the purpose. Unfortunately, the popular, less-expensive 6L6 is an audio tube and is not so well screened and invariably its performance does not conform to that usually predicted for the circuit in use. While it works beautifully in the Tri-tet circuit with the plate circuit tuned to the crystal harmonic, it invariably exhibits characteristics more those of the tetrode circuit at the fundamental, particularly when the circuit is unloaded.

With this condition, crystal current will run high near plate-circuit resonance and oscillation will cease at exact resonance, whereas with the true Tri-tet, crystal current is low when the circuit is not loading and the circuit oscillates without interruption as the plate circuit is tuned through resonance. This high crystal current with the 6L6 explains the frequent crystal fractures which are often experienced with this arrangement. However, as the loading is increased the crystal current does not drop off rapidly as might be expected with tetrode operation, indicating that excitation is being held up by Tri-tet action. In spite of this, however, the plate circuit must be detuned so far to the high-frequency side of resonance for reliable keying or starting of oscillation that we actually end up with less output and poorer efficiency at the fundamental than at the harmonic.

In a multi-stage transmitter or an exciter, this may not always be of great consequence, since the oscillator is often followed by a stage of high power-amplification requiring less than maximum output from the oscillator for adequate excitation. In the simple oscillator transmitter, however, it may become of greater importance, since tests have shown that it may be necessary to detune the plate circuit for such an extent for good keying thta the output is reduced to as low as 20 percent of the maximum obtainable. Such operating, of course, increases enormously the power which the tube is required to dissipate.

With the grid-plate oscillator circuit, we find conditions reversed. The circuit performs well at the fundamental with the oscillating circuit proper functioning quite independently of the plate circuit. The circuit does not perform as well as the Tri-tet at the second harmonic, however. In fact, it is only by virtue of the fact that the screening of the 6L6 is incomplete that any appreciable even harmonic output may be obtained. This may be demonstrated by the substitution of a well-screen tuber such as the 802. Under similarly circumstancs, the harmonic output of the 6L6 will not exceed 60 percent or so of that obtainable wit the Tri-tet arrangement and considerable difficulty may be encountered in obtaining an adjustment which will permit satisfactory keying of the circuit. Experience has shown that it is seldom possible to obtain any second-harmonic output from a 1.7-Mc. crystal in this circuit.

We now come to the conclusion that what we want is a combination of the two circuits—the grid-plate circuit for fundamental output and the Tri-tet for harmonic operation. This can be accomplished by a switching system in the cathode circuit, or perhaps preferably by a plug-in arrangement, which will permit changing the circuit as well as circuit values to optimum figures for each band. Figure 1 shows the circuit.

Figure 1—Circuit diagram of the combination 6L6 oscillator. Crystal frequencies for each combination are shown to column to left of plug-in-unit diagrams

In selecting the best circuit values, several series of tests were made. These tests showed several interesting things. In the Tri-tet circuit, the values of inductance and capacity to be used for each band are fairly critical. For instance, it was found that 7- and 1.75-Mc. crystals operated well with capacity of 100 µµfd. Improvement was shown, however, by increasing this capacity to 200 µµfd. for 3.5-Mc. crystals. With the inductance in use (approximately 3.7 µh.) the circuit resonated at about 6 Mc. A new could was wound to resonate at 6 Mc. with the 100-µµfd. capacity, but this resulted in much reduced output and greatly increased crystal current. The coil sizes given for L2 should be followed closely.

The cathode resistance is quite essential for good keying and ready starting of the oscillator. Its value should not be too high, however, since it results in reduced output, especially when operating at the crystal fundamental. Higher values of grid leak up to 0.1 megohm will result in somewhat better place-circuit efficiency, especially at the harmonic, but the higher values make it impossible to obtain maximum power output and result in much higher orders of crystal current.

With 7-Mc. crystals, it is possible to reduce crystal current appreciably without affecting the output in the least by inserting a small capacity, C1, in series with the crystal. With lower-frequency crystals, however, this condenser must be short-circuited for satisfactory operation. Since crystal currents do not run so high with lower-frequency crystals, the series condenser is unnecessary.

Screen-voltage adjustments did not prove to be critical once optimum circuit values were determined. There seems to be not point in running the screen voltage higher than about sixty per cent of the plate voltage higher than about sixty per cent of the plate voltage, since the output does not increase appreciably and the only results may be unnecessary heating of the screen and increased crystal current.

The same values of grid-leak resistance, cathode resistance and screen voltage seemed to give about optimum output when working at the crystal fundamental with the grid-plate circuit. Somewhat greater output may be obtained by eliminating the grid-leak resistance entirely, but the efficiency suffers.

While the usually recommended value of 100 µµfd. for the capacity shunting the cathode choke in the grid-plate circuit seemed to be optimum for 3.5- and 7-Mc. crystals, best operation with 1.7-Mc. crystals was obtained with a value of 200 µµfd.

In the circuit of Figure 1, an arrangement is shown for shifting the circuit, as well as circuit values for each band, by plugging in the appropriate unit. The extra 100-µµfd. cathode-circuit capacity for fundamental operation with a 1.7-Mc. crystal and 7-Mc. output with a 3.5-Mc. crystal are soldered to the appropriate pins. Contacts 2 and 3 are strapped togehtr when either 1.7- or 3.5-Mc. crystals are in use thereby short-circuit the crystal series condenser, C1. The cathode-circuit choke is short-circuited when the Tri-tet circuit is in use to eliminate lose in output which was especially noticeable with 1.7-Mc. crystals.

With the arrangement shown, it has been possible to operate the 6L6 at inputs up to 55 watts at 600 volts (measured between plate and ground) without exceeding crystal-current ratings; in most cases, the crystal current did not exceed 60 ma. under any adjustment. A 60-ma. dial lamp, which will burn out at about 100 ma., in series with the crystal from excessive current which may result should the wrong cathode-circuit unit be plugged in by mistake.

When either circuit is operating properly there should be no cessation of oscillation at any point over the tuning range of the plate tank condenser. Unloaded, the plate current will run high until resonance is approached when the customary dip will be found, the minimum value of plate current depending upon the frequency, less dip being obtained at the higher frequencies. As the loading is increased, the dip will become less pronounced and, in fact, disappear entirely when heavily loaded so that the plate-current reading no longer serves as a reliable indicator of the circuit tuning. The rectified grid current of a following stage, or antenna ammeter when the oscillator is coupled to an antenna, should be used to tune for maximum output. Since the tank condenser is large enough to cover both the fundamental and harmonic frequencies of any crystal, care should be taken that the circuit is tuned to the desired frequency. Resonance will occur at high capacity for the 1.7- and 3.5-Mc. bands, medium capacity at 7 Mc., and low capacity at 14 Mc.

Outputs as high as 35 to 40 watts at the fundamental and 25 watts at the second harmonic, 14 Mc., included, with good keying characteristics should not be difficult to obtain with a proper power supply.