[ARC5] Converter/front end for Q5-er Sets (long & boring)

[Leslie Smith]

Hello Neil, Yes, although the BC-453 is rather scrappy, it still looks
nice. I'll left a photo in the ARC-5radio 'group' files for a week or
so, along with the circuit. It's in a folder called 'Exptl_(LNS)

I bought it as scrap - for the tuning capacitor gang. At first
inspection it appeared to be a wreck. A good deal of the large (under
chassis) components were missing - including the output transformer. The
heaters had been re-wired to suit a 12 volt supply. One or two valves
were missing. The B.F.O. box was missing. BUT: I found I could buy a
replacement for the B.F.O. and I had a spare output transformer
(somewhere). Sitting in the socket, a new 12A6 looks very shiny.

After some work I had the set working, (and with the heaters re-wired to
a 24V supply.). I don't have the heart to trash it, especially with the
black shiny 12A6 in place. I'll build a new outer cover, and leave the
tuning capacitor w/out a cover.

The set was also used as an experiment in cleaning. Until now I'd read
about cleaning these sets with a hose, but I never had the heart to
'work over' a good set. With this one I had nothing to lose. I cleaned
mine with 'Dot's Bath & Tile Cleaner' - something in a can with
quaternary ammonium compounds. Imagine that - quaternary ammonium
compounds inside a 'command' chassis. I did this using a garden hose, on
the front lawn. There are some cleaner parts of the chassis where the
'Dots' ran out of the holes, and ran down the outside of the chassis.
(There are some 'shiny clean' streaks on the chassis!) Anyway, I have
60V B+ on the set, and it hasn't blown up. I suspect it would work with
180V B+ just as well. One thing about the 60V supply. The set runs
wonderfully cool!

Anyway, the logic of my 'adventure' has two parts - to get 'hands-on'
experience with these sets, and keep them working. It's my view that a
working 'knocked around' set is more valuable than a 'very nice' set
sitting in a box, protected from dust and the ravages of time. But
that's just an idea I have. .

Next part of the little adventure - get a converter with a new config
working with the BC-453 working on 160m.


73 de Les Smith vk2bcu at operamail.com


On Tue, Jan 13, 2015, at 15:41, AKLDGUY . wrote:
> Hi Leslie
>
> Hey, neat picture. BC-453 looks great.
>
> You're certainly going to a lot of trouble to check out various
> arrangements. Well done! I'm not sure I'd have that much patience.
>
> 73 de Neil
>
>
> On Tue, Jan 13, 2015 at 5:29 PM, Leslie Smith
> <vk2bcu at operamail.com> wrote:

>> Hello Neil, I took a photo of the experimental set-up I built for the
>> converter.
>>
>> I added some labels to the photo, although I imagine they aren't
>> really necessary. You can see the plug-in coil, on the front edge of
>> the chassis. The core is FT61-50. With this plug-in coil I can use
>> the experimental rig on different frequencies. The first will be the
>> broadcast band; after that I will try 160m. Then I'm off up to 80m.
>>
>> Your information, that the 100nF caps have a good deal of internal
>> inductance, came as a surprise. I always imagined the inductance
>> would be low, because they were bypass caps, used with digital logic.
>> I learned something. I'll use ceramic caps to decouple in future.
>> (Bonus! I save 1 cent on every cap I use! On a complete board I may
>> save $0.20).
>>
>> I have to build a second converter. I'll make the changes we
>> discussed in the next version - so I'll keep the Franklin oscillator,
>> use a different buffer and change the mixer to a source-fed JFET
>> mixer. I'll isolate the oscillator from the mixer with a source
>> follower. I think the arrangement will work well. After that - I'm
>> going to try a diode ring mixer.
>>
>> Thanks for your comments,
>>
>>
>>
>> 73 de Les Smith vk2bcu at operamail.com
>>
>>
>> On Tue, Jan 13, 2015, at 07:13, AKLDGUY . wrote:
>>> Hi Leslie
>>>
>>> > (A) The load in the drain of the cascode mixer. At the moment the
>>> >     "top" device in the JFET mixer is a 'step-down' bifilar
>>> >     transformer, 4:1 ratio. Bruce Long observed he doesn't favour
>>> >     the cascode mixer, and (in this type of service) I may not
>>> >     need gain. Your observation focuses on the load in the
>>> >     collector circuit. You suggest I replace the transformer with
>>> >     a choke. The change is to simplify the circuit.
>>> >
>>> >I'll try that, and report the result. Unfortunately I don't have
>>> >any test instruments that will allow me to measure benefit. I'm
>>> >using stations on the broadcast band and my ears as 'test
>>> >instruments'. Rather subjective, I think.
>>> >
>>> >I have a 1.5mH "moulded" choke in my parts box, but it's not one of
>>> >those 'old-style' pie wound (or maybe pi-wound) chokes, so it will
>>> >have a lower self-resonant frequency. At 400kHz the load "Z" will
>>> >be about 3700 ohms. Anyway, I'll use it.
>>> >
>>> > I thought the transformer might have a different kind of
>>> > advantage. The converter has too much gain. When I listen to
>>> > Sydney, (about 100km from my QTH) I have to turn the RF gain
>>> > 'back' because I hear an objectionable 'buzz' in the speaker. I
>>> > haven't had a CRO on the signal, but I suspect I'm
>>> > saturating/clipping the signal in one of the later stages in the
>>> > '453. The lower "Z" (and lower output voltage) will assist here.
>>> >
>>> >Bruce Long suggested I use a diode ring mixer. Oh I hate him, very
>>> >much, not because he's right a little, but because he's completely
>>> >right. Ouch, ouch, ouch. I don't need any gain in the mixer stage
>>> >at all, and certainly not the gain that can be got from a cascode
>>> >mixer! Plus, a DBM has other benefits too. Anyway, I have this
>>> >circuit drawn, and (now) working, and I'm going to finish it before
>>> >working on an alternative design.
>>>
>>> OK DON"T BOTHER TRYING THE CHOKE IF THE BIFILAR TRANSFORMER IS
>>> ALREADY WORKING WELL. I THOUGHT THIS WAS A PROPOSED DESIGN, NOT ONE
>>> YOU HAD ALREADY BUILT AND GOT WORKING. IF YOU HAVE TOO MUCH OUTPUT
>>> (AND SUSPECT YOU MIGHT BE OVERDRIVING THE BC-453), TRY A CAPACITIVE
>>> VOLTAGE DIVIDER AT THE OUTPUT.
>>>
>>> > (B)Grid/gate resistor, R2. This is connected between the gate and
>>> >    common on the 'bottom' JFET. In other words, it's in parallel
>>> >    with the input LC 'tank' circuit. I understand your criticism.
>>> >    I don't show how I connected the coil to the gate circuit; I'm
>>> >    using a 'plug-in' coil. I added R2 to stop the gate 'floating'
>>> >    when I unplug the coil. Do you see R2 loading the LC-tank
>>> >    circuit? Should I make it 1 Meg? Should I remove it, or keep
>>> >    it? Is it doing the service I imagine?
>>>
>>> OK I DIDN'T REALISE THE COIL WAS PLUG IN, SO YES YOU CERTAINLY DO
>>> NEED R2 TO PROVIDE A GROUND RETURN WHEN THE COIL IS REMOVED. THAT
>>> VALUE (100K) IS FINE AND WON'T LOAD THE COIL. KEEP IT.
>>>
>>> >(C) Decoupling the mixer (Q1) source resistor. BTW, did you notice
>>> >I cleverly designed my circuit with two Q1s and 2 Q2s? This isn't a
>>> >mistake on my part, no sir. It's intended to confuse the enemy, a
>>> >deliberate, clever and quite un-obvious, trick. Anyway, I think
>>> >your point about the decoupling "C" is that at some frequency the
>>> >series inductance in the cap will create a resonance, and hence a
>>> >problem. Do I understand your point? If so, I'm not sure I
>>> >understand the cause-to effect logic. A high "Z" in the source will
>>> >provide degeneration, not feedback? What do you say? Beyond that, I
>>> >have a lot of confidence in the 33 ohm "stopper" on the gate. A
>>> >'grid-stopper' has been around for a long time. Does any-one else
>>> >have actual experience on this point? I'm writing from 'book
>>> >knowledge' here, not experience. Experience trumps every other type
>>> >of knowledge.
>>>
>>> YES A LITTLE CONFUSING TO HAVE TWO Q1's AND Q2's ON THE SAME
>>> DIAGRAM. THE POINT ABOUT THE 100nF IS THAT IT'S TOO BIG FOR VHF. DUE
>>> TO THE WAY THOSE CAPACITORS ARE CONSTRUCTED, THERE CAN BE ENOUGH
>>> SERIES INDUCTANCE AT VHF TO CAUSE PROBLEMS. SO, YES IT'S BETTER TO
>>> LEAVE IT IN BUT DECOUPLE IT WITH MUCH SMALLER CAPACITORS THAT WILL
>>> DO THE JOB AT VHF.
>>>
>>> > (D)Linearity. Good point. Putting your critique alongside Bruce
>>> >    Long's dislike for the cascode, I'm tempted to re-arrange the
>>> >    circuit as follows. Turn Q2, the 'top' FET into a source
>>> >    follower (for the LO signal) and wire Q1 (the bottom JFET) as a
>>> >    simple source-injected mixer (as shown in many ARRL handbooks).
>>> >    This would answer your point about the supply voltage. I'd like
>>> >    to keep the supply at 12V; this is a common supply voltage, and
>>> >    allows battery operation if the circuit is to be used in other
>>> >    places, eg in a portable set. Should I succumb to tempatation
>>> >    and re-draw/re-wire? (As a bonus, I could re-allocate the
>>> >    device numbers so I have only one Q1 etc.)
>>>
>>> THAT MIGHT BE QUITE A GOOD MOVE. AT HF, THERE'S HARDLY ANY NEED FOR
>>> AN RF AMPLIFIER, SO COUPLING THE TUNED CIRCUIT STRAIGHT TO A MIXER
>>> WOULD WORK QUITE WELL. IT'S NOT LIKE THERE'S A NEED FOR SOME GAIN TO
>>> SET THE NOISE FIGURE - ATMOSPHERIC NOISE WILL BE THE LIMITING
>>> FACTOR, NOT FIRST DEVICE NOISE. BESIDES, THERE'S TONS OF GAIN IN THE
>>> PSEUDO IF SYSTEM (THE BC-453). YOU HAVE MORE GAIN THAN YOU NEED.
>>>
>>> > (E) Circuit simplicity If I make the change suggested in the
>>> >     previous para, I could simplify the L.O. buffer. However, I'm
>>> >     not chasing simplicity; and it seems that many people see
>>> >     adding one extra "device" as 'complicating' the circuit much
>>> >     more than adding one extra resistor to the same circuit. But
>>> >     look at the circuit of the oscillator! Ignoring the buffer -
>>> >     it's simplicity itself. Just a source follower, followed by a
>>> >     common gate amplifier - AND the signal out is a very nice sine
>>> >     wave. Surely the signal out warrants the circuit you see?
>>> >
>>> >An extra device costs five cents (if it's bipolar) or 30c (if it's
>>> >a JFET). I spent days drawing the circuit, cutting aluminium
>>> >channel, drilling holes to mount the capacitor correctly (and so
>>> >on). The cost of an additional transistor is zip! I'm about to add
>>> >'string, dial and drum' to the tuning. Making that work will be
>>> >complicated!
>>>
>>> SOUNDS LIKE YOU HAVE A WELL DESIGNED LO WITH NICE WAVEFORM. HALF THE
>>> BATTLE IS WON. I'M SURE YOU'LL HAVE BETTER SUCCESS BY GOING TO A
>>> CONVENTIONAL RF STAGE OR JUST THE MIXER ALONE. THAT CASCODE CIRCUIT
>>> IS WAY COMPLICATED FOR WHAT'S NEEDED AND LOOKS LIKE A RECIPE FOR
>>> INSTABILITY, WITH MIXER Q2 ABOVE GROUND AT ALL THREE DEVICE
>>> ELEMENTS.
>>>
>>> 73 de Neil ZL1ANM
>>>
>>> On Mon, Jan 12, 2015 at 11:25 AM, Leslie Smith
>>> <vk2bcu at operamail.com> wrote:
>>
>> --
>> http://www.fastmail.com - Does exactly what it says on the tin

-- 
http://www.fastmail.com - Accessible with your email software
                          or over the web