Last night I posted a couple messages regarding the proposed "channel
TRF" AM tube tuner, focusing on the plug-in mini-board idea as one way
among several possibilities to implement it.

The idea underlying the channel TRF concept is to build bandpass
tuning circuitry specific to, and optimized for, each frequency in the
BCB, instead of fixing that circuitry to some "average" value and
trying to vary it using a traditional variable air capacitor (or
variable inductor) for continuous tuning. A switch would be used to
select the bandpass circuitry for the particular frequency channel the
listener wants to hear.

This would allow, in principle if not in practice, the ability to
very precisely optimize the bandpass circuitry (to maintain a quite
constant bandwidth and shape) for every broadcast frequency in the BCB
(from 500 khz to 1800 khz.)

The "mini-board" variation of the concept would place the bandpass
circuitry for each channel (frequency) onto a small plug-in PCB board.
Depending upon the type and order of bandpass filter used, the number
of components on the mini-board may be quite small, maybe a couple
capacitors, a resistor or two, an inductor, etc., having the optimal
values, and with one or more trimmers for fine adjustment of the
center frequency.

Clearly there are several implementations of the general concept, one
of which is a well-known hybrid that allows continuous tuning in the
more traditional and familiar way. The ones I think of at the moment
a

1) Traditional continuous tuning: Divide the wide BCB into several
sub-bands, such as 5 or even more, each sub-band having optimized
bandpass circuitry for the sub-band, and then use the traditional
variable capacitor or inductor to tune within the narrow sub-band.
Although each channel will no longer have the most optimal bandpass
configuration, it will be closer to optimal.

2) Single Board, True Channel: It may be possible, instead of having
120+ totally independent channel circuits each placed on a separate
mini-board, to put them all onto one larger board, but still keep
all circuits otherwise separate on the board. A lot of components,
and probably a lot of trimmers.

3) Single Board, Shared Components: As a combination of items (1) and
(2), channels which are adjacent to each other (in their own
"sub-band") could probably share a lot of common bandpass
components, thereby reducing the number needed on the board. Only
the large number of trimmers for individual channel calibration
will remain.

The original idea of mini-boards is most advantageous when the user
of the TRF tube tuner only plans to listen to 10-20 stations (such
as local, higher-power stations). They only install the channel
mini-boards they want to listen to.


*****

I do have a couple questions of both John and Patrick (and anyone else
caring to chime in) related to this.

1) In the single frequency TRF tube receiver (a TRF designed strictly
to listen to a single frequency), is there a need for double tuned
circuits? Or will singly tuned circuits be sufficient for
excellent performance (audio quality, sensitivity and
selectivity)? If not, how do double tuned circuits benefit the
overall performance of the single frequency TRF receiver?

2) Let's assume that we decide to design a Mark I TRF AM tube tuner
kit designed solely for more local, higher power stations (thus the
sensitivity is less critical than a tuner to also be used for
casual DXing.) How will this further simplify the optimal single
frequency TRF receiver design? Will only one RF amp stage be
necessary, or will we still need two? The focus now will be on very
high-quality audio reproduction of local stations, which I believe
tubeophiles will be most interested in.


Thanks.

Jon Noring

How many stages you need depends on the selectivity you need because of your
geographic location and antenna Two stations close in frequency will
interfere with each other unless you have enough selectivity. A weak local
station that is strong enough to be heard may get splatter from a distance
station close in frequency that has 50KW or more and a pattern that
concentrates on your area! High Q, double tuning, addtional stages all add
to selectivity. The spuerhet solved this problem.

But if you want to satisfy your demons with a mdoular approach you might
want to consider salvaging a couple of turret type tuners from 1950s TVs.
These have clip in moddules with silver contacts and appropriate LC for each
channel. With todays ferrites you should be able to squeeze-in a LC for AM
BCB. Each tuner will give you 12 channels.

--
73
Hank WD5JFR

"Jon Noring" wrote in message
...
Last night I posted a couple messages regarding the proposed "channel
TRF" AM tube tuner, focusing on the plug-in mini-board idea as one way
among several possibilities to implement it.

The idea underlying the channel TRF concept is to build bandpass
tuning circuitry specific to, and optimized for, each frequency in the
BCB, instead of fixing that circuitry to some "average" value and
trying to vary it using a traditional variable air capacitor (or
variable inductor) for continuous tuning. A switch would be used to
select the bandpass circuitry for the particular frequency channel the
listener wants to hear.

This would allow, in principle if not in practice, the ability to
very precisely optimize the bandpass circuitry (to maintain a quite
constant bandwidth and shape) for every broadcast frequency in the BCB
(from 500 khz to 1800 khz.)

The "mini-board" variation of the concept would place the bandpass
circuitry for each channel (frequency) onto a small plug-in PCB board.
Depending upon the type and order of bandpass filter used, the number
of components on the mini-board may be quite small, maybe a couple
capacitors, a resistor or two, an inductor, etc., having the optimal
values, and with one or more trimmers for fine adjustment of the
center frequency.

Clearly there are several implementations of the general concept, one
of which is a well-known hybrid that allows continuous tuning in the
more traditional and familiar way. The ones I think of at the moment
a

1) Traditional continuous tuning: Divide the wide BCB into several
sub-bands, such as 5 or even more, each sub-band having optimized
bandpass circuitry for the sub-band, and then use the traditional
variable capacitor or inductor to tune within the narrow sub-band.
Although each channel will no longer have the most optimal bandpass
configuration, it will be closer to optimal.

2) Single Board, True Channel: It may be possible, instead of having
120+ totally independent channel circuits each placed on a separate
mini-board, to put them all onto one larger board, but still keep
all circuits otherwise separate on the board. A lot of components,
and probably a lot of trimmers.

3) Single Board, Shared Components: As a combination of items (1) and
(2), channels which are adjacent to each other (in their own
"sub-band") could probably share a lot of common bandpass
components, thereby reducing the number needed on the board. Only
the large number of trimmers for individual channel calibration
will remain.

The original idea of mini-boards is most advantageous when the user
of the TRF tube tuner only plans to listen to 10-20 stations (such
as local, higher-power stations). They only install the channel
mini-boards they want to listen to.


*****

I do have a couple questions of both John and Patrick (and anyone else
caring to chime in) related to this.

1) In the single frequency TRF tube receiver (a TRF designed strictly
to listen to a single frequency), is there a need for double tuned
circuits? Or will singly tuned circuits be sufficient for
excellent performance (audio quality, sensitivity and
selectivity)? If not, how do double tuned circuits benefit the
overall performance of the single frequency TRF receiver?

2) Let's assume that we decide to design a Mark I TRF AM tube tuner
kit designed solely for more local, higher power stations (thus the
sensitivity is less critical than a tuner to also be used for
casual DXing.) How will this further simplify the optimal single
frequency TRF receiver design? Will only one RF amp stage be
necessary, or will we still need two? The focus now will be on very
high-quality audio reproduction of local stations, which I believe
tubeophiles will be most interested in.


Thanks.

Jon Noring

Henry Kolesnik wrote:

How many stages you need depends on the selectivity you need because
of your geographic location and antenna. Two stations close in
frequency will interfere with each other unless you have enough
selectivity. A weak local station that is strong enough to be heard
may get splatter from a distance station close in frequency that has
50KW or more and a pattern that concentrates on your area! High Q,
double tuning, addtional stages all add to selectivity. The superhet
solved this problem.

Hmmm, the "channel TRF" approach may help with one stage selectivity
since it appears we can now use a perfectly optimized higher order
tuned filter on the channel mini-board, while in a traditionally tuned
circuit, implementing that same bandpass circuit to apply across the
whole BCB will be much more difficult, and I would guess be near
impossible (too many circuit components which need to be varied
simultaneously as one varies the reception center frequency.)

With the channel TRF approach, the tube-o-phile can mix and match
bandpass filter types from station to station depending upon the
circumstances. For example, they could use the default, wider-band,
gentler, bandpass filter plug-in board (one which has better linear
phase) for a local station which doesn't have adjacent interference,
and for a more difficult station (with adjacent interference) they can
use a bandpass filter plug-in board with a shape factor closer to
unity (which probably has more ripple and worse linear phase). (Even
for the default "wider-band" filter, because we can now use a
frequency optimized higher order filter, we should be able to achieve
reasonably good selectivity, at least sufficient for local station
reception, even with one RF amp stage.)

There appears to be a lot more freedom given to the circuit designer
when the necessity of tuning a fixed set of tuning components over a
frequency range is removed, such as using higher order bandpass
filters. (Of course, this is one reason for IF, but even superhets
have at least one tuned RF amp before the mixer, so the same issue
applies to superhets, but is not as critical.)

I now wonder that with a single TRF RF amp stage, and with a higher
order bandpass filter optimized for a particular frequency, if we can
now dispense with the RF transformer? Or does an RF transformer confer
other benefits that it should remain? I thought its main benefit was
for improved bandpass shaping, but then I may be wrong here (likely
with high probablity -- RF transformers do help with isolation of
stages for DC, so I've read, but don't know how that would benefit
real tuner circuit design.)

Jon Noring

In article , wrote:

*****

I do have a couple questions of both John and Patrick (and anyone else
caring to chime in) related to this.

1) In the single frequency TRF tube receiver (a TRF designed strictly
to listen to a single frequency), is there a need for double tuned
circuits? Or will singly tuned circuits be sufficient for
excellent performance (audio quality, sensitivity and
selectivity)? If not, how do double tuned circuits benefit the
overall performance of the single frequency TRF receiver?

The advantage of the double tuned circuit over two single tuned circuits
is that it has a better shape factor, the pass band response is flatter.
This advantage applies to single frequency as well as tunable receivers.
Even better response curves can be achieved with three, four, or even six
section filters.


Regards,

John Byrns


Surf my web pages at, http://users.rcn.com/jbyrns/

Robert Casey wrote:
Jon Noring wrote:

I guess the question to ask is how much better can be done when
continuous tuning is eliminated, and one uses an optimum bandpass
circuit for each channel frequency? (It appears possible to get
almost uniform bandwidth and shape across the entire BCB spectrum.)
For high-fidelity audio purposes in the channel TRF concept, what
order and type of RF bandpass filter circuitry suggests itself?

The upside of several single channel TRF circuits is that each
channel's TRF circuits can be custom tuned for frequency and
bandwidth. The downside is that this will require a lot of parts
for more than a few selected channels, and that if someone moves to
another radio market a lot of retuning is required if all channels
aren't built to begin with.

Yes, this is a downside, but as I see it now, it may not be that much
of a burden -- it depends upon the use. It also opens up many
interesting opportunities for the hobbyist.

There are two general approaches to wiring up the independent bandpass
filters for some or all of the BCB channels:

1) Hardwire all the filter components for all the BCB channels onto
one large board (we'd have an "American" board and a "European"
board, both premade PCB.) This is not trivial, and we could have
upwards of 1000 small RLC components needing to be soldered on the
board, depending upon the order of the filter we want to use.
That's a whole lot of work. It is also inflexible -- the whole
board must be committed to one particular bandpass filter type and
order (e.g., it must be a 4th order Butterworth -- bandwidth is
adjusted by altering the values of the components soldered in as
the supplied "chart" will indicate.)

2) The mini-board idea, where the filter components for a single
channel frequency are put onto a small PCB mini-board. The user
plugs the mini-board into a slot to connect it to the RF amp
section (probably with antenna tuning as well.) We could imagine
having a large PCB "motherboard" which has up to 130+ plugin slots
(not unlike those used for PCs, but we need only have a small
number of contacts per slot -- the number I can't guess at the
moment.) A switch will also be needed (is an electronic switch a
possibility?) Of course, a smaller board with 20 slots, with a
twenty position switch, could be made for those who do not
anticipate tuning anymore than 20 channels. For local listening
(especially for the simple 1 RF amp stage tuner where it won't
be very sensitive), this is probably more than enough channels.

The advantage of this approach is that the user needs only to get
boards for the BCB channels they will listen to, and will have the
ability to alter the bandpass characteristics for a particular
channel (just wire up a different mini-board tuned to that
frequency.) For example, one could have a 7th order Chebychev for
1130 khz with a bandwidth of 15 khz, and a 4th order Butterworth
for 750 khz with a bandwidth of 10 khz. At a later time, the user
can change the bandpass filter used for any particular channel --
just swap mini-boards.

Now how big does the mini-board have to be? I don't have a good
feel for this, while the experienced radio builders out there will
have a much better idea. But let's look at what the mini-board
will contain. Essentially it will contain the RLC bandpass filter
components (plus a trimmer or two for fine calibration of the
center frequency). Depending upon the order of the filter used,
it may have anywhere from 5 to 10 RLC components (again just a
guess -- the very high order bandpass filters will have more.) So
the mini-board will need to be big enough to hold these components.
Again, I think most of them will be fairly small in size, so it is
not inconceivable for the mini-board to be as small as, for
example, 1" x 2" (again, only a guess -- anyone?)

I also foresee that there will be a standard bandpass filter for
the channel TRF tube tuner (a given type and order -- what would
you use for a single RF amp TRF tube tuner for local listening?)
One can have a large number of PCB boards made for that bandpass
filter. Then, for a given channel frequency (e.g., 830 khz), and
a chosen desired bandwidth, the kit-builder refers to the table of
values for each component (e.g., this resistor will be 50 ohms,
that capacitor 5 pf, etc.), solders them in, then fine calibrates
the center frequency. I don't imagine these boards, when made in
bulk, will be that expensive, neither the components be, nor will
it take much time to solder the components onto the mini-board --
maybe only a few minutes.

The biggest issue I foresee is the fine calibration of the bandpass
center frequency -- can that be done independent of the tuner (thus
allowing the kit-supplier to make them available on order) or must
the mini-board be plugged into the tuner? Since there will be some
distance between the RF amp tubes and the bandpass filter, with
intervening wire, a switch or two, and slot connectors, there will
be interwire resistance, capacitance, etc. For a real world tuner,
how important will this be?


As to the question of what the unused channels will do to the
receiver, I'd do the switching such that all unused channel circuits
are completely switched out, and maybe tied to ground.

Agreed, although there are probably other possibilities.

One interesting aspect of this design is that an enthusiast could add
a more traditional continuous tuner if they wanted to (e.g. with a
multiganged variable air capacitor or inductor -- just build it
separately and plug it into one of the slots.

I really do think there are other interesting things one might do
with the "channel TRF" tuner concept. I think we have just scratched
the surface.

Jon Noring

Henry Kolesnik wrote:

I haven't seen a technical reason whatsoever that a single channel
TRF perhaps switchable would have any verifiable advantage over a
superhet. But the thread continues to perhaps imply that there might
be something. Did I miss something? I'd sure like to know.

Patrick Turner noted the following in a thread from early this year,
when someone asked about a single frequency BCB tuner:

"Since you only want one channel, there is no need for a frequency
converter or any IFTs or IF amps, and a TRF with four tuned circuits
in the form of two critically coupled RF trannies will do nicely."

My understanding of the primary reason why superhet was designed was
to allow the most important amplification to be done at a single
frequency (the IF frequency), so the tuning circuit can be optimized
for that fixed frequency. (I'm sure Patrick and John Byrns will be
able to more accurately explain the advantages of the superhet, but
that's how I understand it in 10 words or less.) But if we already
have a single-frequency tuner, there's no need for an IF stage since
we can optimize the bandpass tuner sections for that particular
frequency.

So for a single frequency tuner, adding an IF stage only complicates
the circuitry -- it is superfluous -- and will add more distortion to
the final audio signal (albeit small, I assume, for a well-designed IF
stage.) So why use it? (It's been said a superhet may confer better
stability, whatever that means -- again a topic for Patrick or John to
address.)

Now, comparing a multichannel TRF tuner (with each channel having its
own optimized bandpass filter circuitry) to superhet tuner, then one
compares the complexity of switching individually tuned optimal RF
bandpass circuits with the complexity of adding a multigang tuning
capacitor (or inductor) and an IF stage. Also, there is the factor of
audio quality.

As I see it at the moment (subject to change as everything comes into
better focus over time), a cross-over point between choosing the
"channel TRF" and the traditional superhet for a tube-based BCB tuner
appears to lie between:

1) Listening to local stations, wanting the highest possible audio
quality, and

2) Casual to medium-serious DXing.

For (1), the listener only needs 20 or so channels, and the number of
RF amps can be kept to one or two (two for some added sensitivity to
pull in fairly weak local stations), so the "channel TRF" is more
attractive for this purpose (particularly for audio fidelity.)

For (2), for a "channel TRF", there'd probably have to be three RF
amps, with the full complement of optimized bandpass filters for all
130+ channels installed, so at this point a traditional superhet is
strongly indicated.

Nevertheless, even for DXing, the "channel TRF" is still intriguing
for those who might want to experiment, especially for the ability to
quickly swap bandpass filters (for changing the type and order, and
not only the bandwidth.)

Just my $0.02 worth.

Jon Noring

[Obligatory Telamon rec.radio.shortwave on-topic statement: The
following reply touches upon aspects of TRF design as they apply to MW
DXing, particularly sensitivity and selectivity of tube-based TRF
designs. MW DXing is on-topic to r.r.s. discussion. I appreciate
Telamon's efforts to keep r.r.s discussion strictly on-topic per the
written r.r.s. Charter.]


Patrick Turner wrote:
Jon Noring wrote:

Patrick Turner noted the following in a thread from early this year,
when someone asked about a single frequency BCB tuner:

"Since you only want one channel, there is no need for a frequency
converter or any IFTs or IF amps, and a TRF with four tuned circuits
in the form of two critically coupled RF trannies will do nicely."

The design I evolved for my variable frequency superhet leaves
behind all of the many AM radios I have repaired as new and tested
in my kitchen for comparison. Even an old Quad AM tuner I have is no
better.

Patrick, have you published the schematic for your radio? How amenable
is your design for turning it into a kit? And what are its overall
specs?

It sounds like a good MW DXer combined with high audio quality.


A TRF fixed F tuner with say two j-fets could be cobbled up
with preset RF transformers, in the same format as the double tuned
IFTs and with necessary couplings and input LC to broaden the pass
band. Miniature sized coils and cans could be used and each module
with the RF amps would fit on a board about 70mm x 40 mm, which is
2,800 sq.mm.

Thanks. The single frequency tuning board you sized is somewhat close
to the size I essentially guessed at. So we'll go with your estimate
of 70mm x 40mm.


120 such modules could then fit on a board about 350,000 sq.mm,
so that 10 boards would each be 35,000 sq mm, and about 190 mm x 190
mm, and thus all fit in a box the same size as a variable tuner
radio from the 1940s. Chuck in an extra board with tubed detector
and audio preamp, and 120 position switch, and a PS, and you're
done.

Another way to look at this is by a volumetric analysis, since one can
take advantage of plugging modules in a motherboard-like fashion.
Let's assume the modules will be 70 mm by 40 mm, and let's assume we
have to space the modules 19 mm apart because of the height of the
components soldered on. (Is this reasonable? -- 19 mm is the spacing
between PCI boards on a PC MB.) This means each module must minimally
take up a volume of 53,200 mm^3. 120 modules therefore will occupy a
minimum of 6,384,000 mm^3 (6.4 liters). This works out to a cubic box
186 mm on a side, or in English units 7.3" on a side (alternatively,
it works out to about 390 cubic inches for those not used to working
in metric.)

Of course, a cube is aesthetically and practically not the way the
modules would be distributed. So let's assume we plug the modules into
the "motherboard", with the long side (70 mm) sticking up. This would
yield a single motherboard footprint of 142 square inches, or about
one foot on a side, and about 3" tall. If we split it into two
motherboards, each holding 60 modules, then the footprint would be
8.4" on a side, with a height of about 6".

Yes, this is not an insignificant volume, but it is not a huge volume.
I notice that the three gang tuning air capacitor on my Philco 37-670
occupies a space of 3"x4"x7", or about 84 cubic inches, about 1/5 the
volume of the 120 module "box" (in a channel TRF, the tuning capacitor
would not be used.) Also remove the small volume taken up by the IF
section (no idea how much volume that typically takes up in a tube
set, but it is not tiny.)

Now, let's look at the 20 module motherboard. Here, the necessary
one-level motherboard would have an area of 24 square inches. Thus, a
6"x4" motherboard (make it 7"x5" for some clearance) will hold 20
modules, plugged in. The height will be 3". Now this seems reasonable
when the end-user only intends to tune in 20 local stations. It is
smaller in volume than the 3-gang tuning air capacitor on my Philco.
Of course, it will only tune 20 stations, and nothing in-between.

From the perspective of tuners in general (not specific to tube type
tuners, but also solid state and digital), the channel TRF does not
make sense. But with respect to a tube-based tuner, it does seem to
make sense for *some* applications. Since the only ones who will even
buy or build an AM tube tuner are tube-o-philes or tube-o-holics
(those who are attracted to tube-based equipment for whatever
aesthetic reason), the aspect of "commercial application" as we
understand it for ordinary radios does not enter the picture.

Those who simply want to get some job done with a radio (listening to
local stations, DXing, etc.), and are not overly enamored with any
particular under-the-hood architecture, will certainly NOT gravitate
to any tube-based tuner because of the much better and cheaper options
out there in the marketplace (digital and SS designs -- I don't know
of any tube-based high-end general coverage receiver being built
today -- and I'd be surprised if someone is attempting it.)

As Patrick noted, and which I agree wholeheartedly, pure digital is
the future of radio for utilitarian purposes (if BCB and FM radio
itself even has a future!)

But that's the point. In this discussion we are not talking about
building a radio for those who want to get a job done, but those who
are enamored with tubes and want the best possible sound out of the AM
tuner. For this purpose, the channel TRF is certainly a viable
candidate, along with a tuned TRF (as John Byrns is apparently working
on), as is the more traditional IF design (which Patrick says he is
working on.)


If the components for each tuner board cost $20, then about $3,000
for the 120 + PS, box, etc, all would be a steal, and a quite cheap
sort of "high-end" price. An asian maker of boards might reduce the
cost by 20 dB to $2 each.

Obviously, the component cost is significantly higher than for a
traditionally tuned circuit because one is using a larger number of
components, most of which will not even be powered while the tuner is
selected to a particular channel frequency. So in a sense, this is a
significant inefficiency.

But for a tuner intended to tune in local stations, the channel TRF
tube tuner appears to have some things in its favor. As a
tube-o-phile myself, one can make several strong arguments in favor of
the channel TRF tube tuner: 1) the circuitry is "clean", no IM mixing,
2) the bandpass filters are *perfectly* optimized for each channel --
no compromises (this is a *huge* attraction), 3) provides the ability
to plugin different bandpass filters for a particular station (if
needed), and 4) *may* be more amenable to a kit than would a full-
blown superhet design.

Now, if a tube-o-phile wants a tube tuner for serious MW DXing (for
whatever reason -- I would not use a tube tuner for *serious* DXing),
then the channel TRF is not down and out, but certainly has its work
cut out for it to try to compete with the continuously-tuned TRF, and
of course with traditional superhet designs. The need to include all
120+ BCB channels does work against the "channel TRF".

(On the other hand, I can see a serious MW DXer build a single-
channel TRF design of three or four RF amp stages where the bandpass
sections are "swappable" to tune the channel wanted to monitor.
Here the design will simply have a single slots for each bandpass
filter stage -- no channel switches. Just swap the mini-boards to
retune to a different frequency.)


I eagerly await your completion of a prototype of just one single
iddy biddy TRF tuner board which has all the discussed and wanted
capabilities with respect to audio BW, distortions, sensitivity,
selectivity to allow local station listening where weak and powerful
stations exist which are only 40 kHz apart, all without spurious
noise, interference, cross modulation, etc.

First, I assume that sensitivity is largely a matter of the RF amp
itself (and number of RF amp stages), not the bandpass filter itself
(although the filter should not overly get in the way of RF amp gain.)
But if the bandpass filter plays a greater role in sensitivity than I
realize, shouldn't an optimally tuned bandpass filter in the channel
TRF concept significantly outperform the limited and sub-optimal
single or double stage bandpass filters one is *forced* to use for
continuous tuning?

Second, each tuning module (for a single frequency) is, by and large,
independent of all the other modules. Thus, this simplifies the design
process since one doesn't have to share the same bandpass component
values from channel to channel, except maybe the RF transformers. This
should make it much easier, not harder, to achieve the performance
goals.

In the channel TRF, we are no longer constrained to single or double
tuning -- we can, for example, have the equivalent of quintuple
"tuning" for a 5th order bandpass filter if we want. I'm assuming
that, for a given frequency, the designer will have full control over
the values of all the LC components (and not just one or two,
excluding the RF transformer, though) in the bandpass filter, thus
making it much easier to achieve selectivity, distortion and other
performance goals, all the while simplifying the main part of the
circuit -- to make it cleaner -- fewer kludges needed.

(I keep looking at advanced radio circuits and see such a spiderweb of
wiring between the various stages, wondering why the hell it is all
there -- I wonder how much of that complexity is due to not being able
to properly optimize the RF bandpass filters for a given frequency,
thus requiring all sorts of work-arounds to get good overall
peformance.)

It is a remarkable achievement for a radio designer to meet the
several specification goals Patrick listed for an AM BCB tuner (his
list appears to be an "all things for all users" dream list) and which
is continuously tunable from 500 khz to 1800 khz (thus necessitating
most of the tuning components be shared.) I have no doubts that
Patrick has come up with a great design.

Superhet definitely helps with accomplishing this feat, but from what
I see, there are a lot of ****ty superhets out there, so if superhet
alone were sufficient, the perfect radio would have been designed
years ago. (Isn't the AA5 that perfect radio? -- it is, depending upon
the definition of "perfect" -- it is "commercially" perfect for the
masses.) IF is not the magic bullet (albeit it is a powerful one), but
simply a nifty tool to get from here to there. But like all nifty
tools, they have their limits and their place. One does not use a
hammer to drive in a screw, for example.


Almost the entire amount of AM radio reception theory that has ever
filled the minds of conscious humans has been repeatedly explained
so far in this thread, so you have all the knowhow you ever wanted,
so what's the hold up? Stop dithering, and go to it man!

Actually the know-how has not been explained in full! :^)


Thanks for your feedback. It is definitely adding useful information
to this thread.

Jon Noring

"Jon Noring" wrote in message
...

snip (here and other places)

Another way to look at this is by a volumetric analysis, since one can
take advantage of plugging modules in a motherboard-like fashion.
Let's assume the modules will be 70 mm by 40 mm, and let's assume we
have to space the modules 19 mm apart because of the height of the
components soldered on. (Is this reasonable? -- 19 mm is the spacing
between PCI boards on a PC MB.) This means each module must minimally
take up a volume of 53,200 mm^3. 120 modules therefore will occupy a
minimum of 6,384,000 mm^3 (6.4 liters). This works out to a cubic box
186 mm on a side, or in English units 7.3" on a side (alternatively,
it works out to about 390 cubic inches for those not used to working
in metric.)


Don't forget you will have to swich those modules in an out of circuit. The
switch and all the associated wiring will significantly add to the space.

The wiring for the switch will also probably affect the tuning of the
modules forcing them to be tuning 'in place'.




Of course, a cube is aesthetically and practically not the way the
modules would be distributed. So let's assume we plug the modules into
the "motherboard", with the long side (70 mm) sticking up. This would
yield a single motherboard footprint of 142 square inches, or about
one foot on a side, and about 3" tall. If we split it into two
motherboards, each holding 60 modules, then the footprint would be
8.4" on a side, with a height of about 6".


If the long end sticks up, you have the inout and output of the modules on
the same end. This could result in unwanted coupling.

Yes, this is not an insignificant volume, but it is not a huge volume.
I notice that the three gang tuning air capacitor on my Philco 37-670
occupies a space of 3"x4"x7", or about 84 cubic inches, about 1/5 the
volume of the 120 module "box" (in a channel TRF, the tuning capacitor
would not be used.) Also remove the small volume taken up by the IF
section (no idea how much volume that typically takes up in a tube
set, but it is not tiny.)


snip


But for a tuner intended to tune in local stations, the channel TRF
tube tuner appears to have some things in its favor. As a
tube-o-phile myself, one can make several strong arguments in favor of
the channel TRF tube tuner: 1) the circuitry is "clean", no IM mixing,

Conceptually it may appear clean, but the proposed switiching of many
modules adds a new complexity. All the wiring associated with a swicth will
cause more problems that the design solves.

2) the bandpass filters are *perfectly* optimized for each channel --
no compromises (this is a *huge* attraction),

If you are looking for broad band with reasonable attenuation of other
stations, a superhet is much better as there is only one signal you are
optimizing for.

3) provides the ability
to plugin different bandpass filters for a particular station (if
needed), and 4) *may* be more amenable to a kit than would a full-
blown superhet design.

I would think that a kit should be simple, the proposed solution is not.



First, I assume that sensitivity is largely a matter of the RF amp
itself (and number of RF amp stages), not the bandpass filter itself
(although the filter should not overly get in the way of RF amp gain.)
But if the bandpass filter plays a greater role in sensitivity than I
realize, shouldn't an optimally tuned bandpass filter in the channel
TRF concept significantly outperform the limited and sub-optimal
single or double stage bandpass filters one is *forced* to use for
continuous tuning?

If you are only looking at a single aspect of the design (get all unwanted
signals out at the earliest point in the radio) then you may be correct. But
if you look at the overall design, then you will see there are tradeoffs
that must be considered. If the front end can tolerate the unwanted signals,
then IF filtering can deal with them and you have a workable solution that
does not have the alignment problems you would look at with a TRF design.


Second, each tuning module (for a single frequency) is, by and large,
independent of all the other modules. Thus, this simplifies the design
process since one doesn't have to share the same bandpass component
values from channel to channel, except maybe the RF transformers. This
should make it much easier, not harder, to achieve the performance
goals.

snip

(I keep looking at advanced radio circuits and see such a spiderweb of
wiring between the various stages, wondering why the hell it is all
there -- I wonder how much of that complexity is due to not being able
to properly optimize the RF bandpass filters for a given frequency,
thus requiring all sorts of work-arounds to get good overall
peformance.)


I think you really need to understand that spiderweb before making that kind
of statement.


snip


Almost the entire amount of AM radio reception theory that has ever
filled the minds of conscious humans has been repeatedly explained
so far in this thread, so you have all the knowhow you ever wanted,
so what's the hold up? Stop dithering, and go to it man!

Actually the know-how has not been explained in full! :^)


And that know-how does not all need to be explained here. There are plenty
of resources on the internet, look for them and study them.


Thanks for your feedback. It is definitely adding useful information
to this thread.

Jon Noring


IMO this thread and the related theads are more on-topic for rrs than much
of the stuff posted by one of the people considering this to be off topic.

craigm

[Obligatory Telamon rec.radio.shortwave on-topic statement: The
following reply touches upon aspects of TRF design as they apply to
MW DXing, particularly sensitivity and selectivity of tube-based TRF
designs. MW DXing is on-topic to r.r.s. discussion. I appreciate
Telamon's efforts to keep r.r.s discussion strictly on-topic per the
written and published r.r.s. Charter.]


craigm wrote:

Don't forget you will have to switch those modules [all 120+] in and
out of circuit. The switch and all the associated wiring will
significantly add to the space.

Granted, it will. Some of the wiring hassle is removed if we use a
ready-made PCB motherboard (see, for example, the ST-35 clone board at
http://www.diytube.com/ ). But, unless one can think of some clever
electronic switching arrangement, there will be wires from the
switch to the board. One would make sure the wires from the switch
would solder on to pins along one side of the motherboard (and
appropropriately shielded), but for 120+ channels could be 480+ wire
connections (assuming we can get by with four wires per channel to
feed the bandpass filter(s), which is unclear at the moment) --
definitely formidable, but not outside the realm of solvability by
clever design and utilization of modern components.

With 15-20 channels (for local listening purposes), however, the
situation is much more reasonable all around.

As I've noted recently (a shift in the requirements), the channel
TRF concept makes the most sense for an audiophile-acceptable tube
tuner to listen to strong, local stations. Even the so-called "RCA
"high-fidelity AM tube tuner" is a TRF design, not an IF. Thus, having
it optimally tuned for 15-20 stations (channels) will be acceptable
for this purpose, and in fact may be a good selling point among these
people.

If one wants all 120+ stations soldered in, that means the user is
interested in tuning the whole BCB, which automatically means DXing.
Now we are in a different ballpark, and the user expectations are
different.

[Here, I would consider John Byrns advice and design a doubly tuned
TRF design, such as his proposed "modernized" Western Electric 10-A.
One interesting twist, to get more optimum bandshaping across the BCB,
divide the wide-ranging BCB into five or more tunable bands; thus, for
example, band 1 would tune from 500-650 khz, band 2 from 650-850 khz,
band 3 from 850-1100 khz, band 4 from 1100-1400 khz, and band 5 from
1400-1800khz (or whatever makes sense.) Each band would have its own
singly or doubly tuned bandpass circuitry, optimized to that band (but
not optimized to any particular frequency). In fact, with a narrow
enough sub-band, one should be able to get pretty good bandshaping
with single tuning, I presume, which is a welcome simplification.]


The wiring for the switch will also probably affect the tuning of
the modules forcing them to be tuning 'in place'.

Yes, that is definitely a consideration which I've noted before, the
affect of the interwire/interconnect RLC on the tuning circuit. With
continuous tuning, this is not an issue.

With the 15-20 channel system for local tuning, it may not be a
problem either if we add to the circuit some calibration indicator
(like the tuning light of old.) For example, the user makes or buys a
mini-board for 830 khz (a local station they want to listen to.) They
plug it in. They then turn on the tuner (letting it warm up fully).
Once sufficiently warmed up, they then calibrate the mini-board by
turning a trimmer on the mini-board to fine tune the station, until
either the station sounds as if it is "in tune", or the calibration
indicator light shows it to be tuned to the correct center frequency.

(Also as I've noted before, it will no doubt be important for there to
be a fine tuning control on the tuner itself, to fine tune +/- 1 khz
(or thereabouts) to account for warmup and for long-term drift of the
component values inbetween calibrations.

But many audiophiles and especially kit-building tube-o-philes love to
tweak their stuff -- they'll enjoy this, and they will also be
enamored in having the most optimum bandpass tuning circuitry for that
frequency -- it's a performance/sound issue. Some may even wish to
swap plugin boards, to try different bandpass types, order and
bandwidth (for some stations they may have to because of adjacent
interference.)

In a sense, an IF design is boring when looked at from this angle.
laugh/.


If the long end sticks up, you have the inout and output of the modules on
the same end. This could result in unwanted coupling.

Granted. One of those problems which needs to be sorted out in the
design of the whole "tuning box" architecture. It is a problem, but so
far does not appear to be a show stopper. It's one of those items that
still falls under the category "to be solved by appropriate board and
wiring design".


Conceptually it may appear clean, but the proposed switiching of
many modules adds a new complexity. All the wiring associated with a
swicth will cause more problems that the design solves.

For 120+ channels, yes, it looks like a plumbing nightmare unless
someone can come up with a clever idea (and that is certainly
possible).

But with 15-20 channels, it is entirely workable. Old TVs worked with
12 channel switches (channels 2-13.)

And I'm not familiar with what could be done with modern electronic
switches.


If you are looking for broad band with reasonable attenuation of
other stations, a superhet is much better as there is only one
signal you are optimizing for.

Well, the same applies to the channel TRF concept. When we switch in a
particular bandpass tuning circuit, it is calibrated for a single
center frequency, with the optimum tuning circuit for that frequency.

Note again that Patrick himself said that IF is not needed when one is
building a single frequency receiver -- and from his comments he is a
very strong advocate of superhet design for a tunable receiver. That's
all the channel TRF is: a single frequency receiver, duplicated n
number of times (where n is the number of channels one wants to tune,
which are switched in and out.)

The downsides of a channel TRF are obvious: plumbing (wiring)
complexity for an all-channel BCB tuner, and not being able to
continuously tune all frequencies within the BCB.


I would think that a kit should be simple, the proposed solution is not.

At 120+ channels, the channel TRF is intimidating (the switch box and
individual tuning circuits), but at 15-20, with the plugin
architecture I am thinking of, it does not look that complicated,
especially if a lot of the architecture and components we see used in
PCs can be utilized.


And that know-how does not all need to be explained here. There are
plenty of resources on the internet, look for them and study them.

I've definitely done that!


IMO this thread and the related theads are more on-topic for rrs
than much of the stuff posted by one of the people considering this
to be off topic.

I've put in the obligatory Telamon preamble stating this message is
on-topic to r.r.s., so all should be fine with the Usenet gods (tm).

Thanks for your informative feedback.

Jon Noring

Jon Noring wrote in message . ..
[Obligatory Telamon rec.radio.shortwave on-topic statement: The
following reply touches upon aspects of TRF design as they apply to MW
DXing, particularly sensitivity and selectivity of tube-based TRF
designs. MW DXing is on-topic to r.r.s. discussion. I appreciate
Telamon's efforts to keep r.r.s discussion strictly on-topic per the
written r.r.s. Charter.]

Telamon is being a petulant freak and the "usenet gods" as you mention
are dealing with him. Meanwhile I'm thinking of killing this file too.
So far I haven't cared to bother about finding out where your nonsense
really emanates from either. Other people have some interest in the
thread and it wouldn't be fair to them. My address is real though. You
won't get any reply without one, even if you do try it.