This is only a preview of the November 2015 issue of Silicon Chip. You can view 31 of the 96 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. Items relevant to "Open Doors With This Fingerprint Access Controller":
Items relevant to "A Universal Loudspeaker Protector":
Items relevant to "A Cheap Programmer For The PIC32 Microcontroller":
Purchase a printed copy of this issue for $10.00. |
Vintage Radio
By Ian Batty
The General Electric
P-807 5-Transistor Set
the references in the September issue.
All germanium devices are vulnerable to surface contamination. Native
germanium will naturally oxidise to
form a surface layer but germanium
dioxide is not impervious to attack.
Silicon dioxide, though, is basically
glass. This meant that manufacturers
could “top off” silicon devices with an
oxide layer and be sure of reliability.
2N170: normal & “top hat”
Many of GE’s early transistors used
a “top hat” outline, as shown in an
accompanying photograph. Much of
GE’s early output was for the military
and so reliability was critical. As a
result, early GE transistors featured a
top “exhaust port” and this allowed
the metal-cased transistors to be completely evacuated during manufacture
to ensure the best possible reliability.
Comparison/highlights/design
The P-807, General Electric’s second transistor
radio, still had only five transistors just like
the earlier model 675. However, unlike the
model 675, it used a fixed bias Class-A output
stage and other aspects of the circuit were also
quite different.
A
S DETAILED in the September
2015 issue, Thomas Edison’s
many technology companies had
consolidated into the Edison Electric
Light Company by 1889. And this in
turn eventually amalgamated with the
Thomson-Houston Electric Light Company to form General Electric in 1892.
In common with several other electronics manufacturers during World
War II, GE worked on microwave diodes for use in radar receiver mixers.
Following the war, the company eventually decided to get into transistor
manufacturing and June 1st, 1951 saw
GE list their point-contact SX-4A and
Z2 types. In today’s dollars, these tran84 Silicon Chip
sistors cost around $480 each.
Intensive development by Hall at
the Schenectady centre and Saby at
Syracuse eventually produced (respectively) the grown-junction type
(as also produced by Texas Instruments for the Regency TR-1) and the
alloyed-junction type. The latter was
released as the 2N43/44/45 in 1953
and this type dominated the industry
until the arrival of advanced diffusion
techniques.
Their Philips/Mullard cousins, the
lower-powered OC70/71, are the types
we are more familiar with in Australia. For a complete history, see the article by Mark P. D. Burgess as listed in
As with the GE 675 radio described
in the September 2015 issue, the P-807
featured here is a 5-transistor design.
However, there are important differences between the two.
First, as noted, the GE-675 uses an
ingenious “sliding bias” volume control/bias circuit for the output stage. By
contrast, the P-807 uses a conventional
diode demodulator and fixed bias on
the Class-A output stage. In addition,
unlike the GE 675, the P-807’s first IF
amplifier stage is wired in a grounded base configuration, while the loudspeaker uses a moving-armature/reed
motor to drive the speaker cone.
This type of speaker is known as a
moving-iron loudspeaker. Basically,
it consists of a solenoid and a thin
ferrous-metal diaphragm which is attached to the cone. When an audio
signal is applied to the solenoid, the
solenoid’s magnetic field varies and
the diaphragm (or reed) moves in response to this varying magnetic field.
As noted previously, a 5-transistor
design may seem like a recipe for poor
performance. However, it’s worth resiliconchip.com.au
Fig.1: GE P-807 is a 5-transistor superhet design. The G, H, S & T models were based on three modules – one for the converter
stage, another for the IF strip and the third for the audio amplifier. Note that the audio amplifier uses a Class-A output stage.
ful removal of the offending module
so that component-level repairs can
be carried out.
Circuit description
The three modules used in the P807 are soldered directly to a larger PCB. In
this photo, the converter module is at bottom left, the IF stage is at top centre,
and the audio amplifier module is at bottom right, immediately to the left of
the volume control pot.
membering that a conventional 6-transistor set has only five amplifying stages, since two of its six transistors are
used in a push-pull output amplifier
stage. By contrast, both the GE-675 and
P-805 sets use a Class-A output stage
when uses just a single transistor.
P807: first look
When I first opened the P-807’s back
and checked the circuit board, there
didn’t appear to be enough tracks for
a 5-transistor superhet design. Removing the board and flipping it over
solved the riddle. Versions P-807G, H,
S & T use three soldered-in modules:
one for the converter stage, another
for the IF strip and a third for the audio stage.
siliconchip.com.au
Basically, the set described here is
an upgrade of the original P-807B/E
models which used a conventional
PCB, with all components mounted
directly on it.
At first glance, modular design
sounds attractive. Instead of an exhaustive component-level troubleshooting procedure, one simply identifies and swaps out the faulty module.
However, this approach does rely on
the availability of replacement modules. In practice, specialised spare
parts such as these become harder to
obtain as the years go by and in this
case, a quick online search turned up
nothing. This means that any repairs
(other than to the antenna rod, tuning
gang or speaker) generally require care-
The GE-675 was the most unusual
transistor radio design I’d previously
encountered, so what surprises did the
P807 have in store? Before we delve
into the circuit, note that the following description is for the modular version, so each module has its own part
numbering. The previous “all on one
board” circuit uses a common numbering scheme.
In particular, note that the transistor, capacitor and resistor numbers
start over again on each module (eg,
there’s a TR1 on the converter module,
a TR1 on the IF module and a TR1 on
the audio amplifier module).
Fig.1 shows the circuit details of the
P807. The converter stage is quite conventional with collector-base feedback
for TR1 via oscillator coil L1. It’s similar to that used in the GE-675 but has
slightly different biasing values due to
a lower 9V supply. In addition, TR1’s
collector load is a single, tapped tuned
winding in the 1st IF transformer (T1).
By contrast, the GE-675 uses a conventional double-coil IF transformer here.
IF module
The first IF amplifier is based on TR1
(in the IF module) and, as mentioned
above, this is connected as a common
base circuit. This circuit is sometimes
preferred for its ability to work well
at both VHF and UHF frequencies. Its
November 2015 85
The three modules used in this P-807 mean that the track layout on the main
PCB is relatively simple. Note the large tuning gang at right.
main advantage is the virtual elimination of feedback capacitance, an advantage that also applies to groundedgrid valve circuits.
Its main disadvantage is lower power gain than from a well-neutralised
common-emitter stage. Although a
grounded-base configuration can give
substantial voltage gain, its current
gain is less than unity. By contrast,
a common-emitter stage can provide
current gain as well as voltage gain
and that means that its power gain
is higher.
Note that although signal injection
voltages and antenna field strengths
are quoted in microvolts per metre,
the critical issue in transistor amplifier stages is power gain. As a result,
there is a near-universal preference for
the common-emitter configuration in
all but VHF/UHF applications.
Unlike the 675, AGC is applied to
the 1st IF amplifier stage (TR2) via its
base bias circuit. In this case. AGC is
derived from the output of detector
diode D1 and fed back to TR1’s base
via resistor R6.
The output from the 1st IF amplifier
feeds the tapped, tuned primary of the
2nd IF transformer (T1). Its untapped,
untuned secondary then feeds the second IF amplifier stage based on PNP
transistor TR2.
As shown in Fig.1, TR2 is connected as a conventional common-emitter
stage and is neutralised via capacitor C4. This IF amplifier stage in turn
feeds the untapped tuned primary of
the 3rd IF transformer T2. Its untuned,
untapped secondary then feeds diode
86 Silicon Chip
demodulator D1 which then feeds the
audio amplifier. D1 also provides the
AGC voltage for the 1st IF amplifier
(TR1) and, as mentioned above, this
is fed back via resistor R6.
There’s just one final detail: TR2 is
also gain-controlled, as its base bias
is affected by TR1’s emitter voltage.
Think of TR1 as an emitter-follower
– it feeds the varying AGC voltage at
its base out via its emitter to the bottom end of TR2’s bias divider which
uses resistors R4 and R5. Note that
this happens in the modular versions
only; the B & E models lack this gaincontrol connection.
Audio amplifier
The first audio stage is based on TR1
and uses simple collector bias (R1 &
R2). As shown on Fig.1, the detected
audio from D1 is fed to TR1’s base via
volume control R102 (a 10kΩ pot) and
capacitor C1. The resulting signal on
Many of GE’s early transistors used a
“top hat” case with an exhaust port as
shown on the right. The exhaust port
allowed the metal-cased transistors
to be completely evacuated during
manufacture to ensure maximum
reliability.
TR1’s collector is then fed to the base
of output stage TR2 via capacitor C3.
Output stage TR2 works with fixed
bias, possibly because the sliding bias
circuit used in GE’s 675 model (see
SILICON CHIP, September 2015) was
unreliable and didn’t offer thermal
protection.
At first glance, TR2 appears to lack
an emitter resistor which would make
it prone to thermal runaway. However, closer examination shows that the
main 9V supply’s dropping resistor
(R6) is also TR2’s emitter resistor, and
it’s bypassed using C102. And that’s
clever – we get a combination of bias
and emitter and supply bypassing in
one circuit.
The stability is also aided by the fact
that TR2 operates as a Class-A stage
and thus has constant current drain.
By contrast, the varying current drain
of a Class-B output stage can be a recipe for audio instability without the
necessary precautions.
Moving-iron speaker motor
Output stage TR2 directly drives a
magnetic speaker with an impedance
of 550Ω. The speaker used is a movingiron type, a design that was common
and popular in the early days of radio.
This type of speaker is capable of driving a large cone to moderate volumes
and replaced the earlier horn types.
While a moving-iron speaker can
give adequate living-room volume levels, its restricted frequency response
and limited power-handling capabilities eventually saw it overtaken by
the moving-coil speaker. These can be
made with a wide frequency response
and power handling capabilities up
to hundreds of watts and remain the
dominant type today.
The author’s GE P-807
I recently purchased a GE P-807
and when I applied power, I was
rewarded with a perfectly-operating
set. A quick spin of the dial brought
in plenty of stations, so a quick clean
and polish was all I needed to do to
restore the set.
So how good is it? The answer is
pretty good. Although it lacks the
sliding-bias “battery miser” output
stage design of the 675 model, it’s a
better radio all round. What’s more, it
runs from a single 9V battery and this
makes it a more attractive vintage set
than some other early transistor radios.
Its audio response from volume consiliconchip.com.au
This close-up view shows the moving-iron speaker motor, a design that was
common in the early days of radio. This one has an impedance of 550Ω.
trol to loudspeaker is 40Hz – 3kHz
at the -3dB points and out to around
10kHz at -10dB. The response from
antenna to speaker is around 45Hz
– 1.5kHz, while the IF selectivity is
±2.5kHz at -3dB and ±30kHz at -60dB.
In practice, its audio performance
is better than the 675’s, with a maximum output of 60mW at clipping. At
5mW output, the distortion is 3.2%
(6% at 50mW).
The set’s RF sensitivity is 50µV/m at
600kHz and 90µV/m at 1400kHz but
with a S/N ratio of only 10dB. In order to achieve a 20dB S/N ratio, it requires around 70µV/m at 600kHz and
130µV/m at 1400kHz.
The 2-stage AGC is outstanding,
there being only a 6dB increase in the
output in response to a 40dB signal
increase. I also monitored TR2’s base
voltage in the IF module during testing
and found that it dropped from about
0.5V to 0.3V in the presence of strong
signals, thus confirming the operation of the 2-stage AGC circuitry. Dur-
ing these tests, I also discovered that I
had to apply some 200mV/m of signal
before the IF channel began to distort
the signal due to overload.
Loudspeaker performance
Compared to similar sets with moving-coil speakers, the P807 performs
pretty well with its moving iron speaker. Applying a pink noise signal to the
audio section resulted in an acoustic response from 200Hz to 3.5kHz
at -10dB, with a pronounced peak
around 3kHz.
This confirms the moving-magnet
mechanism’s problem with mechanical resonance. For such a small set
though, it’s a minor quibble.
How it compares
This is a well-designed little set.
Compared to the outstanding Philips
198, with its 6-transistor design and
Class-B audio output stage, the GE
P-807’s lower audio output (about
-5dB down) suffers only in very noisy
environments. In the workshop, there
really isn’t much difference between
them.
As stated, a 5-transistor set with
a Class-A output stage has the same
number of amplifying stages as a
6-transistor set with a Class-B output
stage. As a result, it follows that their
sensitivity and audio quality at moderate volume levels are pretty similar.
The P-807 does, however, have higher current consumption than similar
sets, at around 20-25mA for all volume
settings. By contrast, the GE-675 varies
its current consumption with volume
due to its sliding-bias Class-A output
stage, while Class-B sets also vary their
current with volume.
So would I buy another one? The
answer is maybe. Until I came across
this set, I had been unaware of the
two different methods used to build
the various models; ie, the P-807B/E
models used all discrete components,
while the P-807G/H/S/T series used
the 3-module approach. Apart from
servicing issues, it would be nice to
know whether the modular construction technique offered any advantages.
Finally, as with the GE-675, P-807s
are often available on eBay, usually
from the United States.
Further reading
(1) For information on the P-807B/E
non-modular sets, see: http://www.
antiqueradios.com/forums/viewtopic.
php?f=4&t=224244 where you’ll find
links to the Sams P-807 Photofact under a post by PBPP.
(2) For information on the G/H/S/T
modularised versions, see: http://
www.radiomuseum.org/r/general_el_
p807sp_807.html
(3) For information on the moving-iron
speaker, see: https://en.wikipedia.org/
SC
wiki/Moving_iron_speaker
Are Your S ILICON C HIP Issues
Getting Dog-Eared?
Are your SILICON CHIP copies getting damaged or dog-eared just
lying around in a cupboard or on a shelf?
REAL
VALUE
AT
$16.95
*
PLUS P
&
P
Keep them safe, secure & always available with
these handy binders
Order now from www.siliconchip.com.au/Shop/4 or call (02) 9939 3295
and quote your credit card number. *See website for overseas prices.
siliconchip.com.au
November 2015 87
|