[Radio Wrangler]

Indeed it is. The requirement for selectivity in the rF path is greatly eased, but not eliminated.

Considering a classic entertainment LW/MW/SW banded radio; on longwave the RF stage selectivity needs to be kept wide enough that it doesn't limit the channel bandwidth more than the IF does, or you get a peaky response in the audio. On this band the RF frequency is less than half that of the IF, so it is easy to get it narrow. And with two tuned circuits which are narrow, alignment and tracking difficulty is exaggerated

On medium wave, you need enough selectivity to block signals at the image frequency, usually around 910kHz higher. This isn't too onerous.

On shortwave, that 910kHz image is a lot closer, in ratiometric terms, and the filter narrowness available for a given Q of tuned circuit has got a lot wider. Rejection of the image has become poor, and gets worse up the frequency range. The classic shortwave sets like the AR88, HRO and CR100 were well mad and designed to do about the best feasible at the time, and they barely got 3dB attenuation of their image frequency at the top of their tuning range.

Tuned stages can be done with either permeability or capacitor tuning and similar bandwidth results obtained, if designed appropriately. Confusion is created because it isn't simply the Q of the naked esonator which counts, it is the in-circuit Q. This is carefully planned in the design of filtering, and the Q is managed by controlling the coupling factors into the impedances of preceding and subsequent stages. Capacitor coupling is popular through cheapness, but it gives tighter coupling at high frequencies and this gives lower Qs. A filter maintaining constant Q would get wider proportionately to frequency (simple scaling) But a filter with 'top capacitor' coupling couples stronger at higher frequency and gets even wider than scaling would predict as frequency goes up.

A filter containing 'Top Inductor' coupling reduces coupling factors ar higher frequencies and tries to combat the widening through scaling effect. These can be ade to have a much more uniform bandwidth across a tuning range.

Other coupling methods can be used such as taps in inductors, or tapped capacitance which are closer to simple scaled variation in bandwidth.

The choice of coupling also determines whether a filter rolls off faster on the top side or the bottom side. So even if you're doing fixed-tuned coupled-resonator filters , you want to be choosy about what sort of coupling yo wish to use. Would you like symmetrical skirts, or would you like it to fall faster on one side to dodge a problem frequency?... even if you want symmetrical skirts, do you want them symmetrical on a linear frequency plot or on a logarithmic one?

There's a whole menu to choose from, but it isn't explained in its entirity anywhere very visible. Most computer programs for generating filters only do the very basics. There's a lot more buried in 'Zverev' but that isn't easy reading.

So the way permeability tuners are implemented often is good for more constant bandwidth. But it isn't implicit in permeability tuning and it can be done with tuning capacitors too, if you're not penny-pinching.

An advantage to permeability tuning that is rarely mentioned is the avoidance of the sliding contacts on tuning capacitor rotors.

David