Quote:
Originally Posted by Synchrodyne
The attached RCA paper provides a simple “worked example” of the benefits of extra front-end selectivity for FM receivers.
That showed that in the case of bipolar transistors, it was better to place the bandpass at the input and sacrifice some noise performance in favour of better spurious response rejection. When dual-gate mosfets arrived, it was more common to place the bandpass at the interstage, with a single-stage at the input. Presumably the better inherent linearity of the mosfets allowed this. Although there were some mosfet examples with a bandpass input, such as the D&W four-gang front ends ( https://vintage-radio.net/forum/show....php?p=1153245).
Cheers, |
The thread you've linked was an interesting read.
In my young days I experimented with prototyping FM front end being capable of DX reception in a "hostile" environment, which was in my case: proximity of a local FM transmitter, FM relay towers, electrical tractions of train station (at ~150m distance), and sparks from tram line and ignition electric of cars passing the street in front of my windows.
Running through various topologies and technologies I dropped BJTs soon, as these were incapable of suppressing these annoying spark interference, even with elaborated filtering (which tended to worsen the noise performance).
The only JFETs I had in hand those days (Japanese 3SK41) performed somewhat better in this respect, but I felt, they had too small transconductance and/or too high NF to achieve the ultimate sensitivity I was after.
The real kick in performance was the use of RCA dual gate MOSFETs I got some day. The circuit was:
single tuned input -> RCA 40672(*) gain stage -> double tuned bandpass -> RCA 40673 mixer with LO injection to gate 2. The LO used a Japanese 2SC668 BJT.
Tuned with use of internally back-to-back connected varicap diodes which were a speciality of those days. These were varicap sets made extra for FM tuning, one type (case dotted green) for RF tanks, the other (case dotted blue) for the LO tank, both types made to have their Ct = f(Ut) characteristics matched to assure best possible tracking along the tuning range. The types I used were Japanese (of Type I don't remember, labels starting with ESM..., if I recall it correctly). Some European companies at that time came with similar device sets (BB204 and BB205, if I'm not in error).
Unfortunately these pearls had gone long ago, sacrificed to the "low cost" philosophy of modern world. (In this discussion therefore, it is worth to note, that today one has to live with an inferior tracking.)
This front end, despite its simplicity, worked exceptionally well. It boosted remarkably the sensitivity as compared to previous prototypes, suppressing the spark interference quite a lot. Good tracking RF vs. LO was possibly essential to achieve that performance.
(") The 40672 were then the types RCA recommended for use in RF stages, and 40673 for mixer. Especially for this project I've bought two sets of 40672 and 40673 from Radio Rim in Munich in 1979.
Interestingly, in the RCA book "RCA Solis State, Linear Integrated Circuits" from 1978 I acquired years later only the 40673 type is included. No spur of 40672 to find today, anywhere.
PS: Of the RCA paper you linked the pages 28-32 seem to be completely blurred. I found a clean copy here:
https://ur.booksc.me/book/14578305/f04202
Adam