This is only a preview of the October 2019 issue of Silicon Chip. You can view 39 of the 112 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 "45V, 8A Bench Power Supply to build":
Items relevant to "High resolution Audio Millivoltmeter/Voltmeter":
Items relevant to "Precision Audio Signal Amplifier":
Purchase a printed copy of this issue for $10.00. |
Vintage Radio
By Ian Batty
Healing M602T transistor
mantel radio
Good performance and long battery life with ‘modern’ 60s styling – a
transistor radio that’s at home in the kitchen.
I picked up this transistor set at an
HRSA auction a while ago. It’s an Australian set, compact and easy to use.
I recently described Healing’s fine
valve portable, the 404B (April 2019;
siliconchip.com.au/Article/11533).
So we can now directly compare that
to the six-transistor M602T from the
same manufacturer.
The M602T was released in 1960
and followed on from the designs that
had matured by the late 1950s. It uses
six alloyed-junction transistors: three
in the RF/IF section and three in the
audio stages. This puts it in the second generation of transistor sets. The
short-lived first generation used inferior grown-junction transistors.
So as well as comparing this set to
the valve radios that were designed
just a few years before it, we can also
compare it to the transistor sets which
came soon after (ignoring the few hybrids which bridged the gap).
siliconchip.com.au
The M602T’s construction uses a design that was passing out of favour at
the time: a punched and pressed steel
chassis using tag strips, transistors
mounted in grommets, and point-topoint wiring. Its mechanical construction is complicated, with three metal
sub-chassis sections.
The plastic cabinet is quite generous, putting it in the ‘small mantel/
portable’ class. The chassis, although
not especially compact, leaves plenty
of room for its 5-inch Rola 5F speaker
and the long-lasting type 276 battery.
It’s a conventional six-transistor set,
using the same cabinet as mains-powered valve models 410E & 411E.
Being a larger set than the 404B, the
M602T has a more relaxed and usable
control layout. The large dial features
station call signs, a reminder of Saturday afternoon footy and Top Forty Hit
Parades. The slow-motion dial makes
tuning easy.
Australia’s electronics magazine
From top to bottom, the knobs are
the on/off switch, volume control and
tone control. Separating the on/off
and volume functions reduces wear
and extends the life of the volume
pot, as it can be left in the same position most of the time. I wish the controls were labelled; maybe you’re just
meant to know.
Compared to the all-valve 404B,
the M602T is a pleasure to work on,
though its complicated construction
sees the tuning gang buried between
the front and back chassis plates,
and the trimmer capacitors partly
obscured, so adjusting it is a bit difficult.
As the IF transformers are all singletuned, the slugs are easily accessible
from the rear.
On my set, they appeared to use wax
to prevent accidental movement, so I
strongly advise against using metaltipped alignment tools. If you need
October 2019 105
to get a slug to move, try using a hair
dryer/heat gun to warm the can and
soften the wax.
All minor components, including
the transistors, are easily accessible for
measurements or replacement.
Circuit details
The circuit of the M602T is shown
above. Note that some sets may have
alternative transistor types to those
shown, especially the ones which were
made in Japan.
The circuit begins with the usual
self-oscillating converter, TR1. This is
a 2N219 or a 2SA15, roughly equivalent to an OC45 rather than the higherperforming OC44 or similar that we’re
used to seeing in this stage. The converter uses emitter injection, so it’s
easy to inject a signal directly into the
base for testing.
The antenna circuit has a ferrite
rod with a separate primary winding
to allow an external antenna connection to be used.
As is usual for converters, the baseemitter forward bias of some 50mV
is lower than the usual 150~200mV
for germanium transistors. This is because converters need to operate in
a non-linear mode close to Class-B,
so that they can create the necessary
sum-and-difference signals from the
incoming radio station and the local
oscillator (LO).
LO transformer L2 has two windings, with the secondary tapped to
106
Silicon Chip
supply feedback to the low-impedance emitter.
As the tuning gang has identical sections, padder C5 (430pF) reduces the
LO section’s capacitance swing to give
a ratio of roughly 4:1 as the set tunes
over the broadcast band.
The converter feeds first intermediate frequency (IF) transformer T1. This
has a tapped, tuned primary and untapped, untuned secondary.
The first IF amplifier transistor, TR2,
is either a 2N218 or a 2SA12, both
slightly lower-performing versions
of the 2N219/2SA15. As the stage is
gain-controlled, bias resistor R5 has a
relatively high value of 82kW. This allows the AGC signal from diode D2 to
reduce TR2’s gain with increasing signal strength. This is filtered by 6.8kW
resistor R11 and 10µF capacitor C51 to
remove the audio component.
TR2 feeds second IF transformer T2,
also with a tapped, tuned primary and
untapped, untuned secondary. The
collector is fed from the supply via a
2.7kW resistor, R7. AGC extension diode D1 has its cathode connected to
R7 and its anode to the primary of first
IF transformer T1.
Compared to
some of the
other radios of
the early 1960s,
the M602T used
point-to-point
wiring, providing
a compact but
messy layout.
Australia’s electronics magazine
siliconchip.com.au
The redrawn circuit diagram for
the Healing M602T. Some models
of this set used different transistors,
most of them made by Hitachi, for the Japanese market. Additionally, the values
of C8 (330nF), C11 (47nF) & C16 (56pF) differ, likely for similar reasons.
In normal operation, there’s a voltage drop of about 1.8V across R7. Since
the converter’s collector sits at about
7.5V, D1 will have a reverse-bias of
around 1.3V. This means that D1 is
cut off with weak signals so it will
have no effect.
As the AGC takes over, and TR2’s
collector current falls, TR2’s DC collector voltage rises. This brings D1’s
cathode voltage closer to 7.5V, so D1
will start to conduct with strong signals. As it does so, it shunts current
from the converter’s collector, further
reducing the set’s gain and giving improved AGC action.
The 2N216~219 series are all alloyed-junction RF transistors, exhibiting collector-base capacitances of
around 9pF. So both IF amplifiers need
neutralisation, with 4.7pF capacitor
C15 providing this for TR2.
Second IF amplifier TR3 uses another 2N218/2SA12 with fixed bias.
The usual emitter bypass capacitor to
ground seems to be missing, but this
stage has its base bypassed back to
the emitter terminal via 10nF capacitor C11, and its collector supply is
bypassed to the emitter via 47nF capacitor C12.
This configuration is most often
used in VHF circuits, as it is more
effective than running everything to
ground. In this circuit, it also saves
one capacitor – the emitter bypass capacitor, such as C9 used by the first IF
amplifier TR2.
siliconchip.com.au
TR3 feeds third IF transformer T3’s
tuned, tapped primary, and T3’s untuned, untapped secondary feeds demodulator diode D2, another GEX34/
OA70. D2’s audio output, filtered by
C13, goes to volume pot R51 and also
provides the AGC signal, as described
earlier.
Note that the AGC filter capacitor,
C51, is an electrolytic type. Electrolytics are not recommended for RF/
IF bypassing, so if you have an M602T
suffering from oscillation or some
other strange RF/IF fault, C51 may be
the culprit.
The signal from the volume control is coupled to the base of audio
driver TR4, a 2N408. TR4 has conventional combination bias, and its
collector drives the primary of output transformer T4. The tone control
pot, R52, and 47nF top cut capacitor
C17 connect between TR4’s collector
and ground.
As R52’s resistance is reduced, C17
progressively shunts more of the high
audio frequencies, giving more and
more top-cut and producing a more
‘mellow’ audio tone.
T4’s secondary provides the pushpull drive to transistors TR5/6, both
2N270s. These have higher power
ratings than the OC72, but less than
the later OC74/AC128 types from
Philips/Mullard. TR5/6 operate in
Class-B, with around 180mV of forward bias.
Don’t be confused by the positive
voltage readings in the emitter/base
sections of the circuit; I’ve measured
relative to chassis ground, and since
R16 is connected between the battery
and chassis, the chassis sits about
one volt below the battery’s positive
terminal.
TR5/6 get their bias from the parallel combination of R19/R41 – again, a
slightly confusing connection, but it
works perfectly. Thermistor R41 compensates for ambient temperature, reducing the forward bias for TR5/6 at
higher temperatures, where their baseemitter junction voltages fall. This
provides a relatively constant collector current, protecting from thermal
runaway.
Quirky decoupling
Class-B output stages draw low
L1
C52/55
C2
T4
C3
C1a/b
From the top of the M602T, you can see two large red 100µF electrolytic
capacitors, used for filtering the supply, at the far right. There are a few other
electros in the circuit, which may cause oscillation problems if they degrade.
Australia’s electronics magazine
October 2019 107
Test results
C3
C2
reduction
drive
T1
Variable trimmer capacitor C3, used to calibrate the oscillator, is shown at
centre left. To its right is trimmer capacitor C2 for the antenna. Again to the
right of C2 is the planetary reduction drive for the dial.
quiescent (idling) currents – it’s the
main reason for using them, despite
their complexity compared to Class-A
stages. But Class-B operation results
in considerably larger swings in supply current, increasing substantially
on output peaks.
These current peaks can impress the
output signal on the supply voltage,
making the entire set prone to audio
feedback as the output signal finds its
way back to driving stages. The simple
remedy is to use decoupling, often just
a simple resistor-capacitor filter, in the
supply line going to the low-level RF/
IF/audio section.
But the M602T applies the full battery negative supply to all stages. The
decoupling circuit is placed in the
positive supply, which in this case, is
ground. It’s odd but effective: output
transistors TR5/6 do get the full battery supply, but the battery positive’s
connection to chassis and set Earth is
via 180W resistor R16.
The battery itself (and thus the
output stage) is bypassed by 100µF
capacitor C55, and the driver/RF/IF
stage supply is bypassed by 100µF
capacitor C52.
The circuit office appears to have
numbered ceramic and paper capacitors consecutively from C1, but started
the electrolytics from C51.
The output stage’s thermistor is renumbered as R41, and the volume pot
as R51, despite there being about 20
108
Silicon Chip
fixed resistors in the circuit.
Cleaning up my set
The set I acquired was in fair cosmetic condition, the only damage being two melted areas on the top of the
case and a missing “Transistor” badge
on the decorative metal panel across
the top of the front panel.
The tuning was very stiff. Inspection
showed that the planetary reduction
drive was stuck tight, so I removed,
dismantled, cleaned and re-assembled
it. It was then time to power up the radio, which still had all of its original
components.
Perhaps unsurprisingly, it was
dead. The culprit, an oxidised power
switch, responded to contact cleaner
and a good number of on-off-on-off
cycles.
Having resurrected the set, it was
time to check its alignment. It seemed
to come up well in the IF department,
and responded well at the low end of
the broadcast band around 600kHz.
But it got progressively more and more
‘deaf’ towards the top end.
I checked all the voltages but found
nothing wrong. So I completely dismantled it and washed the ‘dust of
ages’ from the tuning gang and the
rest of the set with isopropyl alcohol.
Once it dried, I checked it again, but
still found it relatively poor at pulling in stations at the upper end of the
frequency range.
Australia’s electronics magazine
Under my test conditions and for
the standard 50mW output, the M602T
needs around 175µV/m at 600kHz,
300µV/m at 1000kHz and 700µV/m
at 1400kHz. Signal-to-noise ratios exceeded 20dB in each case. That’s a
significant drop-off in sensitivity. Signal injection figures recorded on the
diagram also reflect this loss of sensitivity at the high end, and direct injections to the converter base confirm
these figures.
This seemed unlikely to be a problem with the converter, a case of the
mythical “tired transistor”. Just to be
sure, I replaced it with a new old stock
(NOS) OC44, with no improvement.
You may know that conversion gain
varies significantly with LO injection,
so that there is a fairly narrow span of
injection voltage for best performance.
The LO voltage falls by more than 35%
from 600kHz to 1400kHz, so perhaps
this explains the weak top-end performance.
You may recall Kriesler’s Mini 4147 handheld radio (December 2013;
siliconchip.com.au/Article/5633)
using a germanium diode across the
LO primary to help stabilise oscillator output. Perhaps that’s what this
set needs.
RF bandwidth is around ±1.6kHz at
-3dB; at -60dB, it’s ±33kHz. AGC action is excellent: a 40dB increase at the
input gave an output rise of just 6dB.
This set was excellent on strong signals, needing some 500mV/m before
reaching overload.
Audio response is 150~4600Hz from
volume control to speaker; from antenna to speaker it’s 135~2000Hz. Fully
on, the tone control slashes the upper
-3dB point to just 450Hz.
This set can give 400mW of output
at clipping, although that figure is a
bit misleading. At 10mW, Total harmonic distortion (THD) is just 2.5%,
but it’s 7% at the test figure of 50mW,
rising to the usual cutoff value of 10%
at only 120mW output.
I suspect that mismatched output
transistors are the reason for this, but
my junk box failed to disclose any
2N270s. Rather than substitute, I’ll
leave this set all-original until I can
get proper replacements.
At half the nominal battery voltage,
the output clips at 120mW, which is
still quite loud and enough to usefully
squeeze those last few electrons from
the battery.
siliconchip.com.au
The connections on the back of the set, next to the carry handle, are for an external antenna and ground.
Healing 404B vs M602T
The M602T weighs in at around
2.7kg, with the valve-based 404B a
lightweight at just 1.95kg. The M602T
is also a fair bit larger all around, giving
a volume of 5700cm3 versus 1470cm3.
The M602T is about as sensitive as
the 404B at the low end, but nowhere
near as good at 1400kHz, needing some
four times as much signal for the same
audio quality.
The M602T’s audio performance is
superior, giving over four times the
maximum output of the 404B, with
a better frequency response due to a
larger output transformer and speaker.
But I do like the 404B’s visual design: it looks smart and perky with
hints of Art Deco (despite being made
roughly 20 years after that movement
was popular). It stands out in a way
that the more stolid M602T simply
does not.
This all makes sense in context.
The 404B was aimed at the burgeoning market of the late 1940s, with
each manufacturer spruiking the newfound convenience of “camera case”,
all-miniature portables and hoping
their attractive design would stand out
from the pack. The M602T, with its external antenna and Earth connections,
is clearly aimed at the more everyday
“mantel market”.
Which is the better radio? At moderate volumes, there’s not a lot of difference. The M602T’s transistor design,
siliconchip.com.au
with its type 276 battery, gives over 100
hours of use, while the 404B’s single
filament supply cell runs out in less
than five hours.
Given the 404B’s total power consumption at over 900mW compared
to the M602T’s which is less than onetenth of that, one of the transistor’s
principal advantages over the valve
is confirmed: greatly reduced power
consumption when doing much the
same job.
And there’s a clue in the M602T’s
rear cover. It’s held in place with
screws. That suggests that you’re not
expected to remove it very often to replace the battery.
Special handling
The four knobs (tuning, on/off, volume and tone) are push-fits. Mine were
very hard to remove, so I used some
dial cord looped about the shaft and
re-looped to give four drawstrings. Be
aware that the metal rims on the knobs
are thin, and any attempt to lever under them will cause damage.
Output transistors TR5/6 are mounted in rubber grommets, effectively insulating them and providing even less
heatsinking effect than wiring them
onto tag strips and leaving the cases
unobstructed, in free air.
If this set is delivering its full output of 400mW for any substantial
time, that may cause significant heating of Q5/6, possibly leading to their
destruction.
So if testing for maximum output
with a continuous sinewave signal,
be sure to keep the test brief.
You can find additional information
on this set in the links below:
siliconchip.com.au/link/aau2
siliconchip.com.au/link/aau3
siliconchip.com.au/link/aau4
SC
A size comparison of the Healing M602T and previously described 404B.
Australia’s electronics magazine
October 2019 109
|