Bruel & Kjaer, Beat
Frequency Oscillator Type 1014
|
|
I noticed a few odd things
about this rather fine looking piece of equipment. Bearing in
mind it was buried in a garage under other stuff.. could it be
faulty? The dial felt rather stiff and it was pretty dusty so
it hadn't been used for several years.. and the last test label
was dated 1995. I plugged it into my mains supply and turned
it on having taken the trouble to find a loudspeaker and connect
it to the output terminals. I twiddled various knobs and was
rewarded with a squeal which seemed roughly around the dial frequency,
but the setting wasn't very accurate. The tuning was very stiff
although this improved a lot with the clutch switch tuned to
OFF. The centre tuning knob didn't do anything so something's
amiss in that department. Pressing the 1000Hz button produced
a tone which seemed to add to the indicated dial frequency.
The meter indicated only a small
amount and increasing the output made it increase by about an
inch of movement (to about "10" on the outer scale)
where it remained until the case was tapped.
|
 |
 |
|
Above
is the circuit diagram... click to view full size
After removing the outer case
I could see that the remote tuning spindle could tune the oscillator
if the clutch switch was set to the RH position. Turning this
to the LH position there was a click, slight buzzing and the
tuning knob locked. Setting the switch to the centre released
the clutch, but is it binding and producing stiffness which is
preventing accurate tuning?
Maybe I can fix this? I need
to be careful because the HT is around 320 volts. Getting to
the slow motion tuning drive is going to be tricky because there's
a fine tuning knob connecting to a spindle under the main tuner
via a drum and steel cable..
Below is a set of pictures of
the inside of the oscillator after removing countless screws
to detach the wooden case.
|
 |
 |
 |
 |
From the initial tests
the evidence indicates quite a few problems. The most serious
is probably the tuning mechanism. Because of this setting it
to a specific frequency is tricky because the coarse tuning knob
being stiff results in jerky movements making it impossible to
set a precise frequency other than by using the frequency trimming
knob marked "Frequency Increment".
Next, the 1,000c/s reference
button appears to produce an additive result ie. a 2kHz tone
when the main dial is set to 1KHz. This may be correct and I'll
need to read the user manual to find out.
Turning the modulation control
clockwise changes the frequency of the output by a significant
amount and might be due to a badly leaking capacitor shifting
the master oscillator.
Another fault is a bad meter.
There are in fact two faults in that the meter sticks and also
fails to read the output level correctly.
The remote tuning spindle has
a lot of backlash and this might represent a side effect of the
absence of slow motion tuning via the centre knob. It's possible
that the tuning problems are being caused by dried grease in
the clutch system. This may be preventing the clutch from being
disengaged fully. The drag is then sufficient to defeat the fine
tuning control mechanism.
Looking at the rear of the tuning
mechanism in the picture above those deposits could be dried
grease and it looks similar to the deposits on the electrolytic
capacitor to the left.
The mechanism is a bit odd.
Normally in any tuning arrangement it's not too easy to impose
end stops, but B & K get around this by not bothering to
have any. Instead the tuning pointer can rotate through 360 degrees.
A cam on the rear operates a switch which kills the master oscillator
over the dial setting between the lower and upper limits. Maybe
deactivating the switch would allow the upper limit to increase
above 20KHz and drop below 20Hz.
|
|
The following pictures show the components
To see full size just click on the picture
|
 |
 |
 |
 |
 |
 |
|
I detached the meter
which involved undoing its wires, removing its illuminating lamp
with its bulb securing clips, unscrewing four nuts then carefully
prising it away from the front panel. This latter task wasn't
easy because the rubber cushion between the meter and the panel
was stuck fast. Test with a diode test position on my multimeter
proved it now worked fine without sticking. Either the circuitry
was causing the stickiness or it cured itself. Checking the lead
continuity revealed a fault. The meter is driven from a full
wave bridge but the leads (using a diode checker) read a diode
in one direction and nothing in the other. Finding V3 with its
diodes is next on the agenda.
I was lucky.. someone had marked
the valve locations and I eventually spotted "V3" hidden
away right in the corner of the upper chassis.. but where is
that bridge rectifier that drives the meter? Is it a selenium
or copper rectifier or maybe a set of germanium diodes?
|
 |
|
 |
|
I'm assuming that "Q2"
is external to the meter "I" so finding C38 will hopefully
reveal Q2.. That chassis mounted electrolytic under V3 is marked
20uF so is not C38.
Found them! The diodes were
mounted on the power supply tagstrip (below) miles away from
V3. Not sure about mounting these meter diodes so close to those
HT rectifiers though... they're marked "OA79".
|
|
|
 |
The OA79 is a germanium
diode with a low forward voltage drop, typically 200mV at 1mA
but rises dramatically at 10mA to over 2 volts.
Although I'd expected one or
more to be bad, each one registered as OK on my diode tester.
Because the meter isn't reading very much I'm now looking for
C38, that elusive 2uF electrolytic. Pound to a penny its almost
open circuit...
|
|
 |
The wiring is pretty brittle
and mostly in harnesses making it difficult to trace.
I've now discovered the meter
circuit has been modified as it's different to my circuit diagram.
That pale brown capacitor above
is C20 and is wired to a 50Hz injection voltage via a push-button
but has a 10K resistor connected to it which doesn't figure in
the circuit diagram. In fact further testing revealed there isn't
a C38.
The true circuit is shown here.
My guess is C38 was introducing a slight error in the meter reading
especially at low frequencies. The impedance of C38 at 20KHz
is 4 ohms and at 20Hz it's 4Kohm. In simple terms the meter feed
is therefore 8K or 12K at the ends of the dial.
The very high grid impedance
of the 12AT7 makes the AC input relatively constant.
|
|
|
An easy enough job, as
I have an ESR tester, was to test the electrolytics. The cylindrical
metal cased chassis-mounted ones seemed OK. A check of the circuit
diagram showed that there are 11 electrolytics (I need to check
this figure) of which six are not critical, being HT decouplers.
Five are important and of these, three appear to have significant
roles associated with amplifier gain and load voltage C1, C2
and C6, with the other two being for smoothing HT voltage (C8
and C9).
The next step then is to identify
C1 and C2 (both 100uF),and C6 (32uF) and fit new capacitors and
then discover if any faults are cured.
Below, two capacitors C1 and
C2. The green one, C2 has already been replaced but probably
around 1995.
The result was to enable output
frequencies below about 100Hz to drive the external speakerat
the same level as a 1KHz tone, a fault I'd not really noticed.
I checked the meter readings
and found that maximum output (obviously with the loudspeaker
disconnected) read 0dB on the meter. The modulation swich produced
a tone of about 500Hz and none of the settings altered this either
frequencywise or in amplitude.. I think this voltage reading
is far too low.
|
|
|
|
|
Advice was to replace
the "Jensen" capacitors so I removed several of these
and replaced them with chip capacitors (tiny 100nF components
rated at 500VDC which all measured around 109nF).
For interest I measured all
the Jensen capacitors (they're the gold coloured ones). Five
marked 0.1uF rated at 450VDC measured as 160nF, 191nF, 338nF,
354nF and 367nF at 3.5V but experience tells me (at least with
wax covered capacitors) that with those high figures they might
leak pretty badly at an elevated voltage. I also swapped an 0.001uF
with a modern 1nF although the Jensen tested at 938pF so might
be OK. Later I'll test these at 400 volts and see if they do
leak.
In fact they weren't as bad
as I'd expected. The five 100nF leaked only about 0.5 to 1.5
milliamp although these figures were slowly increasing as the
capacitors were left connected. The smaller 1nF didn't leak
One puzzle is the burnt resistor
connected to V8. This is its screen grid feed and now measures
about 72 ohms (not 100 ohms). My guess is that C23 was leaky
(leading to excessive anode and screen current) and was replaced
by a grey Radiospares capacitor perhaps in 1995.
|
 |
|
|
You can see here the new
chip capacitors used in place of a couple of Jensen 0.1uf components.
I was lucky to spot these offered
for sale by a major supplier at a silly price of a few pence
each and I bought all their current stock of several hundred.
They're 100nF and rated at 500 volts so are good for most valved
equipents and if one needs to maintain originality
I stuff them into the old packages. The only problem I've
discovered is they're so effective that in the example of an
AR88 which I worked on I had to add parallel (leak) resistors
to reduce gain. The excessive gain was making the AR88 unstable.
|
 |
|
|
Two faults are still present
(at least). One is the low meter reading and the other, odd effects
when selecting "Modulation". I noticed another Jensen
capacitor perhaps faulty so I removed this and checked it. This
is C22 and must be a prime candidate for low meter reading as
it's feeding the meter amplifier valve (V3) control grid. It's
marked 0.02uF which is 20nF but it actually measured 135nF. Very
odd. Alas it hardly drew any leakage current and replacing it
with a 22nF made matters worse by halving the meter reading.
The problem then is either the resistors used in V3's circuit
or perhaps there's a fault in the EL84 (V8) circuit? I replaced
V8 with a new valve. No difference so I'll check C23 which looks
like a replacement. The EL84 is running very hot so that might
infer bad anode current. I'll also fit a 100 ohm resistor in
place of burnt R31 which I mentioned earlier.
As far as modulation is concerned
I twiddled most of the preset pots in case any wipers were not
making perfect contact. This did slightly change the effect of
the modulation switch (it actually produced what seemed like
correct modulation although a trifle weak and with intermittencies)
so one of the pots might need a squirt of switch cleaner.
I removed C23 and found it was
marked Radiospares 0.047uF 600V but actually measured 67nF so
must be suspect as this is 57% high. Again, like C22 the higher
value would raise the meter reading rather than lowering it unless
of course it's leaky...
|
|
 |
Here's a view of the rather
scruffy area carrying the EL84 output valve and the meter amplifier
wiring.
There's a diagonal bracing strip
which can be unscrewed at its front mount and slid out of the
way for component access.
|
|
|
Whilst checking the electrolytic
capacitors I noticed that there was a detachable cover over the
tuner and after slackening a pair of screws and sliding off the
cover I found this huge very oddly shaped variable capacitor. |
|
 |
|
Still on the hunt for
the modulation fault I tested an 0.5uF (C34) and a 1uF (C36).
Both were OK. I also changed some of the out of spec resistors
viz. R74, R60 and R31 but nothing changed. Eventually I just
altered the setting of trimmer pot P22 increasing the input to
V7. I was then able to get a two thirds reading on the meter.
Next, because of excessive volume, I connected the loudspeaker
to the attenuated output instead of the load terminals and set
the volume to give a meter reading. Previously, using the load
output, the meter read almost zero at a comfortable level. I
suspect a colour blind technician used this 490K resistor thinking
it was 390K as the original is supposed to be 400K. |
 |
|
|
|
Alas there isn't a copy of the
1014 service manual available anywhere but ChatGPT found me a
close match in the 1022 valve version. Interestingly the meter
circuit has been changed which I guess was a weakness (it was
already modified in my example as already discussed). Anyway
below is a schematic... click for full
size. |
|
 |
|
Oddly the service manual
for the 1022 is a bit mixed up. I'm guessing this is because
it's not the original copy, having been edited by a user of the
transistor version. This means the individual parts of the circuit,
although present for transistor ones, are missing for valves.
However there are some clues so at least I can follow these.
For example the load output voltage should be adjustable to 10V
at 1KHz and 60 ohms using P1 if necessary. All the voltage readings
should correspond with meter readings using an AC voltmeter at
the load terminals. Then confirm readings are good from 20Hz
to 20KHz. Although testing as described in the manual is carried
out using a B & K 2107, for my purposes an AC RMS voltmeter
was used. Here's the results after adjusting P1. I'll add a link
to the service manual later. |
|
|
FREQUENCY |
METER |
EXT METER |
|
20Hz |
10V |
9.93V |
|
1KHz |
10V |
10V |
|
20KHz |
9.4V |
9.76V |
|
|
Back to the faulty modulation
feature.. I found that at some settings vaguely very faint correct
modulation is achieved intermittently but there are two bad things
happening. Firstly the output frequency changes from say 1KHz
to something like 500Hz when the modulation control is set to
ON and the main tuning stops working at a specific frquency which
may typically be 500Hz. Below is the 1014 modulation circuitry
which is substantially identical to that in the 1022 so must
be generally OK designwise. The differences may be related to
improvements to general accuracy over the complete frequency
range? The circuit of the 1022 is below that of the 1014. |
|
 |
 |
|
I guess the next step
may be to swap some valves. V1 with the meter amp and swap over
V5 and V6, but first I'll check the voltages at V1, V5 and V6
in case there's a bad component.. Checking voltages revealed
some odd things. Mainly the voltages around V5 were most peculiar
and resistance measurements didn't make sense. There were loads
of Mohms between R45/R81 and ground so I looked for potentiometer
P3. I found one (unmarked) near to the modulation width control
and it seemed open circuit, but as it was connected to a 1uF
capacitor (C36), I wasn't sure if this was affecting measurements.
Disconnecting the wire to C36 revealed P3 was open circuit...
it should have been 2Mohm but then I noticed it wasn't P3 ....
it was P4. I temporarily fitted a 1Mohm pot. I found that modulation
was now slightly better but still not right. Primarily the frequency
was still dropping to 500Hz from 1KHz if the modulation switch
was turned on. So where is P3? I traced two leads, a blue and
a black and eventually found P3 miles away but it measured too
high. Instead of being 20Kohm it measured 200Kohm, so I twiddled
it down to minimum and modulation started working properly.
The frequency remained at the
dial setting and the modulation kicked in at the marked settings.
I removed P3 and swapped it for a brand new 20Kohm. It's a Chinese
one. I needed a single 100Kohm pot the other week in a hurry
and it was only possible to get it the next day with free postage
if I bought a box of 100 pots for £6.99.
|
|
On the back of P3 (reading 200Kohm)
it's marked "20K", so that and open circuit P4 "2M"
were the crux of the modulation fault... Maybe other pots of
this type are also bad? There's lots of them.
|
 |
|
 |
|
During tests I noticed
it was a bit uncomfortable if I lightly touched the bare metal
chassis. This tingling or rubbery feeling is due to mains leakage
to the chassis. You can see above that the mains connector isn't
polarised meaning that the mains live can connect to either connector
pin.. not that this is important in this case. In this equipment
the designers didn't fit any mains filtering so only the mains
transformer primary carries mains current.
I suppose that the tingling
sensation is due to capacitive effects and this might be different
if the mains plug is reversed. Why is a two-pin mains connector
used? Well, if one looks at say high quality audio equipment,
two pin mains connectors might be used particularly when a specific
layout of items is used. This is because very weak audio signals
would be subject to mains hum if part of the signal is carried
through mains safety earths as would be the case if each chassis
in a system has a separate safety earth carried in a different
mains cable.
The 1014 has a grounding connector
on the front panel and I guess this needs to be wired to mains
safety earth. In my application it would be best if I fitted
a 3-pin IEC mains connector in place of the B & K connector.
The mains connection panel above
is mounted to the chassis via bracket which could be replaced
with one designed for an IEC connector. Another possible solution
is to modify the two pins to fit an IEC connector and add a third
pin for a safety earth.
|
|
|
I chose to fit a snap-in
IEC connector as shown here. This turned out to be quite awkward
as the existing panel was too thick for the connector design.
The first step was to drill
out the tubular rivets and detach the metal mounting bracket.
I had to use a thinnish piece of metal on which to mount the
connector then fit this to the existing mounting bracket and
finally cut a aperture in the bakelite plate to fit the rear
of the IEC connector (without disconnecting the mains selection
wiring).
The parts were assembled using
M3 screws with one carrying a solder tag to which the centre
safety earth pin was soldered.
|
 |
 |
 |
 |
 |
|
|
This video recording shows it
set to 1KHz then the modulation is turned on followed by modulation
swing and resetting to 400Hz etc
You might have to turn up the
audio...
|
|
After sorting out the
last faults I checked the meter readings again and things had
improved quite dramatically. I could now readily overload the
meter and at a 10V reading on my multimeter I had to reduce the
preset gain pot (P21) quite a lot to produce a 10V reading as
at an indicated output of 10V the meter pointer was pinnned against
the endstop. The rnew readings are below with the old readings
in brackets. |
|
FREQUENCY |
METER |
EXT METER |
|
20Hz |
10V (10V) |
10.06V (9.93V) |
|
1KHz |
10V(10V) |
10V (10V) |
|
20KHz |
10V (9.4V) |
9.7V (9.76V) |
|
| |
| |
|
pending |
|
|
