View Single Post
Old 12th Jun 2019, 6:36 am   #2
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Dormer & Wadsworth (D&W)

There were also the Larsholt FM front-end modules that came along a bit later.

The noise issue in the front end was indirect. Designing the selectivity between the antenna socket and the first active device was a compromise. Go for more resonators, or use light coupling to achieve narrowness and the insertion loss went up. This loss acted as an attenuator, reducing your signal, reducing equally noise from the antenna, and adding thermal noise from the resistive effects of the losses of the filter components.

A quick fag-packet calculation woulb be that the loss of the filter in dB would add to the noise figure of the active device following it. but ireality's a bit worse than that. The antenna will see cold sky in half its field of view, and outdoor temperature ground in the other half, so overall its noise output is colder than the 290K assumed in noise figure definition. (This is why it's still worth pursuing noise figures down to 1dB and below) The losses of the input filter are at room temperature and so have a worse effect on the incoming signal to noise ratio than the simple loss in dB would suggest.

In later FM tuner designs the Q of varactor diodes is dramatically lower than that of mechanical variable capacitors, and they become the dominant limitation on resonator Q, no longer the inductor. So the losses of the input filter increase leading to more noise disadvantage unless heavier coupling is used to widen it. A wider filter lets more off-channel signals in to add up and clobber the first active device. Worse, the varactors themselves are non-linear. Narrower filters run their resonators at higher Q's and that scales up the signal voltages across the diodes. Intermod products rise disproportionately fast.

So designers go quite wide with varactor tuned front ends. High-Q filtering would be desirable, but it bites you at both ends of the dynamic range. You need a better active device as the RF amp.... lower noise helps, as does greater linearity at higher levels. The better linearity means narrow preselection is less needed and the low noise figure makes up for a little of the loss of the preselector filter.

A diabolical juggling game.

I developed my own spreadsheet for visualising the tradeoffs and balancing acts of a multi-stage receiver structure many years ago and I've fine-tuned it as I've used it to design various receivers over those years. It did the Agilent Noise Figure Analyser family, so it contributes to the noise figure values in most data sheets, that other people plug into their receiver-designing calculations

When you get to multi-dimensional compromises, there aren't any unique optimal solutions and RF design does become an art.

Artful Bodger would have made a good pseudo-name on the forum, or has it already been used?

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now