Patrick Turner wrote:
Jon Noring wrote:

As I noted in a recent message, it is very intriguing to build a
modernized, high-performance AM tube tuner using the "channel"
approach. This takes advantage of the fact that licensed broadcasters
today must broadcast on specific frequencies, every 10 khz in North
America and 9 khz in Europe and elsewhere. So, instead of trying to
be able to continuously tune across the BCB spectrum, we can think
outside the box for the moment and consider the alternative of
building reasonably optimized tuning circuits for each listened-to
frequency. There'd be a switch to select from a number of channels,
each associated with a specific frequency the user wants to listen to
(suggesting a plugin mini-board for each channel, but there are other
possible configurations.)

The problem is that if you want a channel at 9 kHz intervals to
choose from across the band, you need around 12 perfectly set up
tuning circuits all with multiple LC circuits.
Then you need sitable switching. Far better is to forget all that BS
and use a PC to decode the antenna signal.

Well, if you recall, I did agree with you that the ultimate AM tuner
will be all PC-based DSP as close to the antenna feed as possible,
along with a true Class D digital amp for final output to the
speakers. Everything inbetween will be only real-time digital signal
processing. No need to sell me on that!

So why are we even bothering talking about tube-based equipment?
smile/

Because there is definitely an interest in tube-based equipment, for
various reasons: nostalgia, the challenge, the aesthetics, and in
some cases (such as high end audiophile amplifiers), The Sound (tm).

When true Class D amplifiers mature, they will supplant tube amps for
pure sonic quality. But that's still a few years off until PWM
switching improves.)

And even then, tube equipment is definitely of interest for
aesthetic and nostalgic reasons.

Regarding the channel TRF receiver being "BS", well that's in the
eye of the beholder. smile/


I infer from what Patrick said that it is unnecessary for a single
frequency AM tuner to be a super-het design, and that (I assume) a
much simpler two RF amp TRF design is sufficient for good to excellent
audio quality and good to excellent sensitivity and selectivity. (John
Byrns implies the same in his various comments on TRF AM tuners.)

But you won't sell many kits set up optimally for just one F.
As soon as the owner moves to another area, the radio becomes useless.

If the owner installed a number of "mini-boards" (or whatever) to
receive stations (both local and DX), then moves, he simply either
swaps mini-boards with new ones, or keeps the ones he has and adds new
ones, so now he has even more channels to "surf". The boards don't
become useless at all, especially if they're interested in casual DX.

The mini-boards can be sold either as kit boards (just add the
components of the right value, calibrate and plug-in) or buy them
already made and calibrated from the kit supplier. For simpler
bandpass tuning filters (not the complex ones like nine order
Chebychev, as an extreme example), the mini-board may only have a few
simple components to add. For example, for a given center frequency
(check the chart) just add a capacitor here of a certain value, a
resistor there of a certain value, an inductor over there of a certain
value, etc. Not a big deal. I envision the mini-boards to maybe be as
small as 1" x 2" in size, more like a stick, with terminals on the
narrow end to plug into a slot connected to the main circuitry of the
tuner (hopefully none of the components will be very large -- thus
the question I asked you about making the critically coupled RF
transformers common to all channels -- we don't want to have any more
than two or three of them!)


So, with respect to the channel approach, the next question to ask is
if we can use the same two critically coupled RF transformers (as
Patrick notes), and *independently* vary several of the other smaller
components (e.g., capacitors, resistors, and even inductors) in the
two or three tuning stages (if we include the antenna tuner) so as to
maintain, from channel to channel in the BCB, reasonably optimal
bandwidth and other desirable tuning characteristics?

This has al been investigated before, and the conclusions were about
as simple as possible by about 1927. Try studying basic L,C, & R
theory, and work all this out for yourself.

I have. This channel concept has nothing to do with "basics". It is a
twist to TRF tuner architecture taking advantage of the fact that AM
BCB is done in specific assigned frequencies, just like FM, like TV,
like the CB band, etc. It will not be practical for general shortwave
listening since that is a huge band (from 1.8 mhz to 30.0 mhz) and
amateurs in particular pick their own frequencies (and over time even
commercial SW broadcasts move around a lot, for those only interested
in listening to the majors like Radio Australia, as I do many evenings
on 15.515 mhz. It comes in loud and clear here in Salt Lake City.)

Back in the late 20's and early 30's, on MW there was clearly a need
for continuous tuning since broadcasts could be anywhere on the band.
(And tubes then had poor gain, among other problems.)

Today, a lot of the issues of building TRF circuitry is trying to
overcome the limitations of one-dimensional tuning using, for example,
a multigang air capacitor -- John Byrns is going through agony trying
to find the magic formula to get what he wants with a multigang air
capacitor. But with the channel TRF concept, the sky's the limit as to
how many components in the bandpass tuning filter can be independently
selected and hardwired for any given frequency. So one can optimally
tune the bandpass characteristics for each and every frequency in the
TRF without worrying how that affects other frequencies, since each
channel frequency tuning circuit is now effectively decoupled (made
independent) from the other channel frequencies.


[With traditional continuous tuning, achieved with multiganged air
capacitors, we do indeed vary a few capacitors in the tuning
circuitry, but because all of them track each other, in reality we
only have one degree of freedom, leading to circuit design constraints
for continuous "single knob" tuning. Now imagine, for each channel
frequency, to *independently* vary the value of several components at
the same time -- we now have several degrees of freedom to play with
and thereby hope to achieve reasonably constant (as a function of
frequency) bandpass characteristics.

1925 TRFs had 3 or 4 separate tuning gangs, each set to a certain
numbered position for reception of a given station. Finding stations
was exciting. Try studying the history of radio, and you won't need
to ask such questions here.

With the channel TRF concept, the component values of the bandpass
filter (or parts of the filter circuit) are hardwired on the channel
plug-in board (and trimmed during calibration), so all the person has
to do in listening to the tuner is switch to the channel, and the
radio will be in tune to the desired frequency, with the optimal
bandpass characteristics for that frequency. (There is likely to be a
need for a very fine tuning control, maybe +/- 1 khz, to handle slight
drift, both for tuner warmup, and for the inevitable long-term
drifting of component values.)

I suppose back in 1925 radio stations where in all sorts of weird
locations on the dial, and constantly moving around, so hardwiring
all the tuning components for a particular frequency, and likewise
for other frequencies, was not even an option.


For example, we can imagine having multiple plugin slots, where we
plug into each slot a PCB mini-board specific to a particular
frequency.

?

You probably understand the channel TRF concept, but did not
understand what I wrote the above, so let me restate with an example:

I want my channel TRF tuner to tune in 830 khz (WCCO in Minneapolis),
so I get the mini-board for that frequency already hardwired with the
optimum configuration of the various tuner components, plug it in, and
then listen to that frequency whenever I switch to whatever channel
slot I placed that mini-board in (I am reminded of how components are
plugged into PCs, such as via PCI slots.)

Or, I buy the blank mini-board, check the kit-supplied chart for 830
khz, and then solder in a 50 ohm resistor in this spot, a 200pF
capacitor in that spot, etc. -- probably will take me all of five
minutes. Then calibrate it by tweaking the trimmers. If I instead want
the mini-board to tune 1160 khz (KSL in Salt Lake City), I check the
chart, put in a 75 ohm resistor in this spot, a 150pF capacitor in
that spot, etc. (whatever values are called for.) Then calibrate it.
Plug it in, listen to 1160 khz, knowing that the TRF bandpass tuning
circuitry is now optimized for that frequency, and much better
optimized than could ever be done with the one dimensional limits of a
multigang tuning capacitor.

For those who build tube kits, this will border on the trivial.


And some hobbyists may find the channel TRF AM tube tuner architecture
of real interest, since now they can more easily experiment with new
higher-order bandpass filters of various mathematical functions to see
how they affect TRF performance. This could lead to a revised
mini-board to be issued at some future time based on all this
research,
and the channel tuner owner can, if they so choose, simply buy or
build updated boards for the broadcast stations of interest, and
instantly get better performance. It's possible to mix bandpass
filters for different stations: a third order Butterworth for 1160
khz, and a fifth order Chebychev for 830 khz. The possibilities are
endless.


The board will contain the few components whose values *independently*
change as a function of frequency. They probably will have trimmers
for fine calibration of the center frequency and other bandpass filter
characteristics. We may need multiple mini-boards for each channel
(one for each tuning stage) if necessary for shielding purposes (to
prevent oscillation by stage-to-stage interference if that is a
problem.) And if higher frequency channel boards require some minor
changes in the circuitry configuration, and not just component value
changes, that can easily be done, too. In principle, this tuner might
even be able to extend a little beyond (on both sides) the 500-1800
khz MW band -- just plugin the right mini-board circuitry for the
frequency desired.

This idea is totally impractical for 120 different stations,
and plug ins get lost or broken, or worn out.

I don't believe it is impractical for 120 different stations, for two
reasons:

1) Those tube-o-philes who only want to listen to stronger local
stations, or to particular distant ones, are likely only to want to
have 10-20 stations (with the ability to add more if they want.)

One purpose of picking TRF is its legendary high-fidelity audio
capability which will appeal to audiophiles -- most won't want to
listen to a very weak station 1000 miles away that can only be
picked up some evenings.

And I believe it is easier to sell tube-o-philes on the Channel TRF
concept once it is explained how it maximizes audio performance for
each and every broadcast frequency that cannot be done with a
continuously tuned TRF.

2) Those who would use this for casual DXing (and note the hardcore
MW DXers will use something like a Drake R8B or ICOM R75, or some
digital receiver) will certainly be motivated to add more
mini-boards, and can do so over time. The tuner will work with 1
channel board, or with all 130+ (if enough slots are provided. For
the moment I am imagining the mini-board approach, but the sky's
the limit for other ideas to implement the channel TRF AM tube
tuner.)

3) And as noted above, hobbyists may find the "plugin" bandpass filter
capability of particular interest.


Of course, others here will probably have much better ideas as to
how to implement the channel approach.

You bet there are, and only possible with chip technology,
with press button station selection, and digital station F read out,
with digitally generated oscillator frequency for the F converter of
a superhet, with ceramic filter IF. Grundig have been multiband
radios for about 20 years +.

I have a Radio Shack DX-399 (the Sangean 606A) which is a very good
performer for casual MW (with the Radio Shack MW loop) and shortwave
DXing. So I am very familiar with that hobby, and with the benefits
digital systems bring to tuners. You need not sell me on that!

See my previous note above on "why tubes then?"


Not a tube in sight inh these lightweight plastic radios bought
cheaply by the masses to allow connection to the world's AM, FM, and
HF bands, and even amateur SSB stations.

See my previous note above on "why tubes then?"


(It's interesting to think of doing the same "channel" approach for
an FM tube tuner. Will that also confer several advantages in
simplifying the circuit design for the same overall performance
level?)

Study the way most post 1980 AM/FM tuners are constructed.
Tubes cannot be used with such methods.

O.k. But are you referring to tube-based tuners? Again, if all I
wanted was an audiophile grade AM/FM tuner, and did not care about
what was under the hood, I'd be open to solid state designs, but I'm
specifically looking at tube-based tuners.

I still assume that the channel approach to tube-based FM tuner design
may confer some benefits, but maybe less since the frequency ratio to
tune from the lower to the upper ends of the band (about 1.25) is much
less than that for the AM BCB (a whopping 3.5 or so.) And there are
probably other factors as well specific to frequency modulation.


It's trying to tune the AM band with only one degree of freedom (e.g.,
air tuning capacitor) which is causing all the hassle in tube-based
TRF AM tuner circuit design. One would want to pick a bandpass filter
which is optimally tuned to the specific frequency we want to listen
to, and this involves optimally selecting *several* component values,
not just one as we are limited to by continuous tuning with a
multigang tuning capacitor. The channel TRF approach appears to free
up the TRF designer from the tyranny of having to compromise the
bandpass characteristics over the entire tuning range which only one
degree of freedom allows.

Of course, superheterodyne is one solution to the TRF problem, and
allows for continuous tuning. Note that super-het works because it
uses "one channel" (the IF). So in a sense, superheterodyne supports
the channel TRF approach for those who don't want to build a
super-het, but rather want a pure TRF receiver (e.g., for sound
quality reasons, or whatever.)

*****

Now, I've made the call several times for classic and proven AM tube
tuner designs of the past which have excellent audio quality (and wide
bandwidth capability), are good for casual DX use, and can easily be
"modernized" for a kit. There are no doubt many excellent super-het
designs out there, but I've had very few recommendations. Patrick,
since you appear to much prefer super-het over TRF for AM tube
tuners, which classic super-het tube AM radio designs of the past
would you suggest as candidates to consider?

Anybody?


I reckon you got a pile of reading to do.

Yes, I have been reading. That one-year equivalent of EE training back
in 1974 at the University of Minnesota is slowly coming back to me.
Back then we spent a few weeks on tubes, and only a couple days with
transistors. Things have changed a lot since then.

And it was interesting reading about Chebychev bandpass filters today
since I wrote a lot of Fortran code years ago to do various types of
numerical analytic processing including integration using quadrature
with orthogonal polynomials (mostly Legendre polynomials.) It was
especially cool to see how the higher order Chebychev polynomials U(x)
plot out in the desired shape (well approximately) for a bandpass
filter (but with that slight ripple within the bandwidth.) I'm not
saying all this to brag, but to give a better idea of my background.

Definitely I have a lot to learn, of course, and your posts are
helping me to better understand things. I still believe the channel
TRF concept is viable for those who want to build the best possible
TRF tube tuners where for each frequency the absolute best bandpass
characteristics can be chosen without worrying about how it impacts
the other frequencies since each channel is now largely independent.

The obvious downsides with the tube-based channel TRF concept a

1) The practical, real-world implementation of it (I believe it is
doable, I suggest one approach),

2) Losing the ability to continuously tune, which for BCB is not an
issue as I've noted several times, and

3) Calibration of each mini-board if done by the kit-builder (I think
this is solvable, but it is an issue to consider.)

The upsides are several, as previously noted.


Thanks for your helpful comments!

Jon Noring