Using the Nuvistor on V.H.F. Bands
August 1962 Electronics World

August 1962 Electronics World   Table of Contents  Wax nostalgic about and learn from the history of early electronics. See articles from Electronics World, published May 1959 - December 1971. All copyrights hereby acknowledged.

The nuvistor triode amplifier, a component largely unknown to most people, even during its brief heyday in the late 1950s through the early 1960s, was an attempt to bridge the technology gap between the sunsetting of vacuum tubes and solid state transistors. Its relatively low noise figure, small size, ruggedness, metal shield, and higher operating frequency (near a gigahertz) were the main marketing features when RCA introduced the nuvistor in 1959. Although adopted enthusiastically by many aware of its existence, the nuvistor, as pointed out in this 1962 Electronics World article, tended to oscillate if its metal body was not adequately grounded. You might expect the schematics included in the article to explicitly show a ground connection to the can, but that is not the case. Ultimately, a quick adoption of transistors by industry and hobbyists, with their rapidly improving specifications and decreasing price, rendered the nuvistor to the dustbin of electronic history.

A nuvistor was used in the front end of the HP3400A True RMS Voltmeter (thanks to Michael M. for that). Also, thanks to Bob Davis for pointing out that the Lafayette HB-400 CB radio sported a Nuvistor in the receiver front end. Here is a Nuvistor info resource. See the "Nuvistors and Micro-Modules," "The Tube Family Tree," "Lafayette Radio Electronics," "Hitachi Nuvistor Ad," "Using the Nuvistor on V.H.F. Bands," and the "The NASA 136" articles, too.

Using the Nuvistor on V.H.F. Bands

By Joseph Marshall WA4EPY

Very few tubes have won the rapid and enthusiastic acceptance that the miniature vacuum-tube nuvistor triode, Type 6CW4, has. Although it may have been designed with the idea of furnishing some competition to transistors in compact applications, it is not its small size but its big performance which has won its popularity, especially among radio amateurs operating in the v.h.f. bands.

The 6CW4 is capable of performance as an r.f. amplifier which had previously been possible only with extremely expensive special tubes, such as the 5842/417A. In addition, it makes possible practical and usable receiver sensitivities well below one microvolt in the 50- to 200-mc. region.

However, as is the case with most good things, high performance is not always obtained without some trouble and problems. Many an amateur, or for that matter professional, has run into trouble trying to make the nuvistor live up to its promise. Actually, if a few precautions are taken it presents no really insurmountable problems and is somewhat easier to handle than most tubes at v.h.f. frequencies.

Table 1 gives its operating characteristics as published by RCA. It will be noted that the 6CW4 has a high transconductance and very low grid-plate capacitance. Although the table doesn't show this, it also has very low lead inductances and a high input resistance. It can be used with either grid-leak or cathode biasing. With grid-leak biasing, the equivalent noise resistance is around 200 ohms, and with cathode bias about 260 ohms. This is not remarkably low since there are several larger tubes which equal or better this noise resistance. However, like the 6AK5, the 6CW4 delivers a better noise figure in practical applications than a strict comparison of equivalent noise resistance would indicate. Noise figures of 3 db or less are readily achievable up to 150 mc. and, with care, through the upper TV and 220-mc. amateur bands.

The simplest circuit for the nuvistor is the neutralized triode or neutrode r.f. amplifier. Fig. 1 shows four versions. Fig. 1A uses grid-leak bias and inductive neutralization. Fig. 1B uses capacitive neutralization with grid-leak bias. Fig. 1C combines cathode bias with inductive neutralization while Fig. 1D uses capacitive neutralization and cathode bias. The 6CW4 can provide gains of 25 to 50 in tuned r.f. stages in any of these circuits. The biggest problem is keeping the amplifier stable.

Instability Problems

Many, and perhaps most, of the instability problems can be attributed to one rather simple cause. The tube shell needs to be well grounded. The grounding is achieved through the socket but the socket available for these tubes presents problems in grounding, particularly in home-brewed units. The socket was obviously designed for installation in mass-produced equipment in which proper slots and holes are punched and the socket is mechanically bonded to the chassis. It is difficult with the equipment available to most amateurs and experimenters to mount this socket as it was intended to be mounted. Even when this is managed, the grounding is not sufficiently effective. The author has found that the best solution is to use a thin copper plate for mounting either the socket alone or the entire r.f. stage. A single hole is drilled for the body of the socket; the mounting ears of the socket are turned up; the socket is slipped in the hole from the top and the mounting ears are then soldered to the copper plate. If the nuvistor is to replace a 7- or 9-pin miniature tube, the copper plate can be cut just large enough to replace the socket.

Even when the socket is completely grounded through a good mechanical and solder joint to the copper, the shell of the tube may not ground perfectly because it doesn't always make firm and low-resistance contact with the grounding lugs on the socket. The grounding lugs can be pressed inward to narrow the gap. Also, the grounding lugs on the nuvistor shell should be cleaned with contact cleaner and perhaps a very fine abrasive so that a good low-resistance electrical contact is assured. Once the nuvistor shell is well grounded, the tube presents no problems that do not occur with any other "hot" tube.

Most experimenters are attracted by the grid-leak bias configuration because the higher transconductance promises a better noise figure. The author has had less trouble with cathode biasing probably because (1) the slightly lower transconductance gives a better margin of stability and (2) cathode biasing probably provides a slight degenerative effect. Despite this, gain and noise figures have been obtained which are not significantly poorer. After all, a stable amplifier with slightly lower gain is much more useful and has a better noise figure than an unstable amplifier with higher gain.

Neutralization

Similarly, although inductive neutralization is possibly more attractive, the author has experienced less trouble with capacitive neutralization. There are two problems with inductive neutralization. Because of the very low grid-to-plate capacitance, the neutralizing coil has to be rather large. But it must be positioned so that it is not mutually coupled to the input and output coils. If there is significant coupling, the coil will tend to increase feedback rather than neutralize it. The compactness of the nuvistor fosters the crowding of parts and in the upper half of the v.h.f. range, they must be crowded to keep the distributed inductances low. It isn't easy to place the neutralizing coil in a position which both minimizes coupling to the other coils and keeps the lead lengths short.

Much of the trouble with inductive neutralization is often traceable to a very simple cause - the use of slug-tuned coils in which the slug is rather loose and therefore changes position with vibration. The neutralizing slot for these tubes is very sharp; even a slight movement of the slug can shift the parameters so that the tube goes out of neutralization. A simple but effective solution is to slip a cylinder of paper of appropriate thickness between the core and the inside of the coil form so the slug fits smoothly but snugly and cannot wobble inside the coil. A number of cases of instability have been cured in this simple manner.

Capacitive neutralization fits in nicely with an arrangement of components which also favors isolation between input and output circuits. As in the case of larger tubes, it is a good idea to place a shield across the nuvistor socket to isolate the grid and plate circuits and pins. The 47-μμf. capacitor can then be a feedthrough placed in this partitioning shield and the 2-8-μμf. neutralizing capacitor can be supported at one end by the feedthrough terminal and at the other by a terminal on the input coil, tuning capacitor, or a stiff wire from coil to tube grid. A miniature variable of the MAC type, in its smallest size, is most convenient for neutralization. However, a gimmick of twisted wires can be used although it is more trouble to adjust for complete neutralization. Because no tuned circuit is involved, capacitive neutralization is less critical in original adjustment and maintains complete neutralization over a wider bandwidth.

Whichever method of neutralization is used, the greatest range of stability is achieved when the neutralization is adjusted for the low edge of the band of frequencies to be received, or even slightly below - 45 to 50 mc., for instance, for a 6-meter preselector or converter and 88 to 90 mc. for the FM broadcast band.

The simplest way to neutralize is to break the filament circuit, feed a strong signal into the unit and adjust the coil or capacitor for minimum output. A slightly better noise figure is achieved if a noise generator is available and the neutralization is adjusted for the best noise figure. However, the difference is not significant below the 2-meter band and adjusting for lowest feed-through provides a noise figure well below the antenna noise.

A good enough noise figure can be achieved through the 6-meter and even the FM broadcast band without neutralization in the circuit diagrammed in Fig. 2. Here the plate tank, consisting of the choke and fixed capacitor, is tuned roughly to a frequency well below the operating frequency - 50 or 60 mc. for an FM broadcast-band tuner or 30 to 40 mc. for a 6-meter unit. This provides a low-impedance load at the operating frequency. This low impedance, of course, reduces the gain; however, this is partly compensated by a small amount of regeneration. Thus the gain is not lowered too much and the noise figure is still good enough to produce an IHFM sensitivity of 1.5 μv. or better in the FM broadcast band.

Cascode Circuits

Two nuvistors can be used in the cascode circuit. The author's attempts along this line have not been encouraging and many of the difficulties experienced in working with nuvistors have been due to the attempt to make them operate in cascode. The noise figure of the series cascode is not as good as that of the neutrode unless the plate-to-grid and plate-to-ground capacitance of the first section are both neutralized, as diagrammed in Fig. 3. This, however, requires two neutralizing coils and complicates the problem of placing them so that coupling to the output coil is avoided. Also the over-all gain increases to the point where instability due to mutual coupling is a serious problem.

Finally, nuvistors are no more immune to cross-modulation than other tubes used in series-cascode arrangements. Indeed, the high gain increases the problem and in any area where there are strong local signals, interference from cross-modulation is almost inevitable. The cascaded-cascode diagrammed in Fig. 4 minimizes both the neutralization and the cross-modulation problems, but the circuit is then as complex as it would be with two cascaded neutrodes or pentodes.

Actually, there is little point in trying to cascode nuvistors below the 2-meter band. A neutrode r.f. stage working into a good pentode mixer like the 6U8 or 6X8 produces a noise figure in the 3- to 4-db region through the FM band and this is enough to provide an IHFM sensitivity of 1.5 μv. or better.

There is more point to a cascode in the 2-meter band and higher and since the neutralizing coils can be much smaller, the problem of mutual coupling is not as severe. However, a nuvistor neutrode first r.f. stage into a 6AK5 pentode second r.f. stage gives as good a noise figure at 2 meters with higher gain and no more complication in circuitry.

Converters & Oscillators

An all-nuvistor front-end or converter can be built by using three 6CW4's, one as the r.f., another as a mixer, and the third as an oscillator. There is a tetrode nuvistor which can be used as a mixer for higher gain and, with screen injection, less oscillator pulling.

As a mixer the 6CW4 is connected exactly as it is when used as a grid-leak biased amplifier, with a 47,000-ohm grid-leak resistor and a 50- to 75-μμf. grid-leak capacitor. In the FM band and above, the length of the lead from the plate to the first i.f. transformer may be critical. A small inductance is helpful here to reduce input loading through the Miller effect. In the FM band, a lead length of about one inch is right.

The 6CW4 makes an excellent oscillator in any of the circuits useful in the v.h.f. range. The Colpitts arrangement is especially convenient. The capacitive voltage divider is provided by the grid-to-cathode tube capacitance and the cathode-to-ground capacitance augmented by a small capacitor across the cathode choke. The circuit of Fig. 5 works well through the 2-meter band. The author has had no personal experience with crystal-controlled nuvistor oscillators. It is generally agreed, however, that they present no problems that are not common to any other suitable triode. The circuit of Fig. 6 can be used with overtone-type crystals operating above 30 mc.

Matching & Loading

One thing that is not always taken into consideration when a nuvistor pre-amp is placed ahead of a receiver or tuner is the impedance of the load the receiver input will present to the nuvistor. For example, one of the most logical applications of the nuvistor is as a preamplifier ahead of inexpensive or medium-sensitivity FM tuners using a grounded-grid r.f. stage.

any persons have been surprised and puzzled to discover that the addition of a nuvistor preamp in such a situation has not increased the gain appreciably. Actually, in this instance the gain of the nuvistor cannot be expected to be much more than 2 and may well be unity. In effect, the nuvistor stage becomes the first section of a cascode. The very low impedance of the grounded-grid amplifier loads it so heavily that little gain is realized. The high impedance of the nuvistor should increase the gain of the grounded-grid stage but this is not always the case since quite often the grounded-grid stage has been designed more with an eye to reducing radiation of the oscillator than for providing gain. Thus the resulting increase in gain by the addition of a nuvistor may be very modest. Nevertheless, the noise figure should be improved to some degree and this should increase the effective sensitivity of the tuner.

In such a case, it is often possible to improve matters in two ways. First, a link can be added to the nuvistor output tank so that there is a step-down to match the low impedance of the tuner input. This will, of course, reduce the voltage fed into the tuner. However, this should be more than made up by the increased gain of the nuvistor itself.

In the tuner, the grounded-grid stage should be examined and, if necessary, modified slightly to provide the highest possible input impedance. Specifically, make sure the cathode tank is tuned to the operating frequency. When it is tuned, the impedance will be very much higher than when it is detuned. Usually this tank will tune considerably above the operating frequency when a preamplifier is connected in place of the antenna because the preamp will present a much lower capacitance than the antenna transmission line. A small trimmer can be added, if necessary, across the coil to bring it into resonance at the operating frequency.

Even when the receiver or tuner has a grounded-cathode input stage, matching the nuvistor preamp to the receiver input can have a significant effect on the nuvistor gain. Assuming, as is almost always the case, that the receiver has a 300- or 72-ohm antenna input, the nuvistor preamp should be link-coupled to the receiver or by tapping the output off the output tank so that the low receiver input impedance does not load the nuvistor too heavily. It may pay to try links of various turns and coupling to obtain the best match which will be indicated by the highest gain.

Actually, with the exception of grounding, the nuvistor presents no problems that do not occur with any "hot" tubes at the same frequencies. Indeed, once this particular eccentricity is taken care of, it is rather less critical than any other tube the author has used in the v.h.f. range. If the nuvistor is treated with the same respect as to circuit layout, it is usually easier to neutralize and will remain in neutralization over a wider frequency range. While it will not perform miracles, it will deliver sensitivities in the v.h.f. region which were once idle dreams.

 

 

Posted May 20, 2021