[ARC5] Converter/front end for Q5-er Sets

Sun Jan 11 12:10:32 EST 2015

Hello Bruce, Your letter raises a good number of useful points. To get
the full benefit of your observations I need to deal with them in detail
one at a time. I'll comment briefly on each point, and then return to
each point, one at a time. When I say 'return' I mean 'in the future',
because it will take some time to make the changes you suggest. I
imagine it might take me 2 - months before I deal with each item.

(1) # of resonant circuits before the mixer. You expect poor image
response (or maybe 'good' image response, because I'll get many!) There
is no argument on this point. I have two comments. Correcting this
problem is easy in concept, but I'm going to leave this work to last,
because it's the easiest (I think). Also, I can put up with the odd
birdie/image in the short term. If McCoy is right the images should be
'down' about 40dB. That's not good!

(2) The pot (in the mixer circuit) that goes no-where. You're quite
correct. In the original design I intended to use a varactor (or maybe
a reverse biased zener operating below the zener knee voltage.)
However, having ordered 20 varactors I grew impatient (from the Latin,
'to suffer') and used a common variable capacitor, but didn't re-draw
the circuit.

(3) Complexity of the L.O. I understand your argument about complexity.
Certainly I don't need a cascode stage as a buffer. On the other hand,
transistors are virtually 'free' and the source-coupled oscillator is
very simple; in some ways it's more simple than the Colpitts, despite
using more transistors.

In the past I tried the more common Colpitts, and this circuit does NOT
give an "nice" sinosoidal wave-form. That's the point of my 'more
complex' circuit. I want to feed the mixer with a sine-wave. It's an
experiment based on a hunch. It's a mathematical hunch, based on what
some people call the 'trig frog'. Sin(A) * Cos(B) = 1/2sin(A+B) +
1/2sin(A-B). The "A+B" and "A-B" is a mathematical expression of the
action of mixing. Since a sine wave has no harmonics, and harmonics
contribute to spurious responses I WANT THEM GONE! That's why my circuit
is complicated. I can get a 'nice' sine from this circuit if I put the
correct number of turns on the link coil.

(3b) As an aside, I can get a 'nice' sine from your Colpitts, using the
same trick. Maybe I'll do that, if I get this circuit working to my
satisfaction. My goal is a 'nice' signal for the mixer, not the simplest
circuit possible.

(4)As for your comment about the cascode - I can't comment. I've never
   used this circuit before, and certainly never in the heat of battle.
   However, I think the cascode mixer provides too much 'front-end' gain
   when used before a 'nav/range' receiver. The GREAT thing about my
   set-up here is that I used a plug-in module for the mixer, so I can
   build a converter board using a ring-diode and compare the two.

(5)XTAL-locked or tune-able L.O? Should I use a crystal oscillator,
   instead of the V.F.O.? Sometimes, I think. I used a 3200kHz crystal
   in an earlier version of my experiment. At the moment I'm 'down' on
   the A.M. B'cast band, where stability is easily got. When I lay out
   the art-work for this card I intend to put down tracks for both a
   variable AND xtal-locked oscillator. I think there are times when the
   VFO circuit has an advantage. For example, a lot of people have a
   'command' range receiver 'on the shelf' but don't have a tuning shaft
   (with spline). I'm lucky enough to have inherited a tuning knob.

(6)Double-balanced mixer. You touched on this subject in your point
   about the cascode mixer. If/when I get this circuit going I'm going
   to try a DBM, but note that I can't see ONE SBL-1 on eBay for less
   than about $15.oo I'll probably be winding my own transformers, and
   matching diodes with a multimeter. (I understand I can/should use HP
   'hot carrier' diodes, but there's nothing teaches like rolling my
   own.) I'll bet a dollar to a con-man's wife I can't get the same flat
   response I see on the SBL-1 spec sheet!

(7)Since my original posting, I mounted all the components from the rats
   nest around the mixer tank circuit on a solid chassis. I used a
   length of aluminium channel, 100mm wide (4 inches for all you
   recalcitrant imperial types) x 2 mm thick. Waddya-know? Suddenly the
   instability has gone. I used -61 toroid for the coil (124uH) and I
   haven't had a chance to observe drift, but there WILL be some. I left
   the rig running from 8PM to 4AM (that's the time now), and the drift
   is less than in the IF pass-band. That's encouraging. Anyway, we all
   know good sets need to be built like a battleship, and this
   experience confirms that point. (Thinks: - I have a length of
   aluminium channel that's 3mm thick. THAT would make a solid chassis!)

I think I've dealt with each point you made in the list above. Over the
next few weeks I'll work thru the list, one point at a time. At the
moment, with the newly wound L.O. coil, I am feeding too much signal
into the buffer. The output is badly distorted. I'll work on that until
I have a 'nice' sine wave feeding the mixer.

I repeat - the REALLY good thing about this circuit is that I can re-use
the chassis and tuning arrangement with ANY converter, provided I keep
the same pin connection to the 'base' card. Would you like to see a
photo of the construction?

73 de Les Smith vk2bcu at operamail.com

On Sun, Jan 11, 2015, at 06:03, Bruce Long wrote:
> Thanks for sending a copy of the schematic No I am not a member so I
> don't have access to the ARC5 group
>
> I am not overly impressed with the converter circuit Here are some
> things that caught my eye The mixer stage has only one RF resonant
> circuit so I would expect image rejection to be poor.
>
> Also the pot associated with the rf resonant circuit has no dc
> path through the wiper perhaps a rf varactor was removed or needs
> to be added.
>
> The LO in my opinion is overly complicated.
>
> I like differential pairs ( transistors) but I don't like doing diff
> pairs with discrete fets because the spread of gate threshold voltages
> is too wide to make be sure the fet diff pairs will always work and
> any two random fets
>
> I also have a dislike of the two transistor cascade circuit because
> early in my career as a production test technician I had endless
> problems with - admittedly- poorly designed cascode pairs
>
> The advantage of series cascode pairs is that they can provide
> excellent reverse- output to input isolation. This powerful advantage
> is diminished by the fact the upper- common base transistor is
> susceptible to vhf parasitic oscillations unless you are very careful
> with the base bypass capacitor. You can put a ferrite bead or small
> value resistor in the base lead but this degrades- I think- some of
> the reverse isolation advantage.
>
> I like using the plain old fashion Colpitts oscillator taking the
> output from the low impedance emitter or source node. The waveform
> here is non-sinusodal so some filtering is needed but you get quite
> good interstage reverse isolation and therefore resistance against osc
> pulling with any buffer amplifier that has even a modestly high Zin.
> There is also an esoteric phase noise advantage of taking the signal
> off of the oscillator emitter-source
>
>
>
> Another problem is you cannot set the Collector Emitter current
> independently which does not matter in the least in some application
> but is a real problem in others.
>
> I have also had great success taking the osc output from the cold end
> of the Colpitts inductor or Crystal. Instead of returning these
> components to ground you connect them to the input of a common base or
> common gate stage biased with enough current to had a rather low Zin.
> This has a really good phase noise advantage as well as the osc
> crystal acts as a narrow band filter thereby giving excellent far out
> phase noise levels. But again you need good pcb design practice and
> some other safeguards to brevent the common base-or common gate
> amplifier from going into parasitic vhf oscillaton
>
>
> With respect to mixers, I need a really really good reason not to use
> a DBM. Here in the states you can get a DBM from minicircuits for
> under $3 USD if I remember correctly. I am interested in some the
> newer Mode- H mixers that use fast bus driver chips but IIRC these
> devices are not available in DIP packages which makes their use in
> home brew projects, especially beginner home brew projects
> problematic.
>
>
> good luck with your project You comments on thsi subject have brought
> back to me how much I enjoyed the Q5er- converter combination given to
> be so long ago.
>
> bruce
>
>
>
>
>
>

> From: Leslie Smith <vk2bcu at operamail.com> To: Bruce Long
> <coolbrucelong at yahoo.com> Sent: Friday, January 9, 2015 9:11 PM
> Subject: Re: [ARC5] Converter/front end for Q5-er Sets
>
> Hello Bruce, I send a copy of the circuit of interest. Can you log
> into the page www.groups.yahoo.com/arc-5Radio? I think you need to be
> a member of the group before you can "see" the files?
>
> At any rate, I was able to get my circuit 'back' onto my computer. I
> used IRFANview to open it.
>
> Now to the points you raised. I think you make an excellent point
> about 'fixed tuned', and from my perspective that's a useful addition
> but given the additional cost of laying out an additional 2
> transistors for a xtal oscillator/buffer I'd just put in the tracks on
> the PC card and let a builder make that decision. I need 5 volts (p-p)
> of oscillator signal to drive the mixer - that's all. PC card area and
> transistors are cheap!
>
> Eventually I WILL use a varactor, rather than a variable "C". For the
> moment I have the variable "C"s and I'm using them. (I bought x20
> varactors on eBay, 400pF range. I'm going to experiment with 12 volts
> zeners, reverse biased as tuning caps.) I think the reverse biased
> zeners have a capacity range of about 70 to 140pF, but I have have no
> idea of "Q".
>
> I like the idea of 'common' crystals (or common crystal frequencies),
> and I'll work to that idea when I take the local oscillator. I'm
> working one stage at a time. Thanks for your ideas!
>
>
> 73 de Les Smith vk2bcu at operamail.com
>
>
>
>

> On Sat, Jan 10, 2015, at 11:37, Bruce Long wrote:
>> Leslie When i was a boy someone took pity on a poor novice and gave
>> me a R23 with an attached xtal controlled one tube 40 m converter.
>>
>> It worked great? With the IF transformer coupling rods pulled all the
>> way out the selectivity was quite good, Almost as good as the HW-16 I
>> eventually was able to afford.
>>
>>
>>
>> If it was my project I would make the converter fixed tuned and tune
>> the ARC 5 receiver instead.
>>
>> I would use a two or three pole pre-selector type filter in front of
>> the converter mixer or maybe even multiple varactor tuned resonant
>> circuits.
>>
>> I was unable to get access to the schematic you referred to . If you
>> would email the file to me I'd be happy to comment.
>>
>> To get several fixed tuned bands you could build a simple phase
>> locked loop to lock in - say 100 or 200 kHz intervals This way the
>> ARC 5 receiver dial calibration is still useful
>>
>> Or you could look through the crystals available from Digikey,
>> Mouser, Farnel or whoever is available in NZ land, choose the off
>> ball crystal frequencies closest to the desired decimal integer LO
>> frequencies and then deal with the dial calibration in another way
>> such as supplemental pencil lines on the ARC 5 receiver spaced to the
>> left and or the right of the actual dial indicator line.
>>
>>
>>
>>
>>
>>

>> From: Leslie Smith <vk2bcu at operamail.com> To: ARC-5 List
>> <arc5 at mailman.qth.net>; Ron Scifleet <ron.scifleet at sanitarium.com.au>
>> Sent: Friday, January 9, 2015 5:20 PM Subject: [ARC5] Converter/front
>> end for Q5-er Sets
>>
>>
>>
>>
CNV-412, MF/HF converter for Q5-ers
>>
>>
For some time I wanted to build a general purpose converter to use in
>>
front of a 'command' 'navigation' receiver (e.g. BC-453/946 or R-23/4).
>>
The circuit of a prototype, still under development, is shown in the
>>
ARC-5Radio files, in a folder called Exptl_(LNS).
>>
>>
Find the circuit here:
>>
>>
File Name: "CNV-412, R12 (on 5-0110).png"
>>
Folder: Exptl_(LNS)
>>
Yahoo Group: ARC-5Radio
>>
Web Reference: www.groups.yahoo.com
>>
>>
The circuit is quite conventional, although the local oscillator uses
>>
the not-so-common source coupled oscillator. Some mistakenly call this
>>
a Franklin oscillator, and there are similarities, but the two are
>>
somewhat different.
>>
>>
Operational objective.
>>
To build a generic 'front end' to use with the Q5-er sets that will
>>
operate on the 160/80/40m band and also the AM broadcast band. The
>>
circuit will be a solid state implementation of the popular converter
>>
presented by Don Stone, but have the capability to use both a xtal
>>
controlled and variable local oscillator.
>>
>>
Construction. I built the prototype circuit on strip-board, using the
>>
more-modern 20 way edge connectors used in digital circuits. Using a
>>
plug-in type of construction shortens the develop time. The mixer and
>>
oscillator tank circuits are connected (at the moment) using alligator
>>
clips to the LC tank components. The AM broadcast band has a good
>>
mixture of strong local and weak distant stations, and the oscillator
>>
will be more stable at the lower frequency, so initial design focused on
>>
520 - 1650kHz.
>>
>>
It takes little imagination to see how quickly this could be moved to
>>
160m and after that to 80m.
>>
>>
Present functionality.
>>
The Q5-er used for this experiment is a not-so-badly hacked BC-453. The
>>
outer box and shield over the tuning gang were both gone; the 'common'
>>
mods to the utility box on the front of the set were apparent. The BFO
>>
circuit had been re-wired, but speaking generally the set was quite
>>
function, I thought. After routine troubleshooting I could hear the
>>
local 'nav' broadcast on 362kHz using a few meters of wire on the
>>
aerial.
>>
>>
I run my 'command' sets on a 60V B+ supply; in this particular set I
>>
replaced the 'flower-pot' caps with 100V mylar types, so the B+ absolute
>>
limit is 90V DC. I haven't used a 90V supply enough to say it would
>>
give better performance than a 60V supply, but it may.
>>
>>
To drive a loudspeaker I bought a surplus 110-11V power transformers,
>>
and put this 'in series' with the existing 600 ohm transformer. In
>>
other words, I feed the 600 ohm output to the 110V primary, and connect
>>
my speaker to the secondary. Using this with the 60V B+ supply my set
>>
give useful volume in a quiet room. (More volume might be got by
>>
reducing the value of the cathode resistor on the 12A6 output 'tube',
>>
but I haven't done that yet. I avoid modifying the set as much as
>>
possible.)
>>
>>
Using the converter, I hear broadcast stations. My idea of using the
>>
Q5-er at a fixed
>>
frequency went out the window, because the tuning was 'touchy'. That
>>
surprised me, as a 'lash-up' oscillator I built previously (but using a
>>
negative resitance oscillator) would stay zero-beat on the local ABC
>>
radio station for long periods of time (I mean hours at a time). The
>>
next stage of development will be around the tank circuits - fitting
>>
them to a metal chassis, with shielding.
>>
>>
I suspect the I.F. coupling transformer (at the 'top end' of the mixer)
>>
lacks impedance. In other words, I think I need more turns. I'm going
>>
to plot the response of this transformer, and rewind it if it is found
>>
to lack performance.
>>
>>
The signal from the local oscillator changes (in amplitude) a good deal
>>
over the frequency range. It's too low, even in mid-band. It should
>>
provide 5V p-p to the 'top' gate in the mixer, allowing class "C"
>>
operation. An earlier lash-up gave an output signal that varied no more
>>
than 10% mid-band to the low and high frequency limits. Clearly I must
>>
do more work to the local oscillator circuit.
>>
>>
I built the circuit so the L.O. signal came "off" the card, thru a link
>>
and "onto" the card to the mixer gate circuit. This allowed some
>>
flexibility - if I wanted to substitute a different local oscillator,
>>
all I had to do was open the link and insert the new L.O. signal. This
>>
idea works so well I'm going to use RCA connectors to switch the signal
>>
path in the link circuit.
>>
>>
Problems, deficiencies, bugs, work needed and so on.
>>
First, the stability and tuning rate of the L.O is unsatisfactory. To
>>
tune SSB I need a tuning rate of about 50kHz per revolution of the
>>
tuning knob. I have no idea about a good tuning rate for AM, but
>>
probably only 1/4 that needed for SSB. Perhaps someone here can tell
>>
me?
>>
>>
The mixer tank tuning is hopelessly 'wrong'. It doesn't peak
>>
adequately; it's broad, and I suspect I should reduce the tank inductor
>>
from the present 370uH to about 260 .. 300uH. However, the problem is
>>
more than simply wrong values. When I peak the tank circuit I hear a
>>
noise that (I suspect) comes from instability. Yes .. I think my superb
>>
design is oscillating. I suspect I could replace the cascode mixer with
>>
a straight single source coupled JFET mixer. I don't need the gain the
>>
cascode gives, at least not at the low end of the H.F. spectrum. It may
>>
be useful on 6m, but that wild-eyed idea (using a BC-455) is still a
>>
distant flicker of hope.
>>
>>
So - I need to debug or modify the mixer. Ideas about how this might be
>>
done are welcome.
>>
>>
I think I might get the circuit working to my satisfaction after two or
>>
three more months of work. At the moment my prototype suggests the idea
>>
is sound, but Don Stoner proved that decades ago, so that doesn't
>>
surprise me. Still, it's nice to start with a handful of components and
>>
a circuit and slowly, slowly see the 'thing' come to life.
>>
>>
In the end I want a front-end 'box' that can sit proudly alongside a
>>
'command' radio, knowing it is designed and built as well as the
>>
original set. The prototype might be strip-board and chewing gum, but
>>
the finished rig-owner (me, in this case) should be proud of the final
>>
article.
>>
>>
I want to say a few things about the drawing. First, it's clearly
>>
incomplete. I hope it's sufficiently accurate to not confuse anyone
>>
reading it. Next, I drew this using the 'freeware' program TinyCAD,
>>
running under Windows XP. The program isn't perfect, but I like to use
>>
'freeware' developed by an amateur coder. This is 'neat' software. If
>>
enough people use this software it will survive, and wouldn't that be
>>
nice? The TinyCAD package allows users (I'm a 'user' not a 'loser') to
>>
create libraries & symbols. I created a library (I called it 'common')
>>
and drew most of the symbols in the circuit you see. Useful, eh? As
>>
for the European symbol for the resistor you see - if you don't like it
>>
- it's easily changed, and the the more common 'zig-zag' resistor symbol
>>
IS in the library too.
>>
>>
TinyCAD from SourceForge. Get yours today. Don't leave home without
>>
it! Notice the size of the *.png file produced when I 'export' the
>>
drawing. 27kB. Neat eh?
>>
>>
Observations, suggestions and even hateful criticisms (or flames) about
>>
my converter (the fabulous and exotic - at times quixotic - CNV-412) are
>>
welcome!
>>
>>
Les
>>
formerly VK2BCU.
>>
>>
--
>> http://www.fastmail.com [1]- Send your email first class
>>
>>
______________________________________________________________
>>
ARC5 mailing list
>>
Home: http://mailman.qth.net/mailman/listinfo/arc5
>>
Help: http://mailman.qth.net/mmfaq.htm
>>
Post: mailto:ARC5 at mailman.qth.net
>>
>>
This list hosted by: http://www.qsl.net[2]
>>
Please help support this email list: http://www.qsl.net/donate.html
>>
>>

>
--
http://www.fastmail.com - Faster than the air-speed velocity of an
unladen european swallow
>
>

Links:

  1. http://www.fastmail.com/
  2. http://www.qsl.net/

-- 
http://www.fastmail.com - Same, same, but different...