The Author'S Reply To The Above Discussion: - /-F) Will Not Affect The Modulator Loss If The Input Is

220
DISCUSSION ON THE TWO PAPERS BY DR. TUCKER to swamp the effects of the filter impedance outside the pass band, a resistance pad may be included either between modulator and filter or between input transformer and modulator. Where rectifier capacitance can be ignored the second alternative may be preferred since pads are usually included in any case in front of the input transformer to drop the signal level to a value which can be handled by the modulator. By placing the pads between input transformer and modulator the requirements for the output impedance are relaxed and the pad at the low-level point between modulator and filter may be reduced or omitted.
(or high) at (fc + / ) . Moreover, when the impedance terminating the input matches the characteristic impedance of the modulator at all frequencies, the loss through the modulator will depend only upon the impedance terminating the output at the required frequency (/ c / ) . In other words, the output impedance at (fc -\-f) will not affect the modulator loss if the input is terminated by a resistance. Similarly, when the output is terminated by a resistance the input impedance is important only at the input frequency /. A further point follows: to present a good impedance to a filter following a modulator and
THE AUTHOR'S REPLY TO THE ABOVE DISCUSSION

Dr. D. G. Tucker {in reply): For the most part the discussion has consisted of further contributions to our knowledge of modulator performance, and there have been few criticisms to which I can reply. The keynote of the discussion seems to be that our knowledge of modulator performance and design is still very incomplete, and this does appear to be largely due to the difficult non-linear elements from which modulators are made. The basic principle of operation of a modulator of switching type is independent of the exact characteristics of the rectifiers or valves used, but unfortunately all the finer points of performance, which are becoming so important in modern communication (and other) techniques, depend entirely on these characteristics, which vary greatly from one type of element to another, and worse, from one individual element to another of the same type, and from time to time according to temperature, power supplies or age. Thus Mr. Cooper's stabilized modulator, using a constantcurrent feed to the rectifiers, is a significant development, although its complication may detract from its range of application. I also agree fully with Mr. Peachey that the modulator circuits described in the paper are often not the best, and, like him, I have been developing basically different types for some precision applications. The use of negative feedback in modulator circuits is worth a "good deal of study, and one application is described in Post Office Research Report No. 13041 (July, 1948), "Modulators and Amplifiers Stabilized by the Use of Forward and Feedback Paths of Similar Variability." Mr. M. J. Tucker's extension of the range of application of ordinary ring modulators by the use of very large bias voltages is also an important development; I had myself noticed that even bias voltages of around 0-5 volt decreased the amplitude of the unwanted non-linear terms {fc nf) very considerably. I agree with Mr. Ayers in criticizing the use of modulators also as limiters; such extension of their use reduces their precision, and we should note Mr. Halsey's warning that the tolerance with which several speakers have regarded the modulators in a communication system is rapidly becoming dangerous. 1 have no experience of the modulator over which Mr. Jacobsen is so enthusiastic, but he does seem to think it can be used with precision in such systems. Regarding some of the detailed points raised, many of these refer to the rectifiers themselves. Mr. Crutch and Mr. Halsey refer to the capacitance of copper-oxide rectifiers; this is certainly a non-linear quantity, varying with the bias voltage and with frequency; but the resistance also varies with frequency. Something similar appears to apply to germanium rectifiers, but whereas the effects are obvious with copper oxide at frequencies in the lower carrier range (12-60 kc/s), they are probably negligible with germanium up to 1 Mc/s. Both Mr. Halsey and Mr. Jacobsen mention the effect of ideal rectifiers (i.e. with zero forward resistance) on the modulator operation; actually, although the ring modulator would not work in such a case, the Cowan would work; but the matter is not serious as such rectifiers do not yet exist. I am surprised that Mr. Willisfindsgermanium rectifiers are more uniform than copper-oxide; that is not my experience. I agree with Mr. Ayers that I have been concerned with getting the best total balance against carrier leak; if one is interested only in the fundamental component, a perfect balance is theoretically obtainable with a single potentiometer. But, in practice, reactance unbalances preclude such a result, and I imagine that a carrier balance of 80-90 db is very exceptional, even when some effort is made to correct the reactance unbalances. Mr. Stanesby and Mr. Harris raise a most interesting point in regard to the effect of the terminating impedances of the modulator. I agree with them. An analysis of the problem (with certain restrictions to make the analysis practicable), given in a forthcoming Post Office Research Report,* would probably interest them. Mr. Harris's point about the connection of the usual attenuator pad either before or after the modulator seems an important one, hitherto overlooked.
* No. 13115, "Rectifier Modulators with Frequency-Selective Terminations."
"WATER-COOLING VERSUS AIR-COOLING FOR HIGH-POWER VALVES" RADIO SECTION DISCUSSION MEETING, 15TH FEBRUARY, 1949
In his opening remarks, Mr. J. Bell began by saying that if ordinary tap-water had been suitable for the cooling of highpower valves, it was unlikely that large air-cooled valves would have been developed. But, in general, tap-water was unsuitable for use as the cooling liquid, owing to the resultant "furring" and to its high electrical conductivity. Hence, except in rare cases where sufficiently pure natural water was available, distilled water had to be provided. Such cooling water had to be stored in large tanks, circulated to the valve anodes through suitable piping, cooled by means of a water-to-water or water-to-air heat exchanger, and finally returned to the reservoir. The cooling system, therefore, occupied a considerable amount of space and was a fairly costly installation. It might be asked why tapwater should not be used in spite of its disadvantages, and the

The Author'S Reply To The Above Discussion: - /-F) Will Not Affect The Modulator Loss If The Input Is

Uploaded by

Copyright:

Available Formats

The Author'S Reply To The Above Discussion: - /-F) Will Not Affect The Modulator Loss If The Input Is

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Author'S Reply To The Above Discussion: - /-F) Will Not Affect The Modulator Loss If The Input Is

Uploaded by

Copyright:

Available Formats

220

THE AUTHOR'S REPLY TO THE ABOVE DISCUSSION

You might also like