Wireless Set Canadian
No 52 Receiver
|
|
As you can see from the
picture below this receiver was made in Canada during WW2. |
 |
 |
|
I've never had one of
these sets before. The upper panel is fitted with Zeus fasteners
allowing the panel to be detached for access to the valves, which
appear to be plentiful in number.
There's no case so maybe this
example originally came from a complete set-up including power
supply and transmitter which would have shared a large case? |
 |
|
Below is the block diagram
plus schematic |
 |
 |
But before looking at the circuit
note the pin numbering of the valves below with black as standard
British numbering and red as used in the Canadian documentation
|
|
|
The thirteen valve line-up
(with equivalent codes) is as follows :-
V1A RF Amplifier, ARP3
V1C Mixer, ARP3
V1B Conversion oscillator, ARP3
V1D 1st IF amplifier, ARP3
V1E 2nd IF amplifier, ARP3
V2A Detector and AVC, ARDD1
V2B Noise limiter, ARDD1
V1F Heterodyne oscillator, ARP3
V1G 1st AF amplifier, ARP3
V1H 2nd AF amplifier, ARP3
V3A Crystal calibrator oscillator,
12SC7
V3B Crystal calibrator multivibrator,
12SC7
V3C Crystal calibrator harmonic
exaggerator, 12SC7 |
Valve details:-
Note that I'm quoting British
valve base numbers as per Wireless World Valve Data books, not
those given in the circuit diagram.
ARDD1=10D1=12Y4G=2D13C=CV1300
Heater 13.0 volts at 200mA
Max input 50 volts RMS and max rectified current 1mA
B5: 1=a', 2=a'', 3=h, 4=h, 5=k
ARP3=9D2=12VPA=CV1321=CV1106=VR106
Heater 13.0 volts at 200mA
Anode 250 volts and 10.5mA, ra=0.6Mohm
Screen 125 volts and 2.6mA
Grid -3 volts and gm=1.65 mA/V
Variable mu
B7: 1=Metalising, 2=a, 3=g3, 4=h, 5=h, 6=k, 7=g2, TC=g1
12SC7=CV540
Heater 12.6 volts at 150mA
Anode 250 volts and 2mA, ra=53Kohm
Grid -2 volts and gm=1.3mA/V
IO: 1=Metalising, 2=a'', 3=g'', 4=g', 5=a', 6=k, 7=h, 8=h |
The valves are generally 13
volt/0.2A heater types designed for use with a 12-volt battery
(= circa 13.2 volts).
The 12SC7 is a double triode
metal valve with an International Octal base. This has a 12.6
volt heater taking 150mA
The intermediate frequency (IF)
is 420KHz.
I should mention component identification..
tricky because wiring harnesses are generally used between valve
pins and resistors and condensers. |
 |
|
A couple more views before I
power up the receiver maybe for the first time in over 70 years?
Above the rear view showing
the power connector (top right) for frame-mounting with matching
transmitter etc. and below with front uppermost with loudspeaker
central. No nasty unreliable wax condensers. Where non-standard
resistor values were specified by the designers you'll see instances
of paralleling standard values, for example 600Kohm = 2 x 1.2Mohm
(below centre). |
 |
 |
These low voltage power
supplies are very useful as they allow you to crank up the current
limit to avoid any nasty explosions and smoke. With valve heaters
their resistance when cold is very low so the voltage stays low
and gradually increases as the heaters warm up. In this case
10 heaters at 200mA and 3 heaters at 150mA plus a couple of dial
lamps should be about 2.65 Amps but due to slight voltage drop
in the wiring of half a volt and the fact that the correct heater
voltage should be 13 volts for the 10 B7/B5 valves the measured
current is only 2.3Amps.
Both the main receiver and the
crystal calibrator are operating to draw this current. In order
to reduce battery drain in the field the calibrator can be switched
off, which is the setting I'll be using for most of the testing.
Below, the chassis with the
top section of the front panel detached.
I'm
using my Solartron HT power supply for testing. The front panel meter on the receiver reads
low at 200 volts with 250 volts running 80mA from a variable
HT supply.
The tuning dial is set to 80
metres and I can hear an SSB QSO coming in nice and clearly,
although I found the Het Tone control only works when fully anti-clockwise.
The valves top right (below)
are the RF amplifier, mixer and local oscillator. Access for
fault-finding is pretty good except the power connector will
need removing to get at the wiring to two of the three RF valves.
The set of nine RF trimmers and dust cores are nicely labelled
and arranged as are the IF trimmers. Top left are the all-metal
crystal calibrator valves. |
|
 |
|
The receiver came to life
with 250 volts HT (instead of no more than 150 volts) and running
80mA which is actually very stable but seems too high for comfort.
It's quite likely there are leaky condensers or bad resistors
to blame. The meter also reads low and suggests that resistors
have drifted high. Checking the valves at the meter gives the
following.. Note these valve readings are measured at the valve
cathodes except for V1b whose grid current is monitored. Figures
in brackets are from the handbook. True HT was 250 volts and
LT 12.5 volts.
HT 200(150), LT 9.2(12.5), V1a=2(2.3),
V1b=15(7.0), V1c=2(3.0), V1d=15+(11), V1e=15+(11), V1f=7.5(3.2),
V1g=0(2.3), V1h=13(9.6)
No doubt these readings will
provide some clues to explain the reason for the elevated HT?
First I'll need to check the meter sensitivity, then compare
the various cathode resistor etc values against those given in
the circuit. At first sight V1b, V1d, V1e and probably V1h are
drawing too much current with V1a, V1c and V1g too little, and
the resistor from the HT line looks to be more than 20% high
in value.
After a few minutes checking
resistor values, intially under the lower chassis, I found an
open circuit resistor marked 1Mohm (audio amplifier, V1g screen
grid resistor=R59b). I fitted a replacement and switched on the
receiver, gradually turning up the HT. At 172 volts it became
quite lively and 80m SSB QSOs resolved vey well. A check of the
metered points was now as follows (with the HT now reduced from
250 to 172 volts.. but still slightly higher than the nominal
150 volts):
HT 145(150), LT 9.2(12.5), V1a=1.5(2.3),
V1b=9.5(7.0), V1c=0(3.0), V1d=14(11), V1e=14.5+(11), V1f=5.5(3.2),
V1g=2.2(2.3), V1h=9(9.6)
Now V1c (the mixer) which previously
read 2.0 is now reading low at zero volts. Looking at the circuit
diagram the bad resistor appears to be the anode/screen grid
resistor R48h as one end of the new resistor is sitting at 167
volts and the other at 14.5 volts. A new 100Kohm resistor on
Tagboard AR restored the current through the mixer.
At this point I decided to check
the metering circuit. The meter used in the receiver has an FSD
of 0.5mA so the various resistors to produce correct readings
can be worked out and compared with the marked values and more
importantly their measured values. From some of the readings
above it's likely that several resistors may need changing. As
that point I haven't found a key to the resistors and condensers
so I halted proceedings in order to identify these and mark them
up on pictures, mainly tagstrips. |
Below.. parts identified.. an ongoing process
Note that, as was common in
much WW2 equipment, component numbering was in three parts: Type
of component (V, R, C, L, etc.), a number representing the physical
item (eg. 2, which might represent a resistor value in ohms,
tolerance and wattage) plus its circuit identifier (a, b, etc).
Another point to mention is that it was common in British-designed
receivers to use a minimum number of valve types. This meant
you could reduce the number of spare valves, and also manufacturers
could concentrate on producing many valve types in bulk.
As I identify the parts I'll
add the circuit identifiers. Most small components in this receiver
are fitted to tagboards and sometimes these are remote from other
associated circuit parts. The drawing below shows the location
of the various tagboards which are shown further down this page.
Note that several resistors
are hidden under decoupling condensers, for example R31a and
R20g on Tagboard BR. |
 |
|
Cct Ref |
Value |
Cct Ref |
Value |
Cct Ref |
Value |
Cct Ref |
Value |
|
C1 |
20pF |
C11 |
100pF |
C31 |
2pF |
C48 |
0.002uF |
|
C2 |
4-30pF var |
C20 |
0.01uF |
C34 |
80pF |
C49 |
25pF var |
|
C3 |
0.1uF |
C26 |
0.002uF |
C43 |
250pF |
C50 |
150pF |
|
C4 |
441pF var |
C27 |
500pF |
C44 |
1004pF |
C51 |
1pF |
|
C7 |
100pF var |
C29 |
350uF |
C45 |
1060pF |
|
|
|
C8 |
50pF |
C30 |
150pF |
C46 |
1349pF |
|
|
|
C10 |
12uF |
C31 |
0.01uF |
C47 |
7pF |
|
|
|
|
Cct Ref |
Value Ohms |
Cct Ref |
Value Ohms |
Cct Ref |
Value Ohms |
Cct Ref |
Value Ohms |
|
R4 |
8 |
R25 |
3K |
R43 |
25K |
R52 |
200K |
|
R13 |
300 |
R27 |
5K |
R44 |
30K |
R54 |
250K |
|
R14 |
300 var |
R28 |
5K |
R45 |
30K |
R55 |
300K |
|
R17 |
500 |
R29 |
5K var |
R47 |
50K |
R56 |
500K |
|
R19 |
600 |
R31 |
10K |
R48 |
100K |
R58 |
600K |
|
R20 |
1K |
R32 |
10K |
R50 |
100K var |
R59 |
1M |
|
R23 |
2K var |
R33 |
10K var |
R51 |
150K |
R60 |
4M |
|
 |
 |
 |
|
|
 |
 |
 |
 |
 |
 |
 |
Valves are not too fussy about
resistor values. Often the original values are specified as widely
as +/-20% but components (in this receiver dating back to 1944)
can drift way beyond this (very rarely downwards) and eventually
reach the point where circuits don't work very well.
The first fault was the excessive level
of HT needed to hear anything because the 1Mohm resistor feeding
an AF amplifier screen grid was open circuit. Then the mixer
suddenly ceased to draw much current and then I discovered the
BFO would only work when the meter switch selected the BFO valve.
The 1Kohm cathode resistor measured 4.6Kohm (this resistor was
shunted by the 0.5mA meter circuit when V1f was selected).
When the meter switch was set to LT
it read 9.2 volts instead of 12 volts because the monitor resistor
comprising a parallel pair of 60Kohm resistors measured 69 and
70Kohms. Similarly the HT read 40% less because the parallel
pair of 1.2KMohm resistors measured 1.5 and 1.7Mohms. When this
types of resistor fails it can also vary in resistance depending
on the voltage across it so removing it and measuring its resistance
with an ohmeter may not fully account for the low meter reading. |
|
|
Once the worst resistors
(a sample shown above) had been swapped the receiver worked surprisingly
well and was very stable reading 80m SSB. I noticed that the
selector switch that was supposed to allow AVC to operate had
no effect in either AM or CW settings so a further fault is waiting
to be discovered. V2a is used for AVC by rectifying the IF signal.
The other double diode at V2b as a noise limiter is working properly.
AVC is developed by one of V2a
diodes fed by C8a (50pF), However the AVC is what's termed "delayed
AVC" and this only operates once the detector diode is turned
on. The cathode of V2a is fed by the positive voltage across
the cathode resistor of V1h (R19a) so, if R19a is high in value
AVC may not operate. That voltage is smoothed by a 12uF condenser
(C10a) and these types of condenser rarely survive but it seems
the Canadian version made in 1943 is fine... a bit high in value
but a very low ESR... a little over half an ohm.
Now that more bad resistors
have been swapped here are the monitored readings.. Note: V1f
must be read in the CW setting, otherwise it will be zero, and
LT is slightly low due to cable losses. During fault-finding
and repairs, the HT voltage had been reduced progessively from
250, to 172 and finally 150 volts as the receiver performance
had improved.
HT 150(150), LT 12(12.5), V1a=1.0(2.3),
V1b=1.0(7.0), V1c=3.5(3.0), V1d=12(11), V1e=12(11), V1f=2.5(3.2),
V1g=1.5(2.3), V1h=7.5(9.6)
Still some anomalies.. including
V1a and V1b?
Note that on Tagboard KR you'll
see pairs of resistors in parallel. My guess is that these had
been selected to provide a better tolerance than standard resistors
because they are used for metering important voltages. For example
to provide the correct deflection on the 0.5mA meter (M1a) from
the 150 volt HT line you need 600Kohm. To get close to this value
a pair of 1.2Mohm resistors were used in parallel. Although R58a
is specified as 600 Kohm it's actually 2 x 1.2Mohm.
The same applies to the 10Kohm
metering resistors R32b and R32c which are each made from 2 x
20Kohm. |
|
|
|
The receiver is now working well
enough to check calibration etc. Using a portable receiver I
found the BFO was tuning from 406 to 411KHz, fully anti-clockwise
to fully clockwise which is odd as I can resolve LSB. I then
checked the IF. No luck injecting a signal at the aerial terminals
(see later for the reason) but I connected the signal generator
to the top cap of the mixer and found the IF was on the correct
frequency of 420KHz although I'm not sure if the Sharp setting
is exactly lined up to this. I checked each waveband and the
worst discrepancy was an error of 300KHz where a dial setting
of 8MHz responded to 8.3MHz on Range 2.
The next day I aligned the IF amplifiers
and the BFO to 420KHz. This entailed connecting an audio wattmeter
across the loudspeaker and setting my signal generator to 420KHz
with 80% amplitude modulation, then connecting the signal to
V1b, the mixer top cap. With the volume turned up and the minimum
RF signal I adjusted the sharp and wide IFT trimmers, reducing
the RF input as the audio level increased to minimise any effects
of AVC. The receiver sensitivity, which was already good, improved
dramatically and, after resetting the BFO to centre on 420KHz,
SSB was easier to resolve. The next step is to adjust the IF
response so that it's symmetrical with a good skirts. This used
to be done using a wobbulator (or even manually by plotting the
curves) but using a spectrum analyser it is very easy. After
that I can then align the RF front end so that frequency coverage
matches the dial readings and has reasonably flat response across
each of the three bands.
Below a view of the BFO signal
measured at V2a, Pin 3 (signal rectifier anode). Because the
BFO setting potentiometer does not give a linear tuning range
it's not ideal to set the Het.Tone knob to its centre position
at 420KHz. To get a balanced setting I first measured the tuning
limits, then calculated the average frequency and with the control
at its centre position adjusted the BFO trimmer to this frequency.
As you can see the BFO range is 4.5KHz (equating to +/- 2.25KHz)
so the pot was centred at a reading of about 419.08KHz. The end
result gave a range of 421.332-416.832=4.5KHz. With the Het.Tone
knob in its centre position the BFO frequency was a little over
419KHz. |
 |
 |
 |
|
BFO with setting at maximum
frequency |
BFO with setting at 420KHz |
BFO with setting at minimum
frequency |
|
|
Up to this point, the IF had been
adjusted using an audio wattmeter for maximum output. I've found
through experience that this method rarely gives the correct
results. Usually the centre of a tuned signal ends up either
higher or lower than the correct IF by as much as a couple of
KHz. To get a better idea of the IF response a
very high impedance probe feeding a spectrum analyser was
connected to V2a Pin 3 (signal rectifier anode) and the analyser
tracking generator connected via small condenser to V1b (mixer)
grid. The trimmers seen on the "IF
Tuning" picture above are grouped under Sharp and Flat.
These are very interactive and have extremely sensitive settings
but, by trial and error, the response curves for the two bandwidth
settings were centred on the 420KHz centre frequency. The fundamental
shape of these two curves will be governed by the components
originally fitted in manufacture and it's not sensible to check
these as it would almost certainly involve replacing both condensers
and coils as all will have drifted over the years from their
original characteristics. What we're looking at below is the
best performance of the 1944 receiver some 70 odd years after
manufacture. |
 |
Sharp Setting
As you can see, the overall
scan was 400 to 440KHz making each division 4KHz. The skirts
are 60dB down with the 3dB points about +/- 500Hz around 420KHz. |
|
 |
Flat Setting
In the broader setting the signal
was 4dB stronger. The skirts are better than 65dB down with the
3dB points about +/- 2KHz around 420KHz.
As a comparison my R1155 has
skirts 50dB down so the 52 receiver would seem to have a superior
performance, perhaps because of its more spacious interior layout?
Another reason is of course
degradation of parts, for example the dust cores used in the
R1155 IF transformer coils. |
|
|
It's worth looking at
the EMER in respect of aligning the IF strip. The method used
is to inject the test signal at various points in the circuit
and to peak the readings on the audio output. The method I used
initially was similar to this but when I checked the results
I found that adjustments had resulted in two humps in the wide
setting. This produces an undesirable effect when tuning across
a signal and with SSB may result in one or other peak being a
KHz or two off-centre with the BFO having to compensate. If the
BFO isn't able to track the signal, interference from adjacent
off-frequency transmissions may result. Using the tracking generator
got rid of the double hump by enabling the merging of the peaks
from the IF transformers into a single peak at the correct frequency.
Without a tracking generator it's also possible to inadvertently
tune the narrow and wide responses to two different frequencies
which is annoying because switching to the narrow setting to
improve readability means having to retune the receiver. |
|
|
|
Next, I tackled the RF alignment.
Trimmers and coil adjustments are logically
laid out so it was easy. If anyone decides to tackle one
of these receivers take care with the 3-30pF trimmers as they
can seize and you'll end up with a broken ceramic rotor. Try
rocking the slot right and left very gently using a snug fitting
flat screwdriver. Because these trimmers can be a bit jerky in
operation it can be tricky getting the oscillator trimmer to
peak up a signal, especially if AVC is active so my suggestion
is to readjust the tuning knob once the oscillator trimmer has
been set. This way you can adjust the RF stage trimmers to exactly
the same frequency as the oscillator setting. All three bands
aligned nicely to the tuning dial settings. I noticed the highest
band was slightly deafer then the other two. Something like 4uV
compared with 0.5uV to see a similar decent rise in the audio
wattmeter reading. Either the receiver never was as sensitive
on the highest band or something is wrong (possibly the RF amplifier
V1a because its cathode current as measured through R13a is close
to zero) ?
Before I got underway with alignment
I'd checked around V1a. The meter showed a very low cathode voltage
compared with say V1c and V1d so I measured the cathode resistor
and the resistor feeding g2. R13a, marked 300 ohms read 756 ohms
and screen grid resistor R54c marked 250Kohm measured 2Mohm so
I fitted a new 300 ohm and a new 240K ohm expecting to see a
change in performance. In fact there was no noticeable difference,
with V1a reading around 0.8 volts with the RF gain at max. On
the subject of RF gain control, the WS52 design is slightly odd
in that, rather than inserting extra resistance in the cathode
of V1a and V1d (the 1st IF amplifier), the RF gain pot is used
to select a voltage from zero to about plus 40 volts. As the
voltage is increased it makes the grids of the valves more negative
than their cathodes, reverse biasing and cutting off V1a and
V1d. Looking at V1a, with its cathode at ground via R13 and with
g2 and anode positive, current will flow through R13a developing
a voltage. As V1a and V1d both have a cathode resistor of 300
ohms you'd expect the voltage across their cathode resistors
to be much the same, however V1d has 12 volts and V1a has less
than 1 volt, even after renewing the bad resistors. Both V1a
and V1d have the same AVC voltage applied to their grids so what's
going on? Is V1a low on emission perhaps?
Before leaving alignment, I'll mention
the aerial circuit. There are two connectors on the front panel,
one of which is marked "AE" and the second I'd naturally
assumed was ground. When I started IF alignment I'd connected
the signal generator across these two connectors and discovered
what I'd thought was an astounding degree of attenuation of 420KHz
between the aerial and the IF amplifier, so instead connected
the 420KHz signal to the top cap of the mixer stage. Later, when
I started RF alignment I found I needed tens of millivolts to
hear anything in the loudspeaker, but when I disconnected the
ground connection the speaker blasted out a very loud tone. The
penny dropped and I found the two terminals both connect to the
same input point and no ground connection is provided next to
the aerial terminal. This is similar to the power connector at
the rear of the chassis where no ground connection is provided.
Clearly the overall frame into which the equipments mount deals
with the LT and HT negative returns, as well as any grounding
for the aerial. Because my receiver does not have an outer case
and isn't mounted in a frame and uses a bench power supply, I
lifted off the -1200 connection from its pin and instead wired
this pin to chassis enabling all the power connections to be
made by a single plug. Checking the circuit diagrams (I have
two) revealed the poorer quality circuit shows aerial connections
but the better quality circuit had strangely omitted any aerial
connection. Whoever drew this had forgotten to show the aerial
input terminals so I've now added them. |
|
|
|
Next.. I must tackle the crystal
calibrator because I can only just hear one or two very weak
Mc/s blips. Not too easy as the chassis needs to be detached
from the main receiver. |
|
|
|
pending.. watch this space |
|
|
