Plessey PR1551 Receiver
Feb 2025
|
|
I've shown here the latest
PR1551 (lower) compared with my other a PR155G (above).
These are very similar but the
lower one has a better specification in terms of the number of
modes and the rationalization of features from two to a single
knob.
|
|
PR155G
|
PR1551
Click
the picture to see the Service Manual for the (similar) PR1553
|
Serial Number WL5107
|
|
The receiver didn't come
with a mains lead but after searching through my collection of
such things I eventually found a military style mains lead whose
plug fitted. I plugged it into the PR1551 together with an aerial,
switched on and tuned to 198KHz. The dial lamp came on dimly
but there was complete silence from the speaker so I looked under
the lid and checked the voltage at a large smoothing capacitor.
This seemed to be OK as it had 20 odd volts present, so I checked
the various knobs and when CW was selected I heard a hiss from
the speaker so that's promising. I swapped the aerial for a signal
generator set to 198KHz but still nothing even with tens of mV.
Still set to CW, as that was the only mode I could hear a hiss
from the speaker, I twiddled the signal generator and was rewarded
with a heterodyne at 100KHz. This varied when the BFO was altered
so maybe the receiver has an RF or PLL fault?
Below is a block diagram of
the PR1551 which I discovered is electrically identical to the
PR1553 which uses a nixie display instead of a dial assembly.
Turning to AM mode and switching
the signal generator to AM resulted in a tone from the speaker
and proved the receiver seems to work from its third IF in CW
and AM and the BFO control does what you'd expect. The circuitry
before the 3rd IF amplifier has a problem and that circuitry
is pretty complicated.
My PR155G is set to one side
at present because it has a tricky problem with its phase lock
loop circuit. I did learn a fair bit about that circuitry and
a pound to a penny I'll find a similar puzzle with this receiver.
|
|
 |
|
As usual I'll need to
completely familiarise myself with the intricasies of the receiver.
Fortunately the design is not too dissimilar to that of the PR155G
on which I'm fairly clued up. In fact maybe some of the PR1551
modules may be the same as those in the earlier model?
Things might be a little awkward
because the only manual I can find on the Internet is that for
the PR1553. Some people report that the PR1551 uses identical
modules to those in the PR1553 but the latter's service manual
says that USB and DSB modes can be received yet the PR1551 front
panel clearly shows USB and LSB and so does the front panel of
the PR1553 shown in its service manual. Maybe not too surprising
though, because I found loads of errors in the manuals for the
PR155B and PR155G receivers. Plessey QA must have been slacking...
Looking at that pictures below..
where on earth does one start? Instead of an open chassis with
everything to hand this receiver is constructed from loads of
modules comprising one or more printed circuit boards carried
in metal cases. RF-wise these are interconnected by cables carrying
plugs and sockets. The latter are a mixture of types, of which
the majority are unfamiliar. Well, after discovering that the
receiver could pick up a 100KHz signal at the aerial socket,
was it able to receive signals at its other IFs? These are 10.7MHz
and 37.3MHz but try as I might with my signal generator set to
these frequencies nothing managed to get through, the receiver
being completely deaf at 10.7MHz or 37.3MHz.
|
|
 |
 |
|
Let's start with the good
news. The loudspeaker together with the audio stages and 100KHz
detector plus the BFO must be working for a 100KHz signal to
be received although I didn't see any meter movement so I need
to look at the circuit diagram and see if there's a clue to this.
Maybe there's insufficient signal or perhaps a fault in its circuit?
Maybe there's a power supply
issue? The drawing shown below brings together circuits from
various areas of the receiver. I'll make some checks and see
if all the voltages are OK, then I'll see if the voltages are
getting to all the modules.
I've
attempted to make the power supplies more understandable in the
drawing below. If you click this a larger PDF version will be
visible.
|
 |
|
Click the picture "PR1553
Circuit Diagram" below and you'll see an enlarged PDF of
the complete receiver which can be resized to check different
areas. You can see there's a large number of switches and as
the connections to these are mainly inaccessible due to the mechanical
design of the receiver it makes fault-finding tricky (this design
feature is the main difference between early equipments and modern
equipments which use microproceesors). The fact that those receiver
stages which are working, and which are using the neg 15 volt
supply rail, means that their switching circuits are OK. Later
I'll check the neg 15 volt connections to all the modules.
You can see in the large PDF
a 5 volt regulator driven by a separate 9 volt transformer winding.
There's also a 200 volt rail which I can discount as that's used
only by the digital display in the 1553 version and not in the
PR1551. There's also a 12 volt transformer winding marked 1MHz
oscillator (bottom left corner of the schematic).
|
 |
|
The meter amplifier uses
the neg 15 volt supply so if that's present at the correct voltage
I'd expect the meter to indicate something (which it doesn't).
I need to check the neg 15 volt supply, the 5 volt regulator
and the 12 volt supply hidden in the 1 MHz oscillator box are
all working and distributed properly.
I'm now going to comment on
the quality of the Plessey documentation. It's pretty awful.
Take for example Page 110 in the service manual which shows "Fig
23 Regulator: Circuit and layout" plus the adjacent "Fig.
23 Regulator". There are two regulators.. one for neg 15
volts and one for plus 5 volts. The latter being on Page 114
which does include the term "5V". The parts list for
Fig 23 has VT2, VT3 and VT4 but the schematic also shows VT1
(a second OC35 but where is it?) so there are two OC35s. Pin
7 on the Regulator box is the neg 15 output. I can see a power
transistor mounted on the rear panel, I assume it's the regulator
output transistor VT1 because "VT1" is marked near
to it, although you can see the neg 15 volt regulator circuit
board is fitted in the centre of the drawing labelled "Top"
next to C2 the smoothing capacitor for the supply voltage. What
is going to be ultra-confusing is the use of a negative supply
rail (common in designs using germanium PNP transistors). The
5 volt regulator is top left on the drawing labelled "Bottom".
Also, squirrelled away is another
power supply.. this one for 12 volts and located in the 1 MHz
oscillator module. This provides 12 volts for the crystal oven
plus around 5 volts for the crystal oscillator.
|
|
 |
|
If I knew the history
of the receiver it would be much easier to fault-find, but maybe
the way forward after checking the PSU voltages are present and
correct is to record the voltages on each of the countless modules?
Above you can see the module wiring and below is a table of measurements
generally reading pins from left to right with receiver turned
to CW. The meter amplifier is that unshielded board with haphazard
connections visible below the centre of the chassis with those
two blue capacitors. |
|
|
Volts on pin |
Volts on pin |
Volts on pin |
Volts on pin |
Volts on pin |
Volts on pin |
Volts on pin |
Volts on pin |
See picture above |
|
MODULE |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
NOTES |
|
AUDIO PCB A |
0
E
|
0
VOL CONT
|
0
E
|
0
LS
|
- |
-14.09
PSU
|
- |
- |
MODULE 9 |
|
AUDIO PCB B |
0
E
|
0
FROM DET
|
0
E
|
0
LINE OUT
|
- |
-15.39
PSU
|
- |
- |
MODULE 9 |
|
AGC AMP |
0
E
|
-14.1
PSU
|
-3.2
AGC OUT
|
-14.11
PSU
|
-3.4
IF GAIN
|
6.0
AGC DECAY
|
- |
- |
MODULE 8 |
|
100KHZ |
0
E
|
-4.7
REF
|
-3.2
AGC
|
-14.11
PSU
|
- |
- |
- |
- |
MODULE 7
LOWER
|
|
DETECTOR |
- |
- |
0
E
|
-14.11
SSB/CW
|
0
AM
|
-14.11
AM
|
- |
- |
MODULE 7
UPPER
|
|
ISO AMP |
-15.4
PSU
|
- |
- |
- |
- |
- |
- |
- |
- |
|
1 MHZ OSC |
+7.1
AC FEED
|
+7.2
AC FEED
|
0
E
|
- |
- |
- |
- |
- |
NEEDS CHECKING |
|
-15V REGULATOR |
-11.3
SPECIAL
|
-24
SPECIAL
|
-11
SPECIAL
|
-15
SPECIAL
|
-11
SPECIAL
|
- |
- |
- |
SEE PSU DRG |
|
METER AMP |
-4.1
SPECIAL
|
-1.7
SPECIAL
|
-15.38
SPECIAL
|
-15.38
SPECIAL
|
-3.2
SPECIAL
|
-3.2
SPECIAL
|
-0.1
SPECIAL
|
-15.39
SPECIAL
|
SEE DRAWING |
|
METER AMP |
-15.39
SPECIAL
|
-15.39
SPECIAL
|
-12.27
SPECIAL
|
0
SPECIAL
|
0
SPECIAL
|
- |
- |
- |
SEE DRAWING |
|
INTEGRATOR |
0
E
|
-0.19
R22
|
+3.6
GUARD
|
-0.04
GUARD
|
-0.6
LOCK
|
-0.04
CONTROL
|
-4.6
REF
|
-15.45
PSU
|
MODULE 13 |
|
10.6/10.8MHZ OSC |
0
E
|
-7.9
USB
|
0
LSB
|
-15.39
PSU
|
- |
- |
- |
- |
MODULE 11 |
|
5-volt regulator
+ 1MHZ OSC
|
0
E
|
- |
-13.4
PSU
|
-13.4
PSU
|
0.19
?
|
- |
- |
- |
MODULE 10 |
|
10.7MHZ AMP (2ND IF)
3RD MIXER
|
0 |
-4.7
REF
|
-3.2
AGC
|
-15.4
PSU
|
- |
- |
- |
- |
MODULE 5 |
|
37.3MHZ AMP (1ST IF) |
0
E
|
-4.6
REF
|
-4.7
AGC
|
-15.42
PSU
|
- |
- |
- |
- |
MODULE 4 |
|
2ND MIXER |
0
E
|
-4.6
REF
|
-3.2
AGC
|
-15.43
PSU
|
- |
- |
- |
- |
MODULE 4 |
|
1ST LO |
0
E
|
-6.0
PCB J
|
-14.38
PCB J
|
-15.43
PSU
|
- |
- |
- |
- |
MODULE 3 |
|
RF AMP |
0
E
|
0 |
-3.2
AGC
|
-14.83
PSU
|
- |
- |
- |
- |
MODULE 1 |
|
|
A nice touch by the designers
was to assign the same module pin number to common connections
so you can see the neg 15 volt rail generally connects to Pin
4, "Ref" to Pin 2 and where appropriate AGC to Pin
3.
At first glance at the readings
above there are a few things to check further. The Crystal Oscillator
Module transformer input looks odd and Module 7 mode feeds might
be wrong as CW is selected, not AM. I spotted quite a few interesting
points. From pencilled marks someone has been looking for a fault
and wedged between two modules I found a plastic bag containing
a 10.6MHz crystal. Several wires have been detached and resoldered
and a couple of pink wires have been cut off. Was the fault ever
found and fixed or was the receiver just put on one side and
eventually disposed of? A few pictures below showing evidence
of someone else being busy.
You can also see an added cartridge
fuse in a holder at the top left corner in the picture above.
It's wired in series with the supply voltage to Module 10. Was
this a troublesome area?
|
|
 |
Those pencilled markings and
the new crystal below seem to indicate a problem area? Read on
to see my theory.
|
 |
 |
|
Here's the series pass
transistor for the 5 volt regulator under the blue insulator
.
|
There's the series pass transistor
marked "VT1" for the neg 15 volt regulator hidden away.
It's clearly been quite hot. 
|
|
|
As the tail end of the
receiver seems to be working I'm wondering if I can open some
specific coax cable links (seen in the picture below) and either
monitor their signal or else inject a signal from my TinySA. In
order to do this I need to identify all the cables. |
|
 |
|
The first check is to
see if Module 11 (10.6MHz or 10.8MHz Generator) is producing
either 10.6MHz or 10.8MHz for the 10.7MHz amplifier and 3rd mixer
carried in Module 5. One of these two signals will set the output
of Module 5 to 100KHz where its amplified and detected.
I opened the connectors between
Module 5 and Module 11 and found a comb of signals coming from
module 11. With a span setting of 5MHz and a centre frequency
of 10.7MHz each division is 500KHz making the signals harmonics
of 1MHz. This means that Module 5 isn't producing the correct
signals of either 10.6 or 10.8MHz, instead typically 11MHz.
Switching across the different
modes didn't alter the frequencies of the signals.
I plugged back the cables and
opened the connectors linking Module 4 (37.3MHz Amplifier) and
Module 5. I then injected a 10.7MHz AM signal and found the receiver
worked very well on the AM setting with a heterodyne on the CW
setting. There was nothing heard on the two SSB settings. CW
of course uses the BFO and this worked OK.
I'm beginning to suspect a longstanding
problem on SSB because of the clues in the pictures with pencil
markings above.
|
 |
|
 |
Because AM from the 10.7MHz
input on Module 5 is working proving the 10.7MHz amplifier and
the 3rd mixer are OK I'll leave SSB for the moment and press
ahead and check the 37.3MHz amplifier and the 2nd mixer.
There should be a 48MHz signal
from Module 10 the spectrum generator at cable number 12 and
sure enough (picture left) I found this present at 21dB and 25dB
stronger than adjacent comb signals.
I reconnected the link and injected
a 37.3MHz AM signal at -7dBm into cable number 11 feeding Module
4 from Module 2 and switched to AM.
Absolutely nothing was heard
so it looks like Module 4 is faulty as it's correctly receiving
the 48MHz sgnal from Module 10 and a 37.3MHz test signal. These
should produce 48-37.3=10.7MHz.
|
|
|
I'll remove Module 4 and
see what's going on. The first thing I noticed from scruffy wiring
(right) was that someone had already removed and replaced Module
4 so I guess I should check it was wired back correctly (it was).
I checked the input coax and measured 100ohms from the centre
pin to ground which is correct. Also the voltages on the pins
from the table above are about right.
Module 4 circuit and layout
is reproduced below.
.
|
 |
|
 |
|
Pretty obvious below is
the home-brew addition of three coils in a metal can. These have
numerous tuning capacitors attached and it probably represents
a band-pass filter. I can see that it's used to filter the 48MHz
signal from VT7 replacing the simple filter C34/R34 feeding the
balanced mixer. Once the module is working I can see if the filter
is OK.
Under the ferrite bead is a
zener load resistor which is badly burnt. This may be due to
leakage on the neg 15 volt line or a mishap in the past. The
zener is used to reduce the supply voltage as the ZTX320 transistors
have a maximum Vce of 15 volts.
|
Reading about work carried out
by CT7AWB it seems the Mullard striped 10nF capacitors (at least
in his example) are prone to lose capacity and, as over 20 are
used in Module 4, I'm considering swapping the lot, but I'll
not do this just yet
All the resistors and transistors
seem to be OK and, as is apparent, many originals have already
been replaced.
|
 |
 |
|
|
The module has been modified
by the addition of a 48MHz bandpass filter and I decided to see
if this was connected with the receivers complete deafness. Using
my TinySA I discovered that the input to the filter measured
around 0 dBm and the output at best could be peaked to -13dBm
which represents a significant loss (13dB). I decided to eliminate
the filter (at least until I got everything working) by wiring
in its place an original 47 ohm series resistor and a shunt 18pF
to ground.
Unfortunately this change didn't
really have much effect, so I decided to feed a test signal into
the cable leaving the 1st RF amplifier to see if the latter was
faulty. Again this didn't change matters one iota so I refitted
the cable and disconnected the local oscillator feed to the 1st
mixer but still no change. Then I looked at the local oscillator
and found it was one of two types of signal depending on the
MHz knob setting. One had lots of signals and the other around
3 or 4 at different amplitudes. I noticed that switching the
MHz knob backwards and forwards resulted in what I judged to
be a correct pattern of 3 or 4 signals at some settings, or an
incorrect pattern of many signals. I set it to 24MHz giving an
example of a "correct" pattern. I then connected my
TinySA set to precisely 24MHz with 500Hz AM and -7dBm to the
aerial socket. With the receiver set to CW I could hear the faintest
audio tone from the speaker. I proved this was correctly being
received by either tuning the receiver up and down by a few KHz
or tuning the TinySA. So there's an RF "blockage" in
a fundamentally working receiver. The way the local oscillator
behaves leads me to believe the phase locked loop isn't set up
properly but just happens to work on a few MHz settings, one
of which is 24MHz.
|
|
 |
Setting the receiver tuning
to precisely 24MHz with the megahertz knob set to "24"
gave me this local oscillator pattern of signals.
The frequency reading for the
strongest signal is 61.35MHz. The others are harmonics circa
122.7Mz, 184MHz, 245.4MHz and 307MHz.
61.35MHz-37.3MHz= 24.05MHz
The receiver is tuned to 24MHz
The local oscillator will not
be a pure sinewave signal which might explain the discrepancy?
|
|
|
Testing modules outside
their case and detached from the receiver as shown here is messy
but quite satisfactory because RF connections are made by cables
with only DC connections made via the rear connectors. This is
Module 4 being tested.
I ended up unwiring the filter
in the tinplate case and putting back a 47 ohm resistor with
an 18pF decoupler.
|
 |
|
|
As a break from puzzling
RF problems I tackled the faulty meter circuit. This was an easy
fix. The meter wasn't working in AF or RF so I measured the voltage
across R4 which turned out to be zero volts (ie. VT1 wasn't drawing
current) I removed the switch and was about to cut the wires
off and replace it but then decided to try switch cleaner because
the switch body had a pair of holes either side. It worked a
treat and now the meter works. Alas this fix failed later so
I fitted a brand new switch. |
 |
|
 |
|
|
Now, back to RF fault
finding on the "dead" receiver... There must one of
three possibilities viz. an open circuit or a dead short in a
signal path or a missing local oscillator feeding one of the
mixers (for example the 10.6/10.8 oscillator).
I decided to look first at Module
11, the 10.6/10.8 MHZ generator module. I checked the selection
voltages on the base of the module and found the following (table
opposite).
The values of the lower voltage
(-7.96V) is dependent on resistors connecting Pins 2 and 3.
Neg 15 volts is applied to one
of the two pins (USB or LSB) with one selecting 10.8MHz and the
other 10.6MHz. The resistors and potentiometer connected to the
pins are designed to set a varicap diode, D8, to tune and peak
two specific 1MHz harmonics in a comb of VHF signals.
|
Module 11
|
MODE |
PIN 2 |
PIN 3 |
|
STANDBY |
0 |
0 |
|
USB |
-15.4V |
-11.3V |
|
LSB |
-7.96V |
-15.4 |
|
CW |
-7.95V |
-15.4V |
|
AM |
-7.95V |
-15.4V |
|
F (FSK) |
-15.4V |
-7.95V |
|
|
|
Selection of 10.6MHz is
made by peaking the 53 harmonic from the comb of 50MHz'ish signals
(derived from the 1MHz oscillator) or 10.8MHz from peaking the
54MHz harmonic via a varicap diode driven by a voltage at Pin
2 or Pin 3 (shown in the table above) then passing that signal
through a divide by five circuit.
I looked specifically for a
10.6MHz or a 10.8MHz at the output socket and there was none.
Looking at the input cable to
module 11, I found (correctly) a comb of signals representing
harmonics of 1MHz peaking at around 50MHz which is about right.
|
 |
|
|
Unplugging the coax cables
labelled "15", I plugged the TinySA, set to 10.6MHz,
into the local oscillator input of the third mixer, set the MHz
selector to zero, put a few yards of wire into the aerial socket
and tuned to 198Hz and there was Radio 4 loud and clear... success!
I altered the signal strength at the TinySA and found -20dBm
was the perfect level into the mixer module.
So the problem is in Module
11 (pictured below). Just 15 transistors, 9 diodes, 42 resistors,
18 capacitors, 7 coils and 2 pots not to mention any messing
around by the last owner!! The varactor diode D8 a BA110 might
be awkward to replace if it's failed (but later tests proved
it was working OK).
In fact my worst fears were
realised (having already seen the extra bandpass filter on module
4) when I extracted the board from its case. Someone had completely
messed up Module 11. I say messed up but hopefully once I start
to examine it in detail perhaps it's not as bad as I'd first
thought.
Pictured below is the original
circuit of Module 11. Note.. I had to remove a sizeable metal
screen to see the upper surface and there's a use-by date in
1994 on the back of the tinplate sheet.
Anyway.. the thing isn't working!
Could it be as simple as a short to either the top or bottom
metal shield because the spacing is pretty tiny.
|
|
 |
|
|
Below left is the top
view of module 11 circuit board with its added screen removed.
Why a screen? Instability? Yes!
There's also an added screen
underneath the board (scroll down to see this), and what's that
home-made copper based circuit top right and what else has been
going on??
|
|
|
Below, the modified Module
11... |
Below, the layout of Module
11 in the service manual.... |
 |
 |
|
|
Unfortunately I can't
find a picture of Module 11 other than the drawing above right
and things may not be as dire as I'd imagined. I did a quick
check of the visible diodes and a few transistors and all seemed
to be OK. I noted that most of the (original) divide by five
circuit transistors are specially selected type BSY95A coded
"XK1099". This "special selection exercise"
was a common practice when I worked for Plessey in the 1960s
due to wide variations in the performance of standard off-the-shelf
transistors.
The chap that modified this
board has replaced all of the type XK1099 in the "divide
by five" ring counter with type BF224. These have much higher
Hfe.. perhaps the key parameter in the XK1099? The five 2.7 volt
zener diodes are still present but he's added an extra ten BAT43
Schottky diodes. To handle these changes he'd added that metal
connection plate.
Another change is the addition
of a tiny RF transistor in front of the divide by five counter.
Its number is missing but it looks something like a BFP90A (a
microwave NPN transistor).
There are a couple of options,
bearing in mind it's not working.. Try and get it working "as
is" or revert to the original design. My best bet is to
power it up on the bench, connect a signal generator tuned to
53MHz and see the results.
Below, the underside of the
board. I'll later detach that metal screen.
|
|
 |
|
Where do I start? I set
the board up on my bench and connected a power supply set to
neg 15 volts with a 150mA current limit. I can't readily generate
a comb test signal so I decided to test using a 53MHz then a
54MHz single signal.
I injected a signal of 53MHz,
then connected my spectrum analyser and looked for a 10.6MHz
signal at the output. Very odd things happened and I eventually
realised the thing was self-oscillating at frequencies around
53MHz depending on the settings of the cores in the various coils
and the setting of potentiometer RV2. This probably explains
the reason for the extensive metal screens. Now.. had the modified
board ever worked or was it always temperamental or maybe it
had just failed? Nothing seems to be faulty so I'm putting the
problem down to the high gain amplifier transistor fitted in
the tail end of the input circuit. The reason being that the
original circuit has a single amplifier stage with three emitter
followers feeding the divide-by-five counter. I reckon this would
have been pretty stable. The new modified circuit arrangement
has the input and output physically close together and will be
inherently unstable due to undesired feedback.
|
|
|
Right... with no signal
input I found I could tune coil L2 to produce an relatively unstable
output of 10.7MHz. I tried reducing the PSU voltage. This cleaned
up the shape of the output signal but didn't prevent self-oscillation.
In fact, connecting a signal generator at the input and changing
either frequency or level had no effect on anything. Presumably
the input circuit circuitry is oscillating at 53MHz (or to whatever
its coils are tuned) and the divide-by-five is doing its job.
I now need to examine the modified circuit and see how I can
re-establish the original design. |
|
 |
|
|
I removed the tiny added
microwave transistor and much to my surprise, firstly it didn't
have any base bias resistors and secondly it tested as bad. The
transistor has no markings other than a purple dot and two of
its four legs are a common emitter connection. The emitter is
wired via a 100 ohm resistor to neg 15 volts but the base is
wired to only a 330pF capacitor and as I've mentioned, no sign
of bias resistors... oops. Base to emitter resistance is 320
ohms and, perhaps because it still has some residual transistor
characteristics, it is still capable of some limited degree of
amplification because of internal biasing.
I'm guessing the device is a
BFP90A or something similar and has failed perhaps from an excessive
RF voltage.
See the modified circuitry with
the extra transistor (as found) below.
Is that the problem then? Did
the phantom modifier forget to add base bias resistors and inadvertently
damage the transistor when testing then gave up and set the receiver
on one side?
Just in case the new circuit
was viable I fitted a ZTX314 in place of the BFP90A but added
bias resistors to give about 10mA of emitter current then tried
testing again. But no... things were really bad (as I'd expected)
so I gave up and shall revert to the original front-end design.
Note that the original design uses three emitter followers after
the first amplifying stage so should be tame and not oscillate.
Below, the modified circuit-as
found, but on test it's a self-generating 53 ish MHz signal being
later divided by five.
During abortive testing I shifted
the selection voltage from Pin 2 to Pin 3 and the 10.7MHz oscillation
frequency (shown above) did shift about 200KHz proving with relief
that the varicap diode is correctly altering the tuning of L2.
|
|
 |
 |
|
Can anyone do better than
my guess of a BFP90A?
The dot is purple in colour
and the base lead has a 45 degree end.
|
|
|
|
At this point I'll recap
and try and figure out what's happened here. It began around
1994 when the last user of the receiver decided to either fix
an annoying fault or make it work better or both. Setting aside
the earlier finding.. that extra bandpass filter in Module 4
which is still a puzzle.
At some point he removed the
10.6/10.8MHz generator and decided to swap the "divide by
five" circuit for a slightly different one. Maybe he found
one or more bad XK1099s but, in the event he removed all ten
XK1099 transistors and fitted BF224s, leaving in place the 2.7
volt zeners, but adding a set of Schottky diodes presumably to
improve the ring counter performance? Most of the old resistors
were replaced and an extra stage of amplification using a microwave
transistor was added to the 53MHz-54MHz front end.
The receiver failed to work
but gave some rather puzzling results that without a spectrum
analyser would have been mystifying. What was actually happening
was that the front end of Module 11 was oscillating at around
53MHz or so and completely independent of the 1MHz oscillator.
The oscillation was feeding the ring counter and giving a divide
by five output which was tunable by twiddling coils and pots.
However the output was not locked to the 1MHz crystal oscillator.
This problem may have been realised so extra metal shielding
was added.
I guess the receiver might have
received signals but must have been unstable or just not worked
and then realising the root cause he bought a 10.6MHz crystal
with which to build a new generator, perhaps just providing AM
and CW plus LSB operation. This would have at least got the receiver
working although certainly not to spec. The new crystal was wedged
in its plastic bag between two of the modules. Alas (or thankfully
because the old board is rescuable) this major design was never
implemented. Interestingly, when I looked at the microwave amplifier
stage and noticed he'd forgotten to add bias resistors to its
base this error had brought the whole exercise to a puzzling
conclusion.
|
|
 |
|
Having worked out the
modifications needed to revert to original design, I removed
the most obvious, consisting of the small copper-clad board carrying
the extra amplifier transistor with its supporting components
(see earlier), then looked at the first four transistors. I found
a 2.2Kohm resistor in parallel with the first tuning coil and
R3 (5.6Kohm) then I spotted what looked like a 910ohm resistor
at R5 which the circuit diagram indicated as 470ohms. Maybe the
2.2Kohm was added to reduce the gain of VT1 and the 910 ohm to
reduce the current through VT2? I planned to remove these two
resistors and to fit a 470ohm for R5. Then I spotted a second
910ohm resistor at R31 and at R6 a rather odd resistor.. it was
then I realised the yellow bands had faded to near white and
purple to near brown so the resistors were OK so I'll just remove
that 2.2Kohm.
Surely the original circuit
layout and design would have been stable but maybe amplifying
a comb of RF signals results in a different circuit behaviour
from just a single 53MHz narrow band signal? More testing is
needed.
Below.. testing outside the
module case is quite straightforward because the wiring carries
only DC voltages although some fine tuning may be required later
due to the proximity of the grounded case to RF circuitry.
|
|
 |
|
Further testing showed
that the circuit was still prone to oscillation at around 55MHz
but I decided to try it in the receiver. I twiddled the three
53/54MHz tuning coils fairly randomly with the receiver tuned
to 200KHz but all I got was random squeaks, whistles and hash
but then I remembered that when I first started to test the receiver
I'd heard regular faint clicks. That noise represents lost PLL
lock and then recalled that tuning to 24MHz the clicks in background
noise had stopped. I retuned to 24MHz and connected my TinySA
to the aerial socket similarly tuned to 24MHz (AM) and twiddled
the three coils one again. This time a 10.6MHz signal locked
in place and, by carefully adjusting the coils together with
the setting of RV2, I managed to peak the 24MHz AM signal.
I then retuned to increasing
frequencies and discovered I could hear all the way up to 29MHz
and then decreasing from 24MHz I could hear down to 10MHz. Below
10MHz the receiver lost lock and just ticked quietly.. I guess
its time to re-box Module 11 and continue fault-finding. That
short red wire at 45 degrees is a test point that sticks out
the back of the module and should indicate whether Module 11
output is either 10.6MHz or 10.8MHz.
I refitted Module 11 and the
receiver picked up the 24MHz signal (but with a very strong heterodyne...
which needs investigating) Tapping the underside of the chassis
fairly hard killed the received signal until the power was switched
off and back on when it worked again. I'd also noted from earlier
that the LSB and USB settings sometimes gave hash in the speaker
and usually not so I think I'll tackle that next. Also loss of
lock is a problem. That area is covered by three circuit boards
mounted in a large metal case. I removed the lid and tapped these
boards but this failed to result in anything obvious so I can
rule out a bad connection in that area.
|
|
Turning to the SSB deafnesss.
The mode switch might be resonsible.
This has four 2-pole 8-way switch wafers which are labelled A,
B, C and D. Each one is identified by either F (front) and B
(back) as table below.
|
SWITCH |
FUNCTION |
NOTES |
|
AF |
Waveform generator |
Module 14 |
|
AB |
RF Amplifier |
Turret-1 |
|
BF |
10.6/10.8MHz Oscillator |
Module 11 |
|
BB |
Fixed/Variable BFO/100KHz Amp/Detectors |
Module 7 |
|
CF |
100KHz Amp/Detectors mode |
Module 7 |
|
CB |
Fixed BFO 5 volt supply voltage |
BFO-fixed |
|
DF |
Waveform Generator neg 15 voltage |
Module 14 |
|
DB |
PR1553 display |
Not used in PR1551 |
 |
 |
The switching for AM/CW/SSB
is carried out quite cleverly by S1BB and S1CF but I noticed
an error in the overall circuit diagram. S1BB carries a signal
to Cable 17 into the detectors on Module 7. For example the "Fixed
BFO" is used only for "F" and is shown correctly.
CW uses the BFO output whilst CAL and SSB use an output from
a 100KHz source. AM is carried to the AM detector diode but SSB
goes via a transformer to a product detector (also used by CW).
The AM detector is only powered in the AM setting via S1CF whilst
CW and SSB have their own power supply feed activating the product
detector.
S1BB, left, here shows the two
SSB settings for Cable 17 switched to the main neg 15 volt regulator
output. The blobs are on the wrong wire crossings and there are
two links shown in error. The amended circuit is shown right
and "X" goes to Module 14 picking up 100KHz as
well as a feed to the 5 volt rail via 270 ohms (see later).
It's those dratted technical
authors and Plessey QA slacking again!
|
|
|
|
I'm pretty sure the last
dabbler (I'm the latest) wouldn't have noticed the error and,
if he had, the wiring change would have been very tricky as the
switch is inaccessable without lots of effort. There's a simple
way of checking. I unplugged Cable 17 and looked at the switch
side. There are similar voltage readings of neg 2.8 volts in
CAL and USB settings and looking at the cable from Module 7 I
can see neg 15 volts in S/BY and neg 12.6 volts in A/M. Both
these latter readings indicate that one or both electrolytic
capacitors C22 & C23 (47uf) may have failed. This voltage
would only be present when USB or CW is selected unless there's
a fault or current leak in Module 7 (in fact I added a 470 ohm
resistor to see if the voltage was due to a high impedance current
leak.. and it was... so the capacitors are probably serviceable).
That other anomaly of neg 2.8 volts in CAL & SSB must be
a leak coming from the 100KHz source.
That source is said in the PR1553
manual to be Module 14, the "Waveform Generator" and
that module uses lots of early "54 series" i/cs on
three circuit boards and is not easy to test. I've added pictures
of these below.
|
|
BOARD A
|
BOARD B
|
BOARD C
|
These three circuit boards
are carried in Module 14 which is used in both the PR1551 and
PR1553 but in different configurations to include in the latter
digital display drive signals.
Because of the error in the
main circuit diagram for the receiver it's not a straightforward
exercise to figure out from where the 100KHz local oscillator
for the product detector in Module 7 comes from. If I open the
cable heading towards the mode switch I can see a DC voltage
is present which is from one of the ML7 buffers fed through a
potentiometer R8 & R9 halving the output. The voltage is
only visible on a high impedance grounded voltmeter because of
the diode D1.
Interestingly you can see evidence
of the intended user (military) from the coding on the fifteen
TTL chips (ie. 54 series rather than the usual 74 series). In
fact the reliability of the 54 series i/cs should be much greater
than the much cheaper 74 series.
Looking at the circuit diagam
below, the 1MHz signal from the 1MHz Oscillator module feeds
VT1 (emitter follower) to ML8 Pin1 (SN5490N = decade counter)
and out at Pin 12. This passes to Board C to ML7 Pin11 (SN5400N
quad buffer) and out at Pin13 from where its filtered by L1 &
C4.
R8 and R9 halve the DC voltage
from ML7 output and the sinewave resulting from L1 & C4 is
fed via D1 to the carrier reinsertion feed for SSB and CAL. A
bit puzzling is that D1 appears to be biased off.
|
|
 |
|
For the second time I
removed Module 14. The first time I'd checked by inspecting everything
then checking D1 and the BSY95s.No mean feat as 16 legs need
to be unsoldered, two coax leads unplugged and two internal metal
screens detached (4 x 46BA plus 4 x 8BA screws, washers and spacers).
This time I'd discovered that although the 1MHz signal went in
the 100KHz signal didn't emerge and as the circuitry above was
so simple the fault had to be a bad 5400 or a bad 5490.
As I looked at the three boards
an idea suggested itself. I had measured 5 volts at two of the
16 pins and of course plenty of grounds but I hadn't checked
the two chips in question were connected to their 5-volt power
supply. The 5490 chip is peculiar becase Pin 5 is Vcc and Pin
10 is ground. A quick check revealed the 5 volt feed was fine
but as I buzzed the ground connection I noticed that instead
of the meter reading zero it read 17 ohms. Each of the three
boards has a decent amount of ground copper, but Board C uses
a post and a black wire connecting its ground to Board B via
a second post. As I wiggled the post on Board B the 17 ohms dropped
to zero. Early on in proceedings the SSB modes worked but then
didn't again so is this problem merely a dry joint? I'll power
the boards from 5 volts and see if a 100KHz signal emerges. But
first I'll resolder all those posts!
|
 |
|
|
No luck finding that elusive
100KHz signal so I removed and testing ML9 and ML7 but and both
tested OK... oops I should have tested ML8 not ML9. I checked
it and it was also fine.
I think I've made a breakthrough.
I was checking the 1MHz signal from the 1MHz Oscillator module
and hadn't really bothered much about its amplitude, but I then
noticed it was around -20dBm which is 22mV (into 50 ohms) but
the signal passes through only an emitter follower on Board B
(above) so will have zero effect on the SN5490N. I'd be looking
for say a couple of volts so no wonder the 100KHz signal appears
to be missing!
Time to extract the Oscillator
module and see what's going on. Its solder links are hidden beneath
a pair of coax cables and when I pushed them clear I noticed
an open link. There's also an open link at one end but that's
for the fuse the last chap added. Now things are beginning to
make sense. I soldered the link and checked the 1MHz signal again
and this time it measured about +3dBm but it wasn't stable and
soon dropped to 1.7dBm. Clearly I need to fault-find the Oscillator
module. This incorporates five main areas.. a power supply for
a crystal oven, the oven itself. A crystal circuit operating
at 10MHz and followed by a decade counter for producing 1MHz.
Also a second power supply and finally a buffer amplifier for
1MHz. That missing link (L4 on the main chassis) was affecting
the PSU. But what exactly is the effect of leaving L14 open?
I found something surprising. Below are the circuits for the
module.
|
|
 |
|
The oscillator circuit,
above, unlike most modules isn't easy to work on but I did find
with some difficulty that C2 (100uF) was open circuit (in fact
this capacitor isn't needed). Oddly there's a 4.7 volt zener
for the supply to the decade counter however, I'm not too familiar
with 54 series specs and it seems they're guaranteed to work
with 4.5 volts (74 series need 4.75 volts). The output from VT3
should be around 4 volts with some reduction in sinewave terms.
Now, the (complicated) power
supply..
|
|
 |
|
The Plessey designers made removal
of the modules relatively easy. Each module circuit board has
a set of long pins which are soldered to tinned tags which can
slide backwards when unsoldering to free the pins. So, after
unsoldering say 8 pins the circuit board can be withdrawn after
unplugging coax cables and removing a pair of 6BA nuts.
If the power supply above was
in poor condition maybe VT7 isn't producing a decent output?
Why did the previous dabbler
add a fuse (right)?
The open link to L4 was lurking
under those two black coax cables., but it turned out the link
wasn't needed because the whole module circuit had been redesigned!
Below, compare the layouts of
the original (left) and modified (right).
Basically VT1/VT3/VT5/VT6 are
replaced with an LM317HVK which has a current limit of 1.5A to
2.2A. The old design used a 2N3713 with a 10A rating but I prefer
the new design so I'll leave this in place.
|
 |
|
|
|
|
A couple of days later and I
was still struggling to find the missing carrier insertion signals
for SSB. I'd been aware of an intermittent problem with my TinySA.
The amplitude readings were miles out and could be varied by
putting pressure on the input/output connector. I have a vague
inkling the rear of the connector is shorting one of the multi-layer
planes where they extend to the edge of the printed circuit board.
I'll leave this for now and instead use my Rigol spectrum analyser.
This proved to be very revealing as the 1MHz signals were as
I'd initially predicted at circa 6dBm rather than the best reading
previously of -20dBm. Quite significant!
At the moment I still cannot
see any sign of a 100KHz output from Module 14. I'd had it in
pieces several times and found a dry solder joint in the ground
circuit. I removed the module from the receiver for the nth time
and found not one bad solder joint but all the joints where the
circuit board pins and inter-board posts are soldered... something
like 30 joints. I guess this had happened because the solder
had not been hot enough to bind to the brass posts and, over
50 years of heating and cooling cycles where the brass expanded
and contracted, the solder had parted company leaving a tiny
space or a "dry joint". I resoldered everything at
415C (below Module 14 connection pins).
|
|
|
|
|
Powering the module on the bench
using a 5 volt PSU I was able to inject a 1MHz signal and, finally
at long last, see a 100KHz output. The input level could be reduced
to just -7dBm and increased to +6dBm without any output amplitude
change (ie. the TTL output signals remained at about 5 volts).
I intend to clean up the waveforms
before reassembling Module 14 and the 1MHz oscillator. I discovered,
for example, shunting the 1MHz signal from the decade counter
in the oscillator module increased the gap between fundamental
and 2nd harmonic from 7dB to nearly 30dB with an even greater
reduction of higher harmonics. This extra capacitor was connected
between ground and the base of the BSY95A amplifier.
The 100KHz filter in Module
14 can probably be improved similarly.
|
|
|
This is the signal at
the 1MHz Oscillator. It's derived from a 10MHz crystal passed
through a decade counter and is clearly digital operation is
not ideal.
The amplitude of the leftmost
1MHz spike is about +6dBm and the 2MHz spike is just 0dBm.
|
 |
|
|
The same signal but with
the scan reduced. |
 |
|
|
This is a repeat of the
first scan with a 22nF capacitor from ground to base of the BSY95A
amplifier.
The fundamental is reduced by
9dB and the second harmonic by 37dB.
I'll chose a suitable value
of capacitor later. Thinking about this though... are those spikes
required for CAL operation?
The answer is a categorical
Yes and also for generating the local oscillator for the 3rd
mixer.
|
 |
|
|
Now I've come to a very
odd situation. Powering Module 14 on the bench from a 5 volt
PSU and feeding its input (normally from the 1MHz Oscillator
module) from a signal generator running at 1MHz and 0dBm I can't
see much from the 100KHz output.. in fact typically around -69dBm.
Going beck to the TTL output before its filtered I see a 100KHz
square wave. This is fed from one element of the 5400 inverter
on Board B via a couple of links (a pink lead then a brown lead)
to Board C where it's split into a feed via a diode to the Calibration
setting on the mode switch plus the carrier insertion signal.
The latter is filtered by a small choke followed by a 10nF capacitor
to ground. At this point I measured about +6dBm. This signal
passes through two 220 ohm resistors in series to ground. At
the midpoint of the two resistors a diode feeds the output to
the mode switch. at that point the output has virtually disappeared
having dropped from 14 volts to 2.5mV. I tried forward biasing
the diode by adding a 1Kohm resistor fed by 5 volts at its anode
but that hardly affected the 100KHz signal level. Something is
making the 100KHz almost completely vanish before emerging from
the circuit board, but after resoldering this seems to have fixed
itself (more dry joints!)
I then puzzled over the bias
supply to the diodes for SSB and CAL for ages so decided to check
the receiver because this must have worked at some time. I powered
the set, turned to USB, and looked at the socket into which PL17
plugs and measured plus 5 volts. I then turned off the mains
and checked again. This time I measured the resistance between
the PL17 socket and the 5 volt regulated supply. It measured
270 ohms. The system drawing has an error showing the connection
directly to neg 15 which is clearly wrong should instead show
a 270 ohm resistor to plus 5 volts. That voltage will be for
forward biasing the diodes fitted at Module 14 feeding the 100KHz
signal for SSB carrier re-insertion and 100KHz markers for CAL.
Everything refitted and I turned
to USB to hear precisely... nothing. An easy fix though as I
noticed a pink wire had broken off one of the slidy things that
engage with the module pins and once connected it was fine. Seems
the method of disconnecting modules relies on not too many unsolderings!
LSB needs further work on the 10.6/10.8MHz module. Then there
are loads of other checks of which the biggest will be adjusting
turret tuned circuits as those in my similar PR155G had drifted
badly.
At this stage of testing I half
decided to look at the signals at the various coax connections.
Ideally this would mean leaving the RF path intact but monitoring
the nature and amplitude of each signal using my spectrum analyser
with a high impedance probe to prevent loading from the 50 ohm
termination impedance.
|
|
| FROM |
TO |
SIGNAL |
| MODULE 1-3 |
INTERP OSC |
|
| MODULE 1-10 |
MODULE 2 |
|
| MODULE 2 4 |
MODE SWITCH |
|
| MODULE 2 9 |
MODULE 3 |
|
| MODULE 2 10 |
MODULE 1 |
|
| MODULE 2 11 |
MODULE 4 |
|
| MODULE 3 9 |
MODULE 2 |
|
| MODULE 3 |
MODE SWITCH |
|
| MODULE 4 11 |
MODULE 2 |
|
| MODULE 4 12 |
MODULE 10 |
|
| MODULE 4 13 |
MODULE 5 |
|
| MODULE 5 13 |
MODULE 4 |
|
| MODULE 5 14 |
MODE SWITCH |
|
| MODULE 5 15 |
MODULE 11 |
|
| MODULE 7 16 |
MODE SWITCH |
|
| MODULE 7 17 |
MODE SWITCH |
|
| MODULE 7 18 |
MODULE 8 |
|
| MODULE 7 19 |
REAR PANEL |
|
| MODULE 7 50 |
REAR PANEL |
|
|
| FROM |
TO |
SIGNAL |
| MODULE 8 18 |
MODULE 7 |
|
| MODULE 10 5 |
TURRET |
|
| MODULE 10 12 |
MODULE 4 |
|
| MODULE 10 23 |
5 VOLT REG |
|
| MODULE 10 31 |
MODULE 11 |
|
| MODULE 11 15 |
MODULE 5 |
|
| MODULE 11 31 |
MODULE 10 |
|
| MODULE 14 14 |
5 VOLT REG |
|
| MODULE 14 14 |
5 VOLT REG |
|
| MODULE 14 42 |
5 VOLT REG |
|
| MODULE 14 44 |
5 VOLT REG |
|
| MODULE 14 48 |
PAD |
|
| MODULE 14 101 |
MODE SWITCH |
|
| MODULE 14 102 |
MODE SWITCH |
|
| ISO AMP 1 6 |
ISO AMP 2 |
|
| ISO AMP 1 8 |
MODE SWITCH |
|
| ISO AMP 2 45 |
TURRET |
|
| BFO 20 |
MODE SWITCH |
|
| PAD 26 |
INTERP OSC |
|
|
|
|
Then I changed my mind
and decided to sort out the missing LSB feed to te 3rd mixer.
This was a problem with Module
11 which I'd already rewired to its original design (except the
divide by 5 counter which was working fine). It had been unstable,
bursting into oscillation and difficult to adjust. I'd managed
to get it working on USB but even that was intermittent.
Below shows bench testing external
to the main receiver.
|
|
 |
|
Before going any further
I'd decided I had to fix the 10.6/10.8MHz generator (above).
From the evidence (eg that 10.8MHz crystal wedged in the set)
I suspect the previous owner had also had problems. The board
had been extensively modified and metal screens fitted to stop
self-oscillation. I decided to first check all the components
against the parts lists to see if any wrong parts were fitted,
but apart from a small difference everything looked OK. That
difference was T2 emitter resistor R8 which was 22 ohm instead
of the 10 ohm resistor in the circuit diagram. Could this perhaps
be a clue to earlier work because the extra resistance would
decrease the output voltage and hence the propensity to oscillate?
That 22 ohm resistor was certainly not original so I decided
to increase it initially by adding an extra 150 ohms. Sure enough
the amplifier output dropped and it stopped oscillating. I tried
82 ohms and it oscillated but with an extra 100 ohms it was stable.
Next I tried tuning the amplifier to maximise its output. I found
C15 had a large effect on the tuning of the output transformer
L7 as you'd expect but how should I adjust L1/L2 and L7?
In the receiver a comb of signals
spaced 1MHz apart is applied to the amplifier input but that
type of signal isn't available on my bench set-up so I injected
53.5MHz and connected the LSB selection pin to neg15 volts. After
twiddling L1 and L7 to resonate at 53.5MHz, tuning L2 to 53MHz
gave me a 10.6MHz output from the divide by 5 counter. Moving
the neg 15 volts to the USB selection pin and injecting 54MHz
I got an output of 10.8MHz.
Below are the results. Inputs
54MHz and 53MHz (measured at the loop over L7) were set at 0dBm.
The sinewaves from the decade counter were 10.8MHz and 10.6MHz.
In this test I left the setting of RV1 unchanged but in order
to get the best stability RV1 should be twiddled to get the optimum
sideband oscillator selection.
|
|
|
|
 |
The design of the board
must have taken ages bearing in mind commercial component tolerances,
critical layout and rather tricky adjustments not to mention
receiver operation over a wide ambient temperature range. The
area dealing with varicap tuning in my example is covered with
a hard resin compound presumably to make it as stable as possible.
Below is the method of tuning Module 11 to meet its spec.
TP1 is at the wiper of RV2 used
for setting the decade counter. It will be interesting to see
the results when the board is refitted and the following checked.
L7 is inaccessible.
|
|
|
|
|
With the module outside
its case and its coax leads plugged in and with flying leads
to its four pins I'm getting very close now. By using the comb
of harmonics rather than discrete 50MHz ish signals things had
changed. After several attempts at tuning I'd realised that my
additional 100 ohm resistor used for removing tendency to oscillate
could be removed (reverting to the 82 om resistor). Finally achieving
the following in the pictures above.... USB discrimination was
79.0-58.4=20.6dB and LSB was 78-55.9=22.1dB with amplitudes 58.4-55.9=2.5dB.
Then I noticed a silly mistake.
On the circuit board is a small pin which is obviously supposed
to be used as a test point. I'd connected a red flying lead to
this assuming it was TP1 but, although I'd initially had trouble
finding the signals shown above, by swapping C15 (27pF) for 15pF
and eventually removing the extra 100 ohm resistor at R6, those
signals above had become relatively quite clear. A slight puzzle
was their very low amplitude but I hadn't really thought much
about that.
I was thinking of making TP1
more permanent and easily accessible but then discovered that
what I'd actually measured (above) was "noise" on the
output coax not TP1. I should really solder a wire to the test
point from that central odd-shaped metal plate or the wiper of
RV2 (which is nearer to the board connections). The red wire
I'd used for measurements must have been added by the last owner
when he was having a lot of trouble (adding the extra amplifying
transistor etc) but as it supplies (a copy of) the 10MHz sinewave
heading for the 3rd mixer it's certainly not ideal for checking
the 50MHz'ish signals.. I'd soldered the red wire to the TP1
post but hadn't noticed that the original connection to that
test post was missing.
Once TP1 is available at the
test post I should be able to make final adjustments when the
module is back in its case.
And, I should mention there
was an added bonus. I can now receive Radio 4 on 198KHz loud
and clear.
|
|
| |
| |
| |
pending
|
|
|
