Experiments with the S.G.

Radio Record, Volume IV, Issue 5, 15 August 1930, Page 24

Experiments with the S.G.

A Reader’s Experience

ROM experiment I have found that the screen grid valve is the most efficient valve for the radio frequency stages of the Browning Drake. The two advantages I find it posSesses are: (a) greatest. amplification ; (b) no necessity for neutralisation. I gather that the great objection -to, the screen. grid valve is its high im-. pedance-usually about 200,000 ohms, whereas the normal impedance of a slot wound Browning Drake primary is about 15,000 ohms. At such. a disrepanecy as this the screen grid valve would undoubtedly be unsitisfactory. The solution of the problem is not, however to be found in discarding the yalve, but in adjusting the primary. For experimental purposes I had a rexeneraformer wound with a_ split primary, having an average impedance of. approximately 50,000 ohms. I then tried ‘various valves in front of it, viz., 609, 201A, 630, 685, 200a, PM6D, and screen grid. . The screen grid proved to be the best, being fully 20 per cent. better than the 635, which was the best of the rest. The screen grid was, of course, connected across the whole primary. I used no shielding at this stage, and there did not appear to be any trouble. Selectivity was quite as good as with any of the other valves, and better than with the 609 and 630. t next decided to add a further stuze of radio frequency, and for this prpose I had a radio frequency transformer wound with split primary as before. This radio frequency transformer I shielded with a copper can. Again I found the best valve to be the screen grid. Selectivity with the two stages is just rather too sharp for long-distance work-in fact, the set is quite as sharp as a short-wave set. As a result of the experiments I am satisfied that the screen grid valve can be made eminently suitable for the Browning Drake circuit, provided the primaries are wound to a maximum impedance, It is not to be expected that the full amplification factor of the valve will be obtained, but it is a simple matter to obtain a factor considerably in excess of that of any three electrode valve. :

I propose to experiment with a slo. wound primary, having the Same number of turns as the secondary. Another matter of interest is that I am still using the standard grid leak to filament negative, and it is going so well that I have not bothered to try a positive return. I am also trying the result of applying reaction to the screen grid valve by putting a condenser between, place and grid. In‘dications are to the effect that such an arrangement will be advantageous. but I have not yet completed the experiments. B=XaFoRE continuing some further screen grid experiments, Iam want ing some more information on imped ance, and perhaps your technical edit’ could give me some advice on the fo) lowing points :- (a) If a primary coil at a certaiv frequency has an impedance of 50,000 ohms what would be the effect on im pedance of applying parallel feed to this primary? Condenser .5 and effi cient choke. ° (d) Will the application of reaction either increase the impedance in the plate circuit or decrease the internal impedance of the valve. (c) Why is the grid leak recom mended to be connected for filamem positive in the screen grid circuit. ~(d) What is the maximum numbet of turns possible in a slot wound prim ary? Is there a maximum? What would be the effect or increasing the primary turns beyond the number of the secondary and why should it have that effect?-‘Diogenes" (Cromwell). T will be noted that when a large number of turns are used as a primary the amplification on the

higher frequencies is reduced for the portion of the coil in which currents are in quadraturé is appreciable, and hence the voltage at the grid of the following valve is reduced. After a given number of turns there is no increase in the amplification. Past this further selectivity is impaired. It is generally conceded that this number should not exceed 20 on a three-inch coil, Re the application of reaction to the s.g. valve through a condenser between the plate and the grid. The same pur-

pose would be served if a 0-100,000 ohm resistance is placed in the screen cireuit. This arrangement would probably be better. Concerning the queries when a circuit consists of two or more branches in parallel, as in this case, the total amount of energy in the circuit cannot be obtained by caléulating the branch currents and adding them because of the difference in phase of the various branch currents. ‘The situation is further involved because of the possibility of resolving this fluctuating current into a smooth direct current, modulated by a variable frequency alternating potential. Unfortunately, however, when resistances of the proper value to give us adequate control of the phase angle are -employed, it will be found that any variation of the resistance varies the place voltage applied to the valve, and may seriously effect the amplifaction obtainable. In order to avoid this a comparatively low inductance, (R.F.C.) should be shunted, across the resistor, thereby ensuring a maximum static value of plate voltage at the valve while offering a very high impedance of radio frequency currents. This resistance can now be varied without af--fecting the static value of the plate * voltage, and will be found to serve

very nicely as an oscillation control. Also, if provision is made so that it can be reduced to a low value so as to effectively short-circuit the .5 .mfd., it can also be employed as a volume control for the receiver. (b) The impedance=change in plate voltage divided by the corresponddng change in plate current. (c) Being without gas, the filament return of the screen grid valve is to positive. (d) This has been covered by our introductory remarks.-Tech, Hd.