The Screen-grid Valve as RF Amplifier

Radio Record, Volume II, Issue 44, 17 May 1929, Page 28

The Screen-grid Valve as RF Amplifier

Varied Experiences of Constructors

HAT great interest has been shown in the screen-grid valve as an HR.E. amplifier has been evidenced by the very large number of correspondents who, during the . last few weeks have written in on this score. As stated previously, we have held these over until ‘now, they are published coliectively with some additional notes. The fate of the screen-grid valve still hangs in the balance. It was at first greeted with great enthusiasm, but now there is a.reactive tendency and it is being considered more critically than formerly. The great difficulty with this valve is its flat tuning, and when devices are introduced to sharpen. this, losses occur which render it little bettem than the average valve. It appears that a screen grid set would be most: suitable for use in the coutnry distant from a local station. Even then, tuning would be flat, and there would be great difficulty in separatjing stations whose wavelengths approximate one another. A short time ago "Pentode" discussed the application of the screen grid, recommending transformer coupling.

This has generally proved successful, as the correspondents indicate. The screen-grid valve has been reviewed by one of New Zealand’s radio authorities, Mr. E. R. McCarthy, who states: The science of radio frequency amplification has of late seen retrogression rather than progress. Nothing of any-real value has been introduced since the Haseltine. neutrodyne system several years back, and even that, in so far as the principle of neutralisation by external. capacitative reactance is concerned, was only a Variant of the earlier Rice system. A surprising number of manufacturers have returned to the oldest method of stabilising known-potentio-meter grid control. That there is still a-call for sets which will bring in distant stations was distinctly evidenced by the sudden rush when the screen-grid valve was announced, and it is decidedly unfortunate that this was heralded by somuch misleading. publicity, which Ted builders to expect much that has proven impossible.

Many of the leading publications carried editorial matter describing the great amplification obtainable from a valve which would not oscillate, and even ‘the largest valve manufacturers and ..best engineers in the country allowed ‘statements to appear such as "A voltage amplification -of 200 per stage is obtainable, but at broadcast programmes the resonant impedance is lower, teducing the amplification by 25. per cent. of this value." Prior to ‘commencing any actual work, certain limits were laid down as essential if the results were to be ‘of any value in designing a receiver which could be constructed by ‘the home builder. First, standard apparatus, obtainable by anyone, must be employed. Second, if possible, the need for any complicated balancing by means of expensive accurate apparatus after construction, must be avoided. |'Thirdly, the use of shielding undesirable as introducing superfluous expenses and trouble, Last, no involved or critical adjustments of any kind were allowable, as such a receiver must be infallibly sure to give good results, if the connections are nroperly made, in the hands of the most .inexperienced operator. Types of Coils. TEST was given of coils of sixteen types, aS many of each type being tried as were deemed necessary to determine their worth. Inasmuch as the reproduction of all these curves here would lead to confusion, due to -their numbers, and would serve no. particularly useful purpose, we = will show the results obtained in these most useful types, which -had a bearing on the final result, but as a matter of information, it may perhaps be advisable to outline roughly the types involved and the major reasons for their abandonment. All coils were of the general type. having 77 turns of wire in the secondary circuit, and spaced to conform to an approach to the ideal shape factor, and supported by skeleton bakelite frame, so that the insulation losses are kept at a minimum figure. The actually tested coils as the intro duction of a primary coil or the use of a portion of a secondary coil for coupling, have a decided effect. on both the inductance of the secondary. 1. A tuned impedance, directly from plate to grid to ground. 2. An auto transformer, in which a portion of the secondary is used as primary, and low potential end being common. 8. A transformer in which the primary and secondary are coupled by : by-pass condenser at the low potential ends, the direction of the winding be-

_ ing continuous from plate to grid, and the coils being tuned from plate to grid as in the original "R. B. Lab." circuit and other Betts circuit adaptations, 4, A primary wound to take up a length of one and a-half inches inside the secondary. . 5. A primary wound to take up onehalf inch placed inside, and ‘in the centre of the secondary. 6. A primary wound to take up onehalf inch, placed inside, and opposite the loss potential end of the secondary. 7% A primary wound with a length of one-eighth inch, placed in both positions above described. a 8. A primary wound in a 1,32 inch slot, coupled adjustable to the secondary. 9. A primary wound on the same diameter as, and at an adjustable distance from, the secondary. . 10. A tuned primary with adjustable coupling of the secondary. Types four and six are yery commonly used, and hence we have selected them for detailed presentation in connection with two types. Type four is generally advocated for use with the shield grid valve when the tuned impedance arrangement is not employed. After the number of turns ‘inerease to a certain point,-no further increase in amplification is obtained. Type six is the type most commonly used at the present time, and the results indicate that if conventional cir-\ cuits are to be employed, it is considerably superior to any of the others tried. The results indicate at least as good result as with the widespread primary in amplification, with a considerably improved selectivity factor. Type two is extremely interesting. In this type of coil a varying portion of the secondary acts as ‘primary, with the primary and secondary currents in this portion of the coil in quadrature. The inductive coupling is much greater in this type of coil for the same number of turns, hence a comparatively high primary impedance is built up with a relatively small number of turns with consequent greater overall amplification. . Shifting the Phase Angle. A REVIEW of various past methods of stabilisation indicated ‘that the best results were probably obtained ky shifting the phase -ngle of the cyrrent in various stages by the method ‘sonrewhat similar to that used in the circuit described in the "Listeners’ Guide." In a circuit containing two stages of radio-frequency amplification it will rarely be found necessary to employ more than one: such resistance, as suflicient adjustment of the tune relation ‘can be obtained ‘to avoid oscillation, while still maintaining the satisfactory value of overall. amplification, Maximum results, regardless of the’ type

of yvaive used, can be obtained by varying the proportion of the secondary eoil, which is used in the plate circuit of the preceding valve. Shielding always introduces certain losses in the coil, due to linking a portion of the coil’s field, and also complicatos the mechanical construction of the receiver. ; The necessity for shielding is mitigated as far as interstage coupling effects go, while the diameter of the coils is so small that direct pickup from local station is reduced to a minimum as has been demonstrated in practice. Correspondents Write. [s response to a general invitation in the "Radio Record" re screen grid (writes F.W.L., Alexandra), I can advise that I have been using one for three months The set was originally a five-valve Browning Drake, home as\sembled, with commercial coils, 1-V --3, the three audio stages being impedance coupled. : The set has gradually transformed into a 2-V-1i; 1 screened grid, tuned anode, 1 Browning Drake stage, space wound coils on celluloid strips, phasatrol neutralised, detector, and one audio Stage, with a pentode. The two radio valves have byepass condensers and decoupling resistance in the B leads; the detector and pentode screen are byepassed, and two 100 Henry chokes out of impedaformers are used as decouplers. The two radio stages and the detector are fully screened in upper. The set is used without reaction, I took the tickler right out and when I was using a good choke it was a pleasure to listen to. It isn’t bad now, but won’t carry the same volume ‘without distortion. I found the screen grid a big improvement, but the tuned . anode has broadened the tuning, and I can no longer separate 2BL and JOAK. I sometimes use anode bend detection on 2YA, when it is coming in well (very variable here), and notice the sharpness of tuning of the detector stage as compared to leaky grid. The set is extremely quiet, and cannot be made to oscillate, and I am considering replacing the Browning Drake stage with another screened grid, but the selectivity. question deters. I’m afraid I weet be able to. separate 2YA and HAVE wired my set to your diagram No. 1, "Application of the Screen Grid" in the "Radio: Record," Vol. II, No. 36, writes H. Jacob, Mangaweka. I use the best components procurable. I rewired the set he Waster (1-V-38), and since then I ihavé had the following stations: New Zealand’s 1YA, 2YA, 3YA, 4YA, 2ZM, 2Z¥, and 3Z0; Australia: 2BL, 2FC, 2UB, 2GB, 8AR, 8LO, 4QG, and 7ZL; and three other stations I caunot get the calls-eighteen in all. ‘The first 14 were all too loud when tuned right up to the mark. These stations could be heard about half-mile away. You can see I’m getting all the volume I require. I have shielded the S.G. stage and the Det. stage, and I am now getting a very good tone. It'is very seldom I'use the four valves, as I find that I lose tone if I come down to that. / I think Mangaweka is rather a poor . reception place, as a friend of mine ‘Shas a five-valve set, about nine miles from here, and he ean pick up America and Japan fairly often. There are eight sets here, and not one of us has picked up America or Japan.

I HAVE taken advantage lately of your Queries and Answers Column in connection with the application of the screen grid to the Brown ing Drake. I know that the term "Browning Drake" is somewhat a misfit, but I call it such for the want of a better name. I made "Pentodes" 1 to 1 transformer, as described in an earlier issue of the "Radio Record," and used a 442. I found that with 140 volts on the pla.. and 674 on the sereen that the results were entirely satisfactory. I tried 80 volts on the sereen, but it seemed to choke the valve, and after a little experimenting 673 seemed to be the best. The second dial also tuned broad. [I might state that I joined the grid leak return to A minus, but have altered it to A plus. Would that make any difference to the broad tuning? (It may improve matters depending on conditions particular to the set.) I tried a .0001 fixed condenser in the aerial, but it made the tuning the reverse-too sharp, in faci. (Try bigger valve 00025.) I may also state that the B battery, which I made to "Megohm’s" specifications, is still going strong. The only thing that is worrying me is the broad tuning. I have shielded the’ R.F, stage only, using 615 as detector with 5 megohm leak and usual audio stages.-‘Dials" (Napier). I HAVE been very interested in the discussions for and against transformer coupling and I am sure discussions of this nature are of interest. By closely coupling the primary and secondary of the transformer and tuning the secondary with a parallel condenser, it is stated that indirectly the primary is tuned. Now, as an example, if both consist of 60 turns and the secondary has a parallel condenser to increase its wavelength, how can the primary be tuned any way near the desired frequency without a similar capacity across it or increased turns? I notice this week that an American, a Mr. Marvin H. Thoreau, in a letter on page 8 states that the "best sellers" ave kit sets which tune both grid and plate circuits of 8.G. Perhaps it would be worth the extra to tune both, as it is very desirable using 8.G. valves to gain as much selectivity as possible. About six years ago I used an aperodic transformer (both primary = and secondary) and found it fairly good. I believe it was wound with resistance wire to make it aperodic and I have an idea that perhaps, if in the coupling you suggest the primary was wound with fine resistance wire, it would be aperodic to the tuned secondary. When Browning designed the B.D. Regenaformer, he was careful to place the primary as near as possible to the grid end of the secondary coil, and apply reaction to the filament end, presumably to avoid damping. Now I suggest winding 60 turns fine resistance wire solenoid fashion on a celluloid former to fit closely (say, 1-16th inch) inside a low-loss secondary of 60 turns No. 24 8.W.G. As the secondary would be the longer, the primary could fit under the grid end and reaction could be coupled to the filament end. I am contemplating using a potentiometer in the grid return of the 8.G. valve, so as to get any degree of grid bias. Do you think this worth while? Finally, can I expect a &.G. set of the type diseussed to be as selective az the

ood old Browning-Drake?-Mech." Dannevirke), : Comments. The circuit referred to, tuning both plate and grid, is an altogether different circuit. The primary is tuned, in the close-wound primary and secondary, in centre of its light magnetic coupling with the secondary.

The correspondent is under a misapprehension regarding the placing of the primary in the B.D. It is placed under the filament end of the secondary. This would not act with the screen-grid valve. The potentiometer to control grid bias should be OK. For selectivity, the four-valve . B.D. is hard to excel.