This is only a preview of the May 2015 issue of Silicon Chip. You can view 29 of the 104 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Items relevant to "Appliance Earth Leakage Tester":
Articles in this series:
Articles in this series:
Items relevant to "Balanced Input Attenuator For Audio Analysers & Scopes":
Items relevant to "4-Output Universal Voltage Regulator":
Purchase a printed copy of this issue for $10.00. |
Vintage Radio
By Rodney Champness, VK3UG
The AWA Radiola 523-M: the last
vibrator-powered radio
Battery/vibrator-powered domestic radios
started life in the 1930s and continued to be
manufactured in Australia until the late 1950s.
They were still used in some areas of rural
Australia well into the 1970s.
W
HAT EXACTLY IS a vibrator?
Well, it’s not what you might be
thinking, a least not as used in batterypowered valve radios. In operation, a
vibrator converted a low battery voltage (typically 2-32V) to a much higher
voltage, necessary to power the valves
used in battery-operated receivers.
A vibrator is basically an electromagnetic switch that opens and closes
a set of contacts at a fixed frequency
of 50-150 times per second, depending on the particular circuit it’s used
in. It’s either a double-pole or 4-pole
switch that switches DC power one
92 Silicon Chip
way and then the other through the
centre-tapped primary winding of an
iron-cored transformer.
This rapid switching results in a
waveform across the winding that approximates the waveform from an AC
supply. The secondary winding has
many more turns on it than the primary
and so a much higher voltage is produced across it. The secondary is also
centre-tapped and its AC output is converted to DC by a second set of points
in the vibrator. These are synchronised
with the first set of points, hence the
name “synchronous vibrator”.
Synchronous vibrators are the most
likely type to be found in domestic
radios intended for remote areas where
mains power was unavailable. By
contrast, so-called non-synchronous
vibrators were more likely to be found
in car radios. This latter vibrator type
required an external rectifier to convert its AC output to DC and either a
6X5GT or 6X4 valve was often used
for this task.
So that is basically how vibrator
power supplies work but there are
other things to consider to make them
suitable for powering radio receivers. In operation, a vibrator makes
and breaks the voltage applied to the
transformer and this results in an
abrupt change in the current being
drawn from the supply. As a result,
the transformer’s winding inductance
tries to maintain this current across the
vibrator’s points as they open. Unless
steps are taken to prevent this, the
result is severe sparking which would
completely destroy the points within
a few hours of operation.
To solve this problem, one or more
capacitors are connected across either the primary or the secondary of
the transformer, or both windings in
some cases. By carefully selecting the
capacitor values, the circuit (including
the winding) resonates at the switching frequency and the sparking is
markedly reduced.
If you are repairing a vibrator and
the value of the capacitor is unknown,
the trick is to try a variety of values and
select the value that causes the vibrator
to draw the least current. The voltage
ratings of these capacitors, commonly
called “buffer capacitors”, may need
to be as high as 2000V DC.
Because they are used under quite
arduous conditions, polypropylene
types should be used. Polyester capacisiliconchip.com.au
Fig.1: the circuit of the AWA Radiola 523-M uses the 1R5, 1T4, 1S5
& 3V4 series of valves. V1 is the converter stage, V2 the IF amplifier,
V3 the detector/AGC/first audio amplifier stage and V4 the audio
amplifier output stage. The dashed box contains the vibrator circuit.
tors can have a short life-span when
used as buffer capacitors and so should
not be used. However, they can be used
in all other parts of the power supply
where paper capacitors were used.
One drawback of a vibrator supply
is that while the sparking is reduced
by using suitable buffer capacitors, RF
(radio-frequency) interference can still
be quite evident. To overcome this,
the whole vibrator supply is housed
in a shielded metal enclosure and the
leads going into or out of this enclosure
are filtered to remove interference. In
addition, the supply is mounted on
rubber buffers so that there is little or
no physical noise from the operation
of the vibrator.
In short, designing a vibrator power
supply with low electrical and acoustic noise is not as simple as designing
a conventional power supply.
in remote regional areas.
As shown on Fig.1, the antenna
input circuit has an IF (intermediate
frequency) rejection circuit (L1, C1)
connected across the antenna-earth
terminals. That’s there to prevent IF
signals from being picked up and fed
back in through the converter stage,
which could upset the receiver’s operation.
The rest of the input circuit is conventional, with capacitor C2 giving
some boost to the higher-frequency
signals. C3, one section of the tuning
gang, tunes the incoming signal and
this is then fed to the grid of converter
stage V1.
The oscillator tuned circuit is connected between V1’s grid and chassis,
while feedback winding L4 is con-
Circuit details
Fig.1 shows the circuit details of
the AWA Radiola 523-M. It’s really
quite conventional for a 4-valve battery/vibrator-powered receiver built
around 1949 and uses the economical
1R5, 1T4, 1S5 & 3V4 series of valves.
These valves required only 90V HT
and 50mA of filament current to perform well, the low filament current
being necessary to minimise power
consumption from the dry batteries
used to power the receiver – important
siliconchip.com.au
This photo shows the dilapidated state of the cabinet, dial scale and speaker
cloth before restoration, while the photo on the facing page shows the set after
restoration. The exterior of the set now looks almost like new again.
May 2015 93
ment supply line is series connected
across the 4V supply, with pins 1 & 7
connected to +4V (via L12) and pin 5
connected to the filament of the 1T4.
By doing this and earthing the grid via
resistor R10, the valve is effectively
biased to around -3.25V without further measures. To get the additional
bias voltage required, a portion of the
oscillator’s grid voltage is also applied
to the 3V4 to raise the bias level to
around -6.5V.
Basically, some innovative circuit
variations are needed when the filaments of valves are series connected,
so that correct operating conditions
are achieved. We’ll take a look at the
power supply circuit later on.
Restoration
These photos show the chassis before (top) and after (bottom) restoration. The
valves were cleaned by washing them in soapy water, while the chassis was
cleaned by brushing away the dust, then scrubbing it with a kerosene-soaked
pad. The enclosure in the middle of the chassis houses the vibrator supply.
nected to the screen grid which acts as
the plate for the oscillator. The other
end of L4 is at virtual earth/chassis
since capacitor C6 bypasses any RF
signals (whether IF or local oscillator)
to earth. In addition, L4’s inductance
is low enough that C6 effectively bypasses the lower ends of C11 & L6 to
earth as well.
The output from converter stage V1
is fed through the first IF transformer
consisting of C11, L6, C12 & L7. From
94 Silicon Chip
there, the resulting 455kHz IF signal is
fed to IF amplifier stage V2 (1T4) and
then fed via a second IF transformer to
the detector and AGC diodes in valve
V3 (1S5). The recovered audio is then
amplified by V3’s pentode section after which it is fed to the grid of audio
amplifier stage V4 (3V4). Output stage
V4 then drives the loudspeaker via a
transformer.
The 3V4 needs around -6.5V of bias
in this circuit. To achieve this, the fila-
The chassis is easily removed from
its cabinet by removing two screws
and then sliding it out. Note that the
on-off volume and tuning controls are
concentric and are mounted through
the centre of the dial scale, so they also
come out with the chassis. This would
have to be the easiest set to dismantle
for service that I have come across.
Once the chassis had been removed,
the very grubby cabinet was scrubbed
clean using a nail brush dipped in
soapy water. This was done carefully
though, to avoid wetting the paper
label pasted inside the cabinet (this
label shows the chassis layout). The
cabinet was then carefully rinsed with
clean water and rubbed down with car
cut and polish. It now looks almost
like new again.
Restoring the chassis wasn’t anywhere near as easy, as mice had made
a home in the set and the acid in their
urine had etched through the plating
on the chassis in quite a few places.
Dozens of small pieces of paper had
also been left in the chassis by the mice
but they hadn’t done any damage to
any of the parts or the wiring.
I began by brushing away the dust
and other muck as best I could, then
used a kitchen scourer soaked with
household kerosene to clean the chassis. Restoring the chassis to pristine
condition would have involved removing all the parts and the wiring,
then re-plating the chassis and other
metal parts and rebuilding the set. In
fact, some vintage radio enthusiasts
actually do this and their restored
radio sets look like new.
When it came to this set, I was
happy to leave most of the parts in
siliconchip.com.au
place and simply clean the chassis
as best I could. Removing everything
and completely rebuilding the set is a
time-consuming process.
The dial scale was also dirty so I very
carefully cleaned it with a soft brush.
I then used some soapy water on an
inconspicuous part of the dial scale to
see if the lettering remained in place.
All seemed to go well, so I cleaned
the rest of the dial and all the lettering
remained intact. Unfortunately, it was
still a little dirty when the water dried,
so I tried the same technique again on
the test area and this time some of the
lettering did come away.
Apparently, the letters had been
softened by the first round of cleaning
so I left the remainder of the dial scale
alone and simply left it to dry before
carefully putting it back together. It
wasn’t a tragedy but I still wasn’t at all
pleased with myself as I hadn’t been
careful enough. It’s always important
to be very careful with dial scale markings – some remain on the glass and so
the dial can be easily cleaned while in
other cases, the letters can come away
with very little provocation.
Overhauling the vibrator
Having cleaned the chassis and dial,
I turned my attention to the vibrator
assembly. First, I removed the HT
(high tension) filter choke (L13) and
the LT choke (L12) to improve access
to the vibrator mounting points. That
done, I disconnected the three wires
going into the vibrator supply module
(earth/chassis, +4V input and the HT+
output) and disconnected the earthing
braid that connects the vibrator’s metal
case to the chassis.
The next step was to remove the
three circlips that secure the resilient
mounting to the chassis and remove
the assembly. The plastic sleeves over
each of the three mounting posts were
still in good order but the resilient
mounts were in a bad way. I didn’t
have the correct “spongy” material for
these mounts on hand, so improvisation was necessary when it came to
replacing them.
First, I glued some foam rubber
material to the bottom of the shielded
enclosure, to keep it clear of the chassis. This was simply cut to suit and
mounted near each of the mounting
posts. The material used is approximately 5mm thick and 25mm wide and
is readily available from Clark Rubber
in whatever length you want.
siliconchip.com.au
The chassis and dial scale cleaned up quite well, although some of the lettering at the bottom of the dial scale lifted off during the cleaning process. Great
care needs to be taken when cleaning dial scales to avoid this problem.
Once the glue was dry, I remounted
the vibrator supply on the chassis,
with an 8mm ID rubber grommet fitted to each mounting post. This was
followed by a thick fibre washer, the
original metal washer and the circlip,
to hold the mounting assembly together. I also attached a self-adhesive
felt furniture pad to the side of the
enclosure, so that it could not possibly touch an adjacent IF transformer
which is only a few millimetres away.
As stated above, resilient mounting
of vibrator supplies was routine so that
the physical noise made by the vibrator
was minimised.
The actual vibrator power pack is
also mechanically isolated from the
shielded enclosure. In order to remove
it, it’s necessary to remove the selftapping screw at the top back edge
of the enclosure and the three screws
which go through the side. The supply
can then be lifted out of the enclosure
for restoration.
I began by replacing all the paper
capacitors, even though I found that
they all tested OK, much to my surprise (the same types in the main part
of the receiver all later tested leaky).
Buffer capacitors C26 & C27 were
replaced with polypropylene types
and the other paper capacitors with
polyester types. The 20µF 200V electrolytic capacitor was replaced with a
22µF 160V unit and that was quite safe
to do as the HT voltage won’t exceed
around 120V, even if the valves aren’t
drawing any current.
Basically, I replaced all the capacitors just to be sure and because it’s very
time consuming to access the vibrator
supply to replace any defective parts.
The only part in the supply that is easily accessible is the vibrator itself, as
it was considered to be a consumable
item with a limited life.
My next step was to remove the
mechanical vibrator assembly from
its case. This involved desoldering
the lug at the base of the can, then
removing the circlip that holds the
vibrator’s base in place and sliding the
assembly out.
The first thing I noticed was that a
foam rubber support at the top which
keeps the assembly away from the case
had gone “gooey”. I scraped the goo off
and then got busy with some contact
cleaning strips that I’ve had for years
to clean the points. Quite a lot of black
dust came off the points so the effort
was worthwhile.
If you don’t have contact cleaning
strips, then very fine wet and dry paper can be used instead to carefully
clean between the various points in
the vibrator.
Before cleaning the points, I found
that the vibrator wouldn’t start reliably
but it did so after the points had been
cleaned. Once it was all working, I
found some thin rubber strip around
20mm wide and wrapped this around
the end of the vibrator assembly. This
was then tied in place with thin plastic spaghetti tubing, the idea being to
prevent the vibrator from coming into
May 2015 95
The original paper capacitors were all replaced with polyester and ceramic types, while the electrolytic capacitors were
all replaced with modern equivalents. Several resistors and the speaker transformer also had to be replaced.
contact with the inside of the can and
causing acoustic noise.
The refurbished vibrator supply has
since proved to be reliable although its
output voltage is somewhat less that
I would prefer. But then, this unit is
now over 65 years old.
Electronic repairs
Quite a bit of work was necessary to
restore the chassis to working order.
My first step here was to check the
paper capacitors and these all proved
to be leaky. As a result, they were
replaced with polyester and ceramic
types. The electrolytic capacitors were
also all replaced, as they had been in
this set for many years. Further component checks then revealed three
resistors that were out of tolerance and
so they too were replaced.
In addition, much of the wiring had
perished so I cut the lacing away from
the loomed wires and replaced any
suspect leads. In the end, I replaced
about 80% of the wiring which was a
rather time-consuming task.
Next, the valves were removed and
cleaned in soapy water after checking that their filaments were OK. The
filaments were all intact but not so the
primary winding of the speaker transformer. This meant that the speaker
transformer had to be replaced and I
96 Silicon Chip
then took a close look at the speaker
itself. It initially looked to be a dead
loss as there was grit in the voice coil.
However, after removing the felt pad
in the centre of the cone, I was able to
gently blow out the gritty bits. The felt
pad was then re-glued in place, as was
the outer edge of the voice coil which
had separated from the frame.
Once these repairs had been completed, the speaker worked quite well
again. And that really was good news
because I didn’t have a matching spare.
At this stage, the two filter chokes
were reinstalled and the leads from the
vibrator power supply reconnected.
During this work, the 4V supply cable
running from the set to the battery
was found to be very much the worse
for wear. It has two wires which are
shielded and another two that are
outside the shield.
Obtaining an original replacement
cable would have been impossible,
so I made a replacement cable up. It
doesn’t look like the original but it
functions the same way. Fortunately,
I was able to come up with some braid
of sufficient diameter to accommodate
four wires through its centre – two for
the filament supply and two for the vibrator supply. This braid was obtained
from a length of RG213 coaxial cable
– it was just a matter of removing the
outer cover and pulling out the centre
and the insulation.
Once the new cable had been assembled, I wound electrical tape along
its length to prevent any mishaps due
to short circuits. What was interesting was that once I had the set fully
operational, I found that neither of the
earth/chassis wires inside the cable
performed any useful purpose. The
braid was earthed to the chassis and
went to the negative terminal of the
battery. The two wires that run from
the positive terminal of the battery are
needed to minimise any ripple on the
filament line that runs from the vibrator supply. In effect, the battery acts as
a very big filter capacitor.
The braid is necessary to shield the
vibrator supply lead as quite noticeable
RF is present on this wire. I found that
in order to minimise this interference,
it was necessary to add a 0.1µF capacitor between the wiper of the lefthand
section of S1 and earth (this removed
almost all of the interference).
With the valves reinstalled, it was
time to test the receiver. After applying power, the HT rose to around 75V
which is a bit low but the set certainly
showed signs of life. It worked well
at the bottom end of the band but it
appeared dead when tuned to frequencies above 800kHz. Subsequent
siliconchip.com.au
Silicon Chip
Binders
REAL
VALUE
AT
$16.95
*
PLUS P
&
P
This is the view inside the rear of the cabinet prior to restoration. Unfortunately,
mice had made a home in the set and the acid in their urine had etched through
the plating on the chassis in quite a few places.
adjustment of the antenna tuned circuit trimmer capacitor then allowed
it to pick up stations right across the
band, although it still wasn’t working
all that well.
Suspecting alignment problems,
I then tuned to a weak station and
adjusted the IF transformer coil slugs,
The set’s performance improved quite
markedly and is now quite good – better than I expected in fact from a set
with just four “battery-type” valves.
The interference level from the vibrator supply was less than expected, too.
Electronic vibrator
I’ve always stuck to mechanical
vibrators for my receivers and in fact
have over 100 vibrators in my collection. Some of these are unused and are
in “brand-new” condition, while others are worn out and need servicing.
Over the years, various articles
have appeared in radio and general
electronics magazines describing how
to replace electromechanical vibrators
with solid-state versions. However,
although these did work, some had a
tendency to overheat and most were
too bulky to fit inside the housing used
for the original mechanical vibrator.
To overcome these problems, Tony
Maher of the Historical Radio Society
of Australia (HRSA) developed a solidstate MOSFET-based replacement
module several years ago. It’s designed
to take the place of a variety of vibrators and what’s more, it fits easily into
old vibrator housings.
In fact, Tony has developed two
solid-state vibrator versions. If the
siliconchip.com.au
first version is used, the supply will
put out about the same voltage as for
a mechanical vibrator. The second
V1 version is more efficient and if fitted instead, the supply will have an
output that’s up to 15% higher than
the original.
I decided that it could be an interesting exercise to try one of Tony’s
electronic vibrators in the old AWA
Radiola 523-M receiver. As a result,
I purchased two of the V1 modules.
When fitted with its original mechanical vibrator, the 523-M’s supply
put out around 75V on load. This
clearly indicated that the vibrator
had seen better days, since the voltage should have been nearly 90V. The
current drain of the supply by itself (ie,
when tested out of the set) was around
0.6A at 75V output and a 12mA load.
When I substituted one of Tony’s V1
(second version) modules, the results
were excellent. The output voltage
was now around 107V with a 17mA
load, while the current drawn from
the 4V supply was about the same as
it was for the electromechanical vibrator (which produced just 75V). So the
electronic version is definitely much
more efficient.
In fact, my measurements showed
that the supply has an efficiency of
75% when using the Mosfet solid-state
module. This drops to just 40% when
using the rather tired mechanical
vibrator. A new mechanical vibrator
should result in better than 50% efficiency.
I also checked the interference produced by both vibrators and found
Are your copies of SILICON
CHIP getting damaged
or dog-eared just lying
around in a cupboard or
on a shelf? Can you quickly find a particular issue
that you need to refer to?
Keep your copies
safe, secure and
always available with
these handy binders
These binders will protect your
copies of SILICON CHIP. They
feature heavy-board covers,
hold 12 issues & will look great
on your bookshelf.
H 80mm internal width
H SILICON CHIP logo printed in
gold-coloured lettering on spine
& cover
Silicon Chip Publications
PO Box 139
Collaroy Beach 2097
Order online from www.
siliconchip.com.au/Shop/4
or call (02) 9939 3295 and
quote your credit card
number. *See website for
overseas prices.
May 2015 97
this module. The noise suppression in
some mechanical vibrator sets was not
very good and Tony’s first design using
a 4047 IC generates very little noise, so
it can overcome such problems. Tony
can be contacted by email at tmaher<at>
detection.com.au
Summary
The view inside the rear of the cabinet after restoration. Great care was taken
during cleaning to avoid damaging the label stuck to the inside of the cabinet
and the ARTS&P label on the back of the chassis.
that the electronic version produces
considerably less than the mechanical
version. Added to that, the electronic
version is acoustically silent. These
initial tests, by the way, were all done
with the supply outside the receiver.
However, I was confident that the set
would be reasonably free of RF noise
with either version fitted to the set
and that I could then minimise any
residual interference once the set was
fully operational.
Note that the electronic vibrators
can be wired to replace mechanical
4V, 6V & 12V synchronous, nonsynchronous and split reed vibrators.
In fact, they can be used in any sets in
this voltage range with no alteration
at all. In addition, by replacing the
Mosfets with higher rated types, it’s
possible to use them in 24V and 32V
domestic radio sets.
Both 6V and 12V car radios can be
powered by these modules without the
use of a heatsink. Note, however, that
the electronic vibrators are designed
for negative earth, so some vibrators
in positive earth vehicles cannot be replaced by electronic versions without
making suitable modifications.
By making other modifications, they
could also possibly be used in place
of the large mechanical vibrators fitted
in the 32V DC-to-230VAC inverters
that were used to power TV receivers
in remote rural areas. In this application, because of the large power output
required, it would be necessary to fit
the Mosfets to a heatsink and to adjust
the frequency of oscillation. In short,
this is quite a versatile module.
In-circuit comparisons
With the mechanical vibrator in circuit, the HT was again 75V and the set
drew about 1A (more than the printed
specifications). I then substituted the
electronic vibrator and installed a 1kΩ
resistor in series with the HT rail to
reduce the HT rail to 90V. This time,
the current drain from the 4V source
was 0.75A which is noticeably less
than for the mechanical vibrator.
In the end though, I decided to leave
the old mechanical vibrator in the
receiver. It still does the job, even if
not as well as it did when new, and it
maintains the set’s originality.
However, I was very pleased with
the performance of the electronic vibrator and can confidently recommend
its use. They are available from Tony as
a kit and I’ll probably convert some of
my vintage radio transmitters over to
This electronic
vibrator is small
enough to fit
inside the case
of a mechanical
vibrator.
98 Silicon Chip
The AWA Radiola 523-M is quite
an interesting little set. It works well
when powered from 4V, although
why AWA chose to use this unusual
supply voltage is a bit of a mystery,
especially when virtually every other
manufacturer in the late 1940s used
6V. I suspect that the reason is due to
the earlier use of 2V filament valves,
with one 2V cell of a 6V battery being
used to power the filaments and the
remaining 4V from the other two cells
being used for the vibrator supply.
In operation, the current consumption was balanced between the three
cells. The battery, of course, could
be switched in and out of the car for
charging if the farm was in a remote
area.
Quite frankly, I had expected a lot
more interference from the vibrator
supply, so I was pleasantly surprised
on that front. Other brands filtered the
input and output supply rails more
thoroughly than in this set and there
was no need for a shielded supply
cable to minimise interference.
Physically, the circuit could have
been built onto a larger chassis, as
there is quite a lot of room between
the front of the chassis and the cabinet.
This would have made the set easier
to service, although it’s still relatively
straightforward.
One thing I don’t understand is why
the dial lamp was permanently left in
circuit. Almost all vibrator sets had
dial lights that were controlled by a
pushbutton switch, so that they could
be turned on only when tuning the set.
This was done to minimise current
drain and prolong battery life. Removing the battery when it went flat and
taking it to a local garage for charging
was a real chore, so the longer the radio
worked between recharges the better.
This set was obviously designed for
the cheaper end of the market and it
did a good job there. However, it does
need a good outdoor antenna and earth
to give reasonable performance. It is a
set well worth having in a collection
particularly if, like me, you like battery
SC
valve radios.
siliconchip.com.au
|