[John Byrns]

In article ,
flipper wrote:

On Wed, 27 Jul 2011 13:08:40 -0500, John Byrns
wrote:

In article
,
Patrick Turner wrote:

snip

I do not know if the synchrodyne balanced demodulator can produce a
Vdc which may be used for VFO F control in addition to producing audio
output.

No, you need separate detectors for the ³PLL² and to recover the audio. The
two
detectors must be provided with carrier signals from the oscillator 90
degrees
out of phase with each other. Maintaining the required 90 degree phase
shift
over a 3:1 frequency range is another big potential gotcha for a tube based
direct conversion receiver. In a solid state design this problem would be
simply dealt with by operating the oscillator at 4 or 8 times the received
frequency, and using digital dividers to generate the 90 degree phasing
between
the two local oscillator signals. A large number of tubes would probably be
required to do it this way in a tube based circuit.

Pardon me if this is "not a radio guy" dumb idea but that is one
reason I was thinking about the 6ME8 because you can get at it single
ended.

Looking at the bottom of page 9 here...

http://jlandrigan.com/files/Receiver...0Using%20the%2
07360.pdf

The 6ME8 has different gain and biasing but it's functionally the same
thing so I'm pondering that product detector as a sort of starting
point.

Add a reactance tube to make a VCO of the LO.

I'm Following you to this point, beyond here I'm either not following you, or
it's a "not a radio guy" dumb idea, or most likely a little bit of both. I'm
also not familiar with the application of this tube, although I have at least
heard of it.

Now, seems to me you might could get the control signal from a PP
transformer on the 6ME8 outputs into a single secondary. (You'd also
need to at least R isolate the single ended outputs rather than just
RF ground them). Filtered, of course.

You mention getting the control signal from the secondary of a PP transformer, I
don't see how that could work as the control signal is DC? In any case what are
the "single ended outputs" and the Rs to isolate them all about?

I think you're still 90 degrees but that just means a lower single
ended amplitude, doesn't it? Which I'm thinking is a lesser problem
than the other.

If it is at 90 degrees, which is what the PLL would be driving towards, the
output would be zip. Say it isn't quite 90, the problem is that the phase angle
changes the amplitude of the recovered modulation, but the uncorrelated noise on
the received signal remains constant, so the SNR of the recovered audio goes to
pot when the phase angle differs from zero.

Obviously just an 'idea' and not a design.

Unfortunately I'm not "a radio guy" either, I just played one while I was in
college and for a year after I graduated, so what I say could easily be total BS.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
flipper wrote:

On Wed, 27 Jul 2011 20:35:32 -0500, John Byrns
wrote:

In article ,
flipper wrote:

On Wed, 27 Jul 2011 13:08:40 -0500, John Byrns
wrote:

In article
,
Patrick Turner wrote:

snip

I do not know if the synchrodyne balanced demodulator can produce a
Vdc which may be used for VFO F control in addition to producing audio
output.

No, you need separate detectors for the ³PLL² and to recover the audio.
The
two
detectors must be provided with carrier signals from the oscillator 90
degrees
out of phase with each other. Maintaining the required 90 degree phase
shift
over a 3:1 frequency range is another big potential gotcha for a tube
based
direct conversion receiver. In a solid state design this problem would
be
simply dealt with by operating the oscillator at 4 or 8 times the
received
frequency, and using digital dividers to generate the 90 degree phasing
between
the two local oscillator signals. A large number of tubes would probably
be
required to do it this way in a tube based circuit.

Pardon me if this is "not a radio guy" dumb idea but that is one
reason I was thinking about the 6ME8 because you can get at it single
ended.

Looking at the bottom of page 9 here...

http://jlandrigan.com/files/Receiver...s%20Using%20th
e%2
07360.pdf

The 6ME8 has different gain and biasing but it's functionally the same
thing so I'm pondering that product detector as a sort of starting
point.

Add a reactance tube to make a VCO of the LO.

I'm Following you to this point, beyond here I'm either not following you,
or
it's a "not a radio guy" dumb idea, or most likely a little bit of both.
I'm
also not familiar with the application of this tube, although I have at
least
heard of it.

Sorry, I was way too brief about my "not a radio guy" dumb idea.

And my reading comprehension was compromised last night, I failed to notice your
statement ³Looking at the bottom of page 9 here...², and even though I read
through the entire paper, I ended up looking at the schematic on page 2.

Now, seems to me you might could get the control signal from a PP
transformer on the 6ME8 outputs into a single secondary. (You'd also
need to at least R isolate the single ended outputs rather than just
RF ground them). Filtered, of course.

You mention getting the control signal from the secondary of a PP
transformer, I
don't see how that could work as the control signal is DC? In any case what
are
the "single ended outputs" and the Rs to isolate them all about?

First, there's two take offs in my crazy idea with one being 'like'
(but not exactly) the shown single ended audio and then the 'added on'
PP transformer. The added on transformer passes IF and is post
filtered to DC.

Can you explain how the IF would be ³post filtered to DC²?

I'm thinking each half of the primary in series with the plates, like
your typical PP amp output stage, which means the plate signals can't
be AC bypassed to ground in the single ended audio take off or else
the transformer feed is screwed. Might could put a second transformer
in there for the audio but I was thinking simple R isolation, between
the plates and filter, might be sufficient. I'd duplicate it on both
sides, even though one is 'unused' for audio, to keep plate balance.


I think you're still 90 degrees but that just means a lower single
ended amplitude, doesn't it? Which I'm thinking is a lesser problem
than the other.

If it is at 90 degrees, which is what the PLL would be driving towards, the
output would be zip. Say it isn't quite 90, the problem is that the phase
angle
changes the amplitude of the recovered modulation, but the uncorrelated
noise on
the received signal remains constant, so the SNR of the recovered audio goes
to
pot when the phase angle differs from zero.

I think the question is output of what? There's 2 in my wacky idea: a
PP transformer with the other being a single ended audio take off.

Output of the demodulated audio.

At 90 degrees the PP transformer's single secondary should average
(filtered) to 0, which is right for 0 phase error. Phase error to
either side should produce a DC (after filtered) error in that
polarity to drive the VCO.

I still don't understand how the DC is produced, through a transformer?

At 90 degrees we have equal but opposite audio and carrier on each
plate (which is why it nulls in the PP tranny) but the audio takeoff
is single ended, which isn't 'null'. I'm guessing it's half, 6 dB
down, from what you'd have in the 'ideal case' (assuming all else was
equal) of creating a quadrature Lo signal into a second mixer but it's
a heck of a lot simpler and maybe 6 dB isn't too horrible a sacrifice.

The wanted audio, although not the uncorrelated noise, is attenuated to zero at
90 degrees, not just 6 dB, even from a single ended output. Remember we aren't
talking SSB here, where the phase of the reinserted carrier doesn't affect the
amplitude of the demodulated audio, we are talking DSB where carrier phase is
critical.

An alternate might be to use a dual primary, dual secondary,
transformer for the audio and flip the phase on one secondary so it
adds.

That's fine for the audio, but it still doesn't explain how the control signal
is derived.

Did I manage to sensibly convey the idea?

Yes, except for the part about how the DC control signal gets through the
transformer. Even ignoring that, I don't think the idea will work because it
violates what we learned in High School Trigonometry class. I don't believe
that the system can be made to work without using two separate ³modulators² or
³phase detectors², one operating at 90 degrees to provide the control signal and
a second operating in phase to recover the audio.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
John L Stewart wrote:

Did anyone mention IBOC in North America? That will pretty well ZAP any
improved AM detector & many ordinary ones as well. Listen in for IBUZ!

I don't think anyone mentioned that, however the OP who asked the question is
located in OZ, do they use ³IBOC² in OZ?

The wording of your comment suggests that IBOC may not ³ZAP² some ³ordinary² AM
detectors, was it your intention to imply that, or was it just an artifact of
your writing style? If it was intentional could you reveal which ³ordinary² AM
detectors are not ZAPPED by IBOC.

Maybe you, or someone else on the group, can enlighten us about the nature of
the IBOC signal and how it affects various types of detectors? I get the
impression that there are both in-phase and quadrature IBOC subcarriers in the
areas greater than 5 kHz above and below the main carrier which must be removed
by filtering. I originally had the impression that the IBOC subcarriers within
5 kHz of the main carrier were in quadrature with it, or otherwise arranged so
that they would have minimal impact of the SNR of the demodulated audio below 5
kHz. Is this true, I seem to remember reading something more recently that said
it wasn't true? If it is true that the subcarriers within 5 kHz are in
quadrature, a true synchronous detector would reject the ZAPPING better than an
envelope detector. Also if a post detection filter is used to eliminate the
noise above 5 kHz, a synchronous detector would also be more effective. If very
narrow IF filtering is used to eliminate the subcarriers more than 5 kHz from
the main carrier, then the detector probably has little effect, although we are
left with typical AM radio frequency response beginning to cutoff at 2 kHz,
rather than getting a full 5 kHz audio response.

I would be surprised if a true synchronous detector doesn't resist ZAPPING by
IBOC better than an ³ordinary² detector.

Does any one know the details of the IBOC signal, and how the subcarriers are
constructed?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
flipper wrote:

On Thu, 28 Jul 2011 08:28:06 -0500, John Byrns
wrote:

In article ,
flipper wrote:

Sorry, I was way too brief about my "not a radio guy" dumb idea.

And my reading comprehension was compromised last night, I failed to notice
your
statement ³Looking at the bottom of page 9 here...², and even though I read
through the entire paper, I ended up looking at the schematic on page 2.

Is that maybe still confusing the AM demod issue?

???

I've tried doing some simulations substituting NPNs because I don't
have a sheet beam tube model but I'm so ****ed at this POS
Circuitmaker.

Now, seems to me you might could get the control signal from a PP
transformer on the 6ME8 outputs into a single secondary. (You'd also
need to at least R isolate the single ended outputs rather than just
RF ground them). Filtered, of course.

You mention getting the control signal from the secondary of a PP
transformer, I
don't see how that could work as the control signal is DC? In any case
what
are
the "single ended outputs" and the Rs to isolate them all about?

First, there's two take offs in my crazy idea with one being 'like'
(but not exactly) the shown single ended audio and then the 'added on'
PP transformer. The added on transformer passes IF and is post
filtered to DC.

Can you explain how the IF would be ³post filtered to DC²?

It might need to be rectified but basically you have 455kHz phase
pulses coming out. When 90 degrees they will be equal and opposite,
for a net of 0. On either side of 90 they aren't equal so you can
derive a DC offset from that.

As far as I know there are only three signals at each anode:

1. A DC signal, riding on the B+ present at the anode, that is proportional to
the carrier level multiplied by the cosine of the phase angle between the
received carrier and the locally generated carrier. Relative to the B+ on the
anode this DC component reaches a maximum when the phase offset between the
received and locally generated carriers is zero degrees, becomes zero when the
phase offset is 90 or 270 degrees, and becomes a maximum in the negative
direction at 180 degrees phase offset.

This DC signal can be used to control a VCO in a "PLL" by using a differencing
circuit to subtract the DC voltage at one anode from the other thereby canceling
the B+ voltage, this differencing works because the desired DC voltages on the
two anodes move in opposite directions since the IF phase driving the second
anode is 180 degrees out of phase with that driving the first anode. As a
control voltage for a VCO in a "PLL" the VCO phase would be driven towards 90
degrees where the DC difference signal is zero, positive on one side and
negative on the other.

2: The recovered modulation, a.k.a. the audio, voice music or whatever. The
amplitude of the recovered modulation varies with the same cosine function of
the phase offset as the DC component does. The recovered audio is maximum at
zero degrees phase offset, inverted at 180 degrees phase offset, and is nulled,
or becomes zero, at a 90, or 270 degree phase offset. This is in direct
conflict with the action of the DC control voltage which drives the loop close
to a 90 degree phase offset. At this point the modulation recovery is minimal,
and to make matters worse the quadrature noise component of the received signal
is maximized, seriously degrading the Signal to Noise Ratio.

3. A component at twice the carrier, or IF, frequency. I would have to get out
my High School Trigonometry book to be sure, but I think the amplitude of this
2Fc signal is constant, independent of the phase offset between the received
carrier and the local carrier, with only the phase angle changing as the phase
offset changes.

If a transformer is used in the anode circuits of the tube, only the recovered
modulation and the 2Fc signal would make it through to the transformer
secondary. The audio signal doesn't seem like it would be particularly useful
as a control signal for our "PLL", especially considering that it may not always
be present, as in quiet spots in a program, although granted quiet spots are
frowned upon in modern radio broadcasting.

The 2Fc signal could be compared with a 2Fc reference, say 910 kHz in the IF
case, and used to lock the loop. I believe this loop would lock at either of
two phase angles, which would result in a potential ambiguity in the phase of
the recovered audio, which may not be important, except that the phase could
flip from time to time. Also, without getting out my High School Trigonometry
book, which by now it should be obvious I am trying to avoid, it is not obvious
to me what the phase offset at 455 kHz would be, at lock. If it is not the
desired zero or 180 degrees, a simple phase shift network in the 910 kHz circuit
should fix the problem. The 455 kHz signal is of course derived from the 910
kHz oscillator circuit by a divider circuit, or more likely the 910 kHz signal
would be derived from the 455 kHz oscillator by using a frequency doubler
circuit.

Why would we want to go to all this extra trouble, we still need two balanced
modulators? It doesn't save us anything, indeed we have added an additional
circuit, the frequency doubler, and all we have gained is a 180 degree phase
ambiguity.

Did you have a different "post filtering" technique in mind?

I'm thinking each half of the primary in series with the plates, like
your typical PP amp output stage, which means the plate signals can't
be AC bypassed to ground in the single ended audio take off or else
the transformer feed is screwed. Might could put a second transformer
in there for the audio but I was thinking simple R isolation, between
the plates and filter, might be sufficient. I'd duplicate it on both
sides, even though one is 'unused' for audio, to keep plate balance.


I think you're still 90 degrees but that just means a lower single
ended amplitude, doesn't it? Which I'm thinking is a lesser problem
than the other.

If it is at 90 degrees, which is what the PLL would be driving towards,
the
output would be zip. Say it isn't quite 90, the problem is that the
phase
angle
changes the amplitude of the recovered modulation, but the uncorrelated
noise on
the received signal remains constant, so the SNR of the recovered audio
goes
to
pot when the phase angle differs from zero.

I think the question is output of what? There's 2 in my wacky idea: a
PP transformer with the other being a single ended audio take off.

Output of the demodulated audio.

Are you on the right schematic now?

I'm on page 9, is that the right one?

At 90 degrees the PP transformer's single secondary should average
(filtered) to 0, which is right for 0 phase error. Phase error to
either side should produce a DC (after filtered) error in that
polarity to drive the VCO.

I still don't understand how the DC is produced, through a transformer?

See above.

That doesn't help without a more detailed description of the "post filtering"
that you are proposing.

At 90 degrees we have equal but opposite audio and carrier on each
plate (which is why it nulls in the PP tranny) but the audio takeoff
is single ended, which isn't 'null'. I'm guessing it's half, 6 dB
down, from what you'd have in the 'ideal case' (assuming all else was
equal) of creating a quadrature Lo signal into a second mixer but it's
a heck of a lot simpler and maybe 6 dB isn't too horrible a sacrifice.

The wanted audio, although not the uncorrelated noise, is attenuated to zero
at
90 degrees, not just 6 dB, even from a single ended output. Remember we
aren't
talking SSB here, where the phase of the reinserted carrier doesn't affect
the
amplitude of the demodulated audio, we are talking DSB where carrier phase
is
critical.

We aren't inserting a carrier.

That's just a matter of semantics, what I was referring to was the local carrier
used to demodulate the signal by whatever means. This could actually be a large
amplitude carrier signal added to the IF signal, or reinserted, and feed to a
simple diode detector.

Anyway, when I ran simulations I did get single ended audio with 90
degrees Lo-carrier, or at least close. I can't say much else about it
because Circuitmaker keeps going loco and I'm not getting consistent
results so something is 'wrong'. Maybe I was getting a rectification
effect.

Hmm, makes one wonder how quadrature modulation systems, AM stereo, digital
Television over cable, and others, manage(d) to work?

I'm also not convinced my circuit is representative of the sheet beam
tube.

I don't think the details of the sheet beam model matter to this discussion, as
long as it sort of acts like a multiplier.

Speaking of "sheet beam" tubes, isn't this essentially the tube version of the
Gilbert Cell that Patrick is looking for?

So, maybe it's a great idea with the minor inconvenience of not
working

That's possible, although not yet proven.

An alternate might be to use a dual primary, dual secondary,
transformer for the audio and flip the phase on one secondary so it
adds.

That's fine for the audio, but it still doesn't explain how the control
signal
is derived.

Did I manage to sensibly convey the idea?

Yes, except for the part about how the DC control signal gets through the
transformer.

I didn't say put DC 'though' the transformer. You have 2 phased 455kHz
signals going into it and I said DC came from post filtering.

Yes, but I am waiting for an explanation of how the post filtering works?

After looking at it again I think you'd need to rectify to get a
second order filter.

Or more specifically feed it through another phase detector of some sort with a
2Fc reference.

Even ignoring that, I don't think the idea will work because it
violates what we learned in High School Trigonometry class. I don't believe
that the system can be made to work without using two separate ³modulators²
or
³phase detectors², one operating at 90 degrees to provide the control signal
and
a second operating in phase to recover the audio.

I dunno. My high school trig class didn't talk much about single ended
sheet beam tubes.

Of course not, neither did mine, just trigonometric equations and related
theory. AM DSB modulation is an example of such an equation.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 28, 4:08*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:

Anyway, I tried to look at the schematic of the PLL mentioned on page
461 within the .pdf for the 37 tube colour TV at the site at
http://antiqueradio.org/sitemap.htm
But the pdf was largely un-readable, having been scanned by an
incompetent. And so I found far too much time would be needed to de-
cypher the schematic to see just what flipper was talking about. I
ain't no fukkin *expert on tubed color TV sets from 1950s, and I'll
stay that way with very scant information, and posted so you can't
read what you're lookin at.

I guess flipper or JB will NEVER provide a copied version of the tubed
PLL schematic with full ammended details to allow it to be used in a
synchrodyne radio. Its WAAAAAAAAAAAAAAAAY beyond their willingness to
give, while they remain extremely willing to shoot everyone else down
who challenges them in in any way.

How am I trying to shoot your PLL efforts down? I'm trying to encourage you to
experiment with a PLL based radio, and your latest post gives me hope that
perhaps you are giving the matter some serious consideration. *I am trying to
give you a leg up by filling in a few details that you seem to be unaware of,
you are free to use, or not use, this information in any way you are comfortable
with.

Just exactly what details are you giving anyone here on the group?

Where is the schematic, preliminary test results?

[Snip]

Now a phase detector **could be** be fairly easily be built using a
suitable 1:1 RF transformer with secondary having a CT, and two
diodes, plus some R&C bits, all very much like an FM radio ratio
detector which is arranged to produce a Vdc output which is applicable
to a reactance triode, 1/2 12AT7, which then controls the FM set RF
oscillator F at over 100MHz.

The VDC generated by the phase detector goes +ve for where VFO F goes
higher than carrier, -ve if VFO F goes lower, and we might *HOPE* that
the Vdc change is enough to lock a VFO and that the time constant is
long enough with Vdc LPF network that VFO will run on without
significant F change if the carrier signal drops to nothing during
100% modulation.

The ³LPF² is called the ³Loop Filter², the design of this filter is something
you should read up on. *I'm not a ³PLL² expert but I think you want a double
time constant filter here, this typically involves two resistors and two
capacitors. *This consists of a short time constant filter to remove the carrier
frequency components and other grunge from the output of the phase detector. *
This is followed by a long time constant which serves as a modified integrator
to allow the VCO to track through the periods of carrier lock. *The long time
constant part of the filter is modified so that while it integrates the phase
error signal over the long term, it allows some higher frequency components to
pass through to provide faster locking and better loop stability, remember this
is a negative feedback loop so there may be stability issues. *Hopefully there
is a fully fledged PLL expert reading this who can better explain this.

I have not noticed any experts here. None. Zip. Zero. Lotsa wannabes.
Lotsa people suggesting I go build something, while they sit back and
watch.

I'm too busy re-engineering people's gear to re-invent a wheel with
many more wooden spokes than the wheel I already about.

The only reason I see for use of PLL techniques for a synchrodyne are
to get over the problem of intermittent locking of a locked oscillator
even with a limiter stage present.

That sounds about right, we need a second order loop to ride us through the
periods of carrier loss, or even effective phase reversal as may happen with
heavy modulation during selective fading.

I don't care about short wave listening to signals that fade and
amateur radio and so on, I only care about strong local BC band hi-fi
performance. Seems to me there's nothing wrong with envelope detection
with diode + R&C when its done my way with a simple pair of CF. 1 x
12AU7, and that ain't ever going to offend Bean Kounter. The only
exception is when the BC station is at low levels. So "exalted
carrier" is all that's needed, ie, add a 455kHz oscillator signal to
whatever recovered IF signal is there, and effectively reduce the %
modulation of everything detected by the diode + R&C, and thus there's
little distortion to the audio detected if the current discharge from
C is kept nearly constant, thus keeping ripple voltage constant, thus
keeping AF wave close to envelope shape.

So a 455khz oscillator would be good if locked by PLL a highly limited
version of any IF signal present. The limiter could be a pair of 6AU6.
There would be a phase detector tranny, 2 diodes plus some sort of
veractor tuning of oscilator plus a slow time constant filter are
needed. The oscillator can be another 12AU7.
At least 3 tubes are needed.

But Tucker's radio has its VFO tunable for all of the BC band. This
requires a large change in C between say 400pF and 30pF. Let us assume
we might find that a limiter stage using a pentode would help to give
a better signal to use to lock the VFO, and if you have a tuned VFO
with a tuning gang and you have a parallel tuning diode, then getting
the C change needed by a Vdc applied might be difficult to get working
right, so perhaps all the tuning might be done using a pot and varied
DC and nothing but veractors.

Yes, there are certainly problems here, it's not immediately obvious how to
connect a varactor to the oscillator section of a standard multi gang tuning
condenser while providing both the needed capacity change at the low end of the
band, and also not unduly increasing the minimum capacity at the high end of the
band, that is required to provide the 9:1 capacity change needed to tune the
entire MW band.

Some tuning diodes give large variations in C. If such diodes are not
used, a mechanical C is used and must have a parallel tuning diode
strapped across it for synchro. The tuning diode must be arrange to
work with its low C value to alter the high end of the band and with
its high C value at the low end.
Don't ask me how, I ain't done it.

But its easy to swing an IF frequency each side of just one F.


This 9:1 impedance change across the band will also affect the gain of the PLL
circuit, influencing its performance including lock range and stability.

It's not clear to me how much substituting varactors for a mechanical tuning
gang simplifies the problem, it presumably would if we were willing to use
opamps and piecewise linear summing circuits in the DC circuits controlling the
varactors. *Maybe there is a simple solution that I am just not seeing.

Most expeditions into chip driven circuits involve simple ideas
achieved with hundreds of devices hidden in the chip. Hence we see a
typical AM/FM radio tuner PCB with so many parts and nobody really
knows what's going on. The chip count is larger than the tube count.
Most people look into a an AM/FM tube tuner and its all baffling. Most
old guys in amateuer radio hate FM, and exclaim, "Damn it man, we just
want F to remain constant." None I ever talked to knew about mutiplex
stereo decoding, and very few knew about TV sets.
But they knew how to turn on the set and tell everyone about their
latest medical condition. I live in a world where most details and
principles behind how most things work are unknown.

I searched for a simple explanation about how digital radio works,
with all waveforms and schematics. No way, its secret stuff. Don't
even try to change anything on the board with several layers and such
small chips with so many pins you can't count them.

In any event there is something very satisfying about a mechanical tuning gang
in a tube radio.

Indeed.

I cannot list all forseable bothers trying to do all this.

Yes, while the direct conversion, or ³Zero-IF², approach has a strongly
seductive allure, its realization does raise a multitude of bothers. I'm not
sure how common direct conversion radios are outside of the ³Ham² radio
community, does anyone know? *I suspect that most of the so called ³Zero-IF²
receivers are actually superhetrodynes, where what at one time would have been
double conversion superheterodynes have had their second IFs replaced with a
³Zero-IF² and detector.

All these bothers are why I like my crawl before walking superhetrodyne with PLL
synchronous detector approach, at least for a first project.

I agree that's the best approach, but you maybe need 3 tube sockets
and a fair bit of skill. Tuckers double balanced ring diode de-
modulator looks simple until you try to make one. Quite a bit about
that on the web now and considerable stuff in british amateur radio
book, RSGB 5th edition, which has stuff with tubes and SS and chips.
Not one single working PLL with tubes for BC band.

But if it can be done for where VFO must make F between 530kHz and
1,710kHz, then its better to do it than retain the superhet mixer
stage, which we might want to chuck out BECAUSE we wish to use a
synchrodyne instead, and gain selectivity offered by the audio LPF
rather than by the IFT selectivity.

Yes, the direct conversion approach is very seductive, no question. *With the
superheterodyne approach, if you make the IF very wide, perhaps using a 2 MHz
IF, then you can control the audio bandwidth with the audio LPF just as with the
direct conversion approach.

And 2MHz IF does mean wide audio BW, but perhaps an extra IF stage, ie
6 tuned circuits to get skirt selectivity are needed. Exalted carrier
at 2MHz with PLL control and envelope detection would work to allow
the extra AF BW without slew.

I have a schematic of a Miller "Hi-fidelity" TRF radio tuner using 4
tuned circuits with a 4 gang cap.
http://www.indianaradios.com/Miller%...AM%20Tuner.htm
There are 4 RF coils, one per can, 200uH each, arranged in 2 pairs.
with bottoms of the two coils not grounded but are connected to each
other via a 12uH coil with CT which is taken to 0.05uF and the AGC
voltage circuit. I assume the slight mutual coupling improves the
selectivity curves for the claimed 7.5kHz to 10kHz BW while getting
enough skirt selectivity for closest locals which here would be 45kHz
apart.

Many ppl, myself included will have two identical 2 gang caps to make
4 gangs. They may be fairly easily made to have the same sized tuning
wheels and the one dial cord will swing them all the same.


So if people want to know how I might get better hi-fi from a tubed
synchrodyne set which I would want to work better than Tucker's 1947
job, then they might see agreement about using a following summary of
stages :-

TRF input stage, ferrite rod antenna input coil, AVC applied to grid
of V1 pentode.
One tuning gang, 30pF-400pF.
RFT at anode, secondary tuned with second tuning gang, 30pF-400pF.
*V2 CF buffer stage.
Q of LC should not be too high to caused side band cutting and RF BW
should be at least 8kHz at the low end of band.

V2 buffer drives balanced modulator using Tucker's circuit with
transformers using details to be worked out.

V2 buffer also drives V3 pentode limiter to make near constant
amplitude output signal.
V4 CF buffer used after V3 to drive phase detector to produce Vdc to
change veractor C of oscillator.
LPF for dc path keeps Vdc constant for limiter signal drop outs when
AM % goes to 100%.

V5 triode CF oscillator with tunable tank using the 3rd available
tuning gang.
V6 triode oscillator output buffer used to drive VFO input to balanced
demodulator.
Veractor used in parallel with tuning gang.
Application of Vdc from phase detector to ensure VFO stays locked to
limiter carrier F.
Tuning of oscillator should track RF stage tuning

The AF from balanced de-mod stage buffered by V7, and active second
order RC + V8 CF LPF filter used before V9, V10 audio output stages.

It obviously might be easier to use a PLL and oscillator which worked
at 455kHz only, to avoid having to tune the oscillator over a wide
range.

Then the RF input stage might be simplified.

There are probably 101 other ways which might be utilised, but the
synchrodyne will probably end up more complex than the standard 2 tube
superhet AM tuner used in most AM radios for strong local stations.

That's a very good start!

Using a two tube superhetrodyne as a base line is bean counter thinking. *We
want the best possible High Fidelity reception under as many conditions as
possible, this is only possible using a true synchronous detector, pretenders
need not apply

The two tube superhet works well for most ppl for locals if they don't
mind 3kHz AF BW. But it can be made to work a lot better with a
switched tertiary coil on IFT1 to give twin peaked 455kHz which then
widens overall AF BW to 8kHz. Details are in RDH4, and such things
were not uncommon on AM tuners of 1950s and 60s and you'd tune with
the tuning eye on lo-fi, then press the hi-fi button to switch in the
tertiary and BW went to 8kHz. A dedicated tone control could increase
it. The tuner was cheap with just 6BE6, 6BA6. I've done a few samples
with the mod, not bad, and if fitted with the CF detector and ungraded
following audio stages, and with ferrite antenna to minimise hum when
tuned to stations affected by compact fluorescent lamps.


I do not know if the synchrodyne balanced demodulator can produce a
Vdc which may be used for VFO F control in addition to producing audio
output.

No, you need separate detectors for the ³PLL² and to recover the audio. *The two
detectors must be provided with carrier signals from the oscillator 90 degrees
out of phase with each other. *Maintaining the required 90 degree phase shift
over a 3:1 frequency range is another big potential gotcha for a tube based
direct conversion receiver. *In a solid state design this problem would be
simply dealt with by operating the oscillator at 4 or 8 times the received
frequency, and using digital dividers to generate the 90 degree phasing between
the two local oscillator signals. *A large number of tubes would probably be
required to do it this way in a tube based circuit.

You may also want to add a lock detector circuit, and an envelope detector to
your list above. *The lock detector could switch the parameters of the ³Loop
Filter² when out of lock to facilitate locking, and also switch to an envelope
detector during tuning.

But one thing is for sure, to make a synchrodyne with tubes to meet
modern use expectations is a real big ask.

Yes, it's a very big ask!

The TRF looks a lot easier. There's always enough signal on strong
locals. No real need for "exalted" carrier injection and envelope
detection, or syncro-anything.

Patrick Turner.

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/- Hide quoted text -

- Show quoted text -

[John Byrns]

In article ,
Patrick Turner wrote:

On Jul 28, 4:08*am, John Byrns wrote:

That sounds about right, we need a second order loop to ride us through the
periods of carrier loss, or even effective phase reversal as may happen
with
heavy modulation during selective fading.

I don't care about short wave listening to signals that fade and
amateur radio and so on, I only care about strong local BC band hi-fi
performance.

At least here in the US, the same sort of distortion from greater than 100%
negative modulation, or skewed sidebands, occurs on local stations if you happen
to be unlucky enough to be in one of the nulls of a directional broadcasting
antenna.

Seems to me there's nothing wrong with envelope detection
with diode + R&C when its done my way with a simple pair of CF. 1 x
12AU7, and that ain't ever going to offend Bean Kounter. The only
exception is when the BC station is at low levels. So "exalted
carrier" is all that's needed, ie, add a 455kHz oscillator signal to
whatever recovered IF signal is there, and effectively reduce the %
modulation of everything detected by the diode + R&C, and thus there's
little distortion to the audio detected if the current discharge from
C is kept nearly constant, thus keeping ripple voltage constant, thus
keeping AF wave close to envelope shape.

Sounds good, but you are describing a syncrodyne detector.

So a 455khz oscillator would be good if locked by PLL a highly limited
version of any IF signal present. The limiter could be a pair of 6AU6.
There would be a phase detector tranny, 2 diodes plus some sort of
veractor tuning of oscilator plus a slow time constant filter are
needed. The oscillator can be another 12AU7.
At least 3 tubes are needed.

And presto you have just built the synchronous detector that say isn't needed.

I have a schematic of a Miller "Hi-fidelity" TRF radio tuner using 4
tuned circuits with a 4 gang cap.
http://www.indianaradios.com/Miller%...AM%20Tuner.htm

I'll raise you two, I have a schematic of a Western Electric TRF radio tuner
using 6 tuned circuits with a 6 gang cap.
http://louise.hallikainen.org/~harol...oads/We10a.pdf
and
http://www.radiomuseum.org/r/western_el_10_a10.html

There are 4 RF coils, one per can, 200uH each, arranged in 2 pairs.
with bottoms of the two coils not grounded but are connected to each
other via a 12uH coil with CT which is taken to 0.05uF and the AGC
voltage circuit. I assume the slight mutual coupling improves the
selectivity curves for the claimed 7.5kHz to 10kHz BW while getting
enough skirt selectivity for closest locals which here would be 45kHz
apart.

I have one of these J.W.Miller TRF tuners. To maintain constant bandwidth from
the bottom to the top of the MW broadcast band the circuit Q of the coils must
increase in proportion to the frequency, this can be accomplished by arranging
the circuit so that series resistance, either part of the coil, or external,
dominates the in circuit Q of the coil. Then when coupled into band-pass pairs
the product of the coupling factor k, and the Q, or kQ, must be kept constant
across the MW band. Since the Q is increasing with frequency the coupling
coefficient k must be inversely proportional to frequency. This is done by
keeping the mutual coupling reactance constant across the band. Unfortunately
ordinary capacitors and inductors have reactance which varies with frequency.
This problem is solved by using a normal capacitor in series with an inverse
capacitor who's reactance varies proportionally to frequency, the two
"capacitors" approximating a constant capacitive reactance across the MW band.
The inverse capacitor is simulated by using the so called "negative mutual
coupling coil", the center tapped coil that you describe which acts as the
inverse capacitor in the mutual coupling circuit.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 30, 2:28*pm, John Byrns wrote:
In article ,
*Patrick Turner wrote:

On Jul 28, 4:08*am, John Byrns wrote:

That sounds about right, we need a second order loop to ride us through the
periods of carrier loss, or even effective phase reversal as may happen
with
heavy modulation during selective fading.

I don't care about short wave listening to signals that fade and
amateur radio and so on, I only care about strong local BC band hi-fi
performance.

At least here in the US, the same sort of distortion from greater than 100%
negative modulation, or skewed sidebands, occurs on local stations if you happen
to be unlucky enough to be in one of the nulls of a directional broadcasting
antenna.

Seems to me there's nothing wrong with envelope detection
with diode + R&C when its done my way with a simple pair of CF. 1 x
12AU7, and that ain't ever going to offend Bean Kounter. The only
exception is when the BC station is at low levels. So "exalted
carrier" is all that's needed, ie, add a 455kHz oscillator signal to
whatever recovered IF signal is there, and effectively reduce the %
modulation of everything detected by the diode + R&C, and thus there's
little distortion to the audio detected if the current discharge from
C is kept nearly constant, thus keeping ripple voltage constant, thus
keeping AF wave close to envelope shape.

Sounds good, but you are describing a syncrodyne detector.

So a 455khz oscillator would be good if locked by PLL a highly limited
version of any IF signal present. The limiter could be a pair of 6AU6.
There would be a phase detector tranny, 2 diodes plus some sort of
veractor tuning of oscilator plus a slow time constant filter are
needed. The oscillator can be another 12AU7.
At least 3 tubes are needed.

And presto you have just built the synchronous detector that say isn't needed.

I have a schematic of a Miller "Hi-fidelity" TRF radio tuner using 4
tuned circuits with a 4 gang cap.
http://www.indianaradios.com/Miller%...ty%20AM%20Tune...

I'll raise you two, I have a schematic of a Western Electric TRF radio tuner
using 6 tuned circuits with a 6 gang cap.http://louise.hallikainen.org/~harol...oads/We10a.pdf
andhttp://www.radiomuseum.org/r/western_el_10_a10.html

There are 4 RF coils, one per can, 200uH each, arranged in 2 pairs.
with bottoms of the two coils not grounded but are connected to each
other via a 12uH coil with CT which is taken to 0.05uF and the AGC
voltage circuit. I assume the slight mutual coupling improves the
selectivity curves for the claimed 7.5kHz to 10kHz BW while getting
enough skirt selectivity for closest locals which here would be 45kHz
apart.

I have one of these J.W.Miller TRF tuners. *To maintain constant bandwidth from
the bottom to the top of the MW broadcast band the circuit Q of the coils must
increase in proportion to the frequency, this can be accomplished by arranging
the circuit so that series resistance, either part of the coil, or external,
dominates the in circuit Q of the coil. *Then when coupled into band-pass pairs
the product of the coupling factor k, and the Q, or kQ, must be kept constant
across the MW band. *Since the Q is increasing with frequency the coupling
coefficient k must be inversely proportional to frequency. *This is done by
keeping the mutual coupling reactance constant across the band. *Unfortunately
ordinary capacitors and inductors have reactance which varies with frequency. *
This problem is solved by using a normal capacitor in series with an inverse
capacitor who's reactance varies proportionally to frequency, the two
"capacitors" approximating a constant capacitive reactance across the MW band. *
The inverse capacitor is simulated by using the so called "negative mutual
coupling coil", the center tapped coil that you describe which acts as the
inverse capacitor in the mutual coupling circuit.

I like the Miller tuner more than I like the other WE sets.

But what you said in your concluding paragraph has an abundance of
information which leaves the obscurity around the subject described at
least highly persistant.

I found the Q got higher at the low end of the BC band so in one radio
I used two input coils and two gangs on one set so that each coil
tuned each side of a centre F at the low end, and both together at the
top end where whatever the Q was, it wasn't enough to cut sidebands
much for the wanted 10kHz of AF BW. The two coils had their own copper
can and were bits of ferrite rod and solid wire. The first had a
loosely coupled input primary winding from a bittowire antenna, with
the tuned winding coupling to the second tuned LC via a carbon comp R
= 39k. Seemed to work just fine. But fluorescent lamps ruined
reception and I changed to one hand made coil on a single ferrite rod.
Audio HF remained OK and hum ****ed off, so I is 'appy wiff wotteye
got now.

Patrick Turner.



--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/

[John Byrns]

In article ,
Patrick Turner wrote:

On Jul 30, 2:28*pm, John Byrns wrote:

I have one of these J.W.Miller TRF tuners. *To maintain constant bandwidth
from
the bottom to the top of the MW broadcast band the circuit Q of the coils
must
increase in proportion to the frequency, this can be accomplished by
arranging
the circuit so that series resistance, either part of the coil, or
external,
dominates the in circuit Q of the coil. *Then when coupled into band-pass
pairs
the product of the coupling factor k, and the Q, or kQ, must be kept
constant
across the MW band. *Since the Q is increasing with frequency the coupling
coefficient k must be inversely proportional to frequency. *This is done by
keeping the mutual coupling reactance constant across the band.
*Unfortunately
ordinary capacitors and inductors have reactance which varies with
frequency. *
This problem is solved by using a normal capacitor in series with an
inverse
capacitor who's reactance varies proportionally to frequency, the two
"capacitors" approximating a constant capacitive reactance across the MW
band. *
The inverse capacitor is simulated by using the so called "negative mutual
coupling coil", the center tapped coil that you describe which acts as the
inverse capacitor in the mutual coupling circuit.

I like the Miller tuner more than I like the other WE sets.

But what you said in your concluding paragraph has an abundance of
information which leaves the obscurity around the subject described at
least highly persistant.

I found the Q got higher at the low end of the BC band so in one radio
I used two input coils and two gangs on one set so that each coil
tuned each side of a centre F at the low end, and both together at the
top end where whatever the Q was, it wasn't enough to cut sidebands
much for the wanted 10kHz of AF BW. The two coils had their own copper
can and were bits of ferrite rod and solid wire. The first had a
loosely coupled input primary winding from a bittowire antenna, with
the tuned winding coupling to the second tuned LC via a carbon comp R
= 39k. Seemed to work just fine. But fluorescent lamps ruined
reception and I changed to one hand made coil on a single ferrite rod.
Audio HF remained OK and hum ****ed off, so I is 'appy wiff wotteye
got now.

I think you need to use Litz wire instead of solid wire to wind the coils, like
the crystal set guys do, to get the required Q at the high end of the band.
Then add a small series resistor to each coil to reduce the Q at the low end of
the band; this series resistance has little effect at the high end of the band
when variable capacitor tuning is used. Then you need to do some reading on
filter theory to find out the proper way to couple the two coils. This approach
will provide a flatter response, with less finicky alignment and lower losses
than your ad hoc approach.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/