This is only a preview of the January 2001 issue of Silicon Chip. You can view 34 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 "The LP Doctor: Cleaning Up Clicks & Pops; Pt.1":
Items relevant to "The WaveMaker: An Arbitrary Waveform Generator":
Items relevant to "2-Channel Guitar Preamplifier, Pt.3":
Items relevant to "Digital Reverb - The Missing Pages":
Items relevant to "PIC Programmer & TestBed":
Purchase a printed copy of this issue for $10.00. |
Clean up the clicks & pops on your vinyl records with the
Gotta lotta records? Never play ’em because of the dreaded
clicks and pops? Well, now you can start enjoying your records again.
How? Just play them through the LP Doctor and it will get rid of
most of the clicks and pops so you can enjoy them at any time.
By LEO SIMPSON & JOHN CLARKE
M
any people have big collections of vinyl records but
they can’t be bothered playing them because of the surface noise,
clicks and pops. The truth is, we’ve all
been spoilt by the pristine, noise-free
sound of CDs.
In fact, if you have not listened
to some of your old LPs for a while,
you might be shocked at just how bad
they sound...
There used to be two solutions to
the problem of clicks, pops and noise
on LP (long-play) records. You
could (try to!) ignore them – not
particularly satisfactory.
Or you could download your
!
LP to your computer via your
!
sound card and use one of the
!
many software programs which
!
allow you to edit audio files before you dub them to CDs. You
!
can then play the new disk on
!
your home CD or DVD player, in
your car or portable disk player,
and never have to play the old
vinyl record again.
To find out more about this approach, go to our feature article on
this subject starting on page 4.
For many people, the above approach is the complete answer. They
get a CD which sounds better and
they can play it any time they like.
But as always, there are drawbacks to
the computer/software/CD approach.
First, you do need a reasonably
fast Pentium or equivalent computer
22 Silicon Chip
with a late model sound card, a CD
Writer and the necessary software, all
of which can add up to a pretty big
investment.
Second, as good as these sound
editing programs are, only the more
expensive let you listen to the disk as
the processing takes place. In other
words, they don’t all do this magic
processing “on the fly”. Instead, you
have to play the record into your
computer via an RIAA preamplifier
and sound card and store the various
Features
Stereo RIAA preamplifier included
Bypass mode
Treble filter for noisy recordings
Output level adjustment
Signal clipping indicator
Indicators for click muting in each channel
tracks as .WAV files on your hard disk.
Then you set the software to work to
process the .WAV files and finally,
they can be dubbed onto a CD with
your CD Writer.
By the way, you are going to need a
pretty big hard disk (say with at least
1 or 2 Gigabytes to spare) to store and
process these .WAV files.
Third, and this is more of a subjective problem than a technical one,
even if you do create CDs to replace
your old vinyl records, you will not
have the satisfaction of looking at the
sleeve notes while you are listening
or deciding which track to play next.
That’s one of the nice features of the
old 12-inch disks. Sigh...
Finally, even if the process of CD
dubbing is entirely satisfactory, who
is going to go to the trouble of dubbing
their entire LP collection to CD, just
for the benefit of being able to listen
to any or all of them occasionally?
So if you a large collection of records, you need an alternative
to CD dubbing, one where you
can listen to any of them at any
time, substantially free of clicks
and pops.
You can do this by playing
them through the LP Doctor.
LP Doctor + CD Writer
Maybe you don’t have computer but you may want to
consider a CD Recorder such as
the Marantz DR6000 pictured
elsewhere in this article.
These make dubbing all sorts of
audio material to a CD a real doddle
but they do not include any means of
removing noise, clicks and pops from
your LP records.
This is where the LP Doctor can also
play a part. You dub your LPs to a CD
Recorder via the LP Doctor and get a
very worthwhile sound improvement,
especially on jazz and classical music
records; ie, those with quiet passages
The LP Doctor can be
teamed with any turntable having
a magnetic cartridge so you can listen to
your LPs without clicks and pops or you can feed its
signal into a PC sound card for processing and dubbing to CD.
which can be really plagued with
clicks and pops.
Or maybe you do have a computer,
CD Writer and so on but you are without a sound editing program. Here
again, you can dub your records to CD
on your computer via the LP Doctor.
LP Doctor – what it does
OK, we’ve talked about where and
why the LP Doctor might be used
but we have not been specific about
what it does, apart from the general
Fig.1; this block
diagram shows the
left channel only. The
signal from the RIAA
preamplifier is fed via
a delay so that the click
detector has sufficient
time to detect and fully
mute signal transients.
theme of click and pop removal. Let’s
describe specifically what it does.
The LP Doctor has two independent
channels which monitor the left and
right signals. When a click or pop
is sensed, the signal is muted for a
brief interval to greatly reduce its
amplitude or completely remove it.
At the same time, it can apply a slight
degree of treble cut filtering, to reduce
surface noise and also make the click
attenuation more effective.
While it does work well at greatly
reducing bad clicks and pops, it does
not do well on low level clicks which
are difficult to discern from the general program signal. In other words,
if you have trouble hearing clicks and
pops unless you listen up close to the
speakers (or wear headphones) then
the LP Doctor will also have trouble.
On the other hand, while many
pop and rock records might have a
fairly serious amount of clicks and
pops, the general signal level is usually so consistently high that it does
not matter – the music
masks the noise. So really
it is records that have a
wide dynamic range, such
as classical music and
jazz, with quiet passages
among the louder ones,
that are more likely to
be plagued with
clicks and pops.
This is where the
LP Doctor can
work very well.
But the LP Doctor does not fix
bad surface noise
which may be caused by
lots of dirt being ingrained
into the record grooves or
may be the result of fungus
attack on records which
have been stored for long
periods in fairly humid
conditions. In fact, if you
have a record which has
January 2001 23
Fig.2: this is the circuit of the left channel. IC3 is the digital delay
which is set to 1ms at switch-on by ICs17, 18, 19 & 20.
24 Silicon Chip
January 2001 25
There are also three indicators on
the front panel. Two indicate whenever a click is detected in either channel
while the third is a clipping indicator.
The output level control needs to be
adjusted so that there is normally no
clipping.
As well as providing click suppression, the LP Doctor includes a
high performance RIAA preamplifier,
making it compatible with virtually any amplifier, PC sound card
or free-standing CD Recorder. The
preamplifier is designed to suit the
majority of moving magnet cartridges
but may not have enough gain to suit
moving coil cartridges.
On the rear panel, there are two
pairs of RCA sockets. One pair is for
the magnetic cartridge signals while
the other pair is for the line level
output signals to a stereo amplifier,
sound card or CD Recorder.
Fig.3: the overall performance of the magnetic cartridge preamplifier, measured
by applying an inverse RIAA signal to the input. Our preamplifier is not ideal
but it is pretty close – the curve deviates by no more than ±0.3dB over the whole
frequency range from 20Hz to 20kHz.
been subject to a bad fungus attack,
there is nothing you can do because
the damage is permanent.
And even if you do use the LP
Doctor on a regular basis, there is
still no substitute for keeping your
records as clean as possible and also
making sure that your stylus is free of
any gunk that may be picked up from
the record grooves. In fact, you really
need to examine the stylus after every
record side has been played, to make
sure that it is clean.
By the way, if you do lots of critical listening to CDs via headphones,
you probably still won’t enjoy vinyl
records after they have been processed
by the LP Doctor. Let’s be honest: there
is still going to be a huge difference
in sound quality between an old LP
and a pristine CD.
Operating features
The LP Doctor is housed in
rack-mounting case measuring
426mm (W) x 44.5mm (H) x 277mm
(D), not including the side-mounting
flanges. Of course, you don’t have to
mount a rack-mounting case in a rack!
There are three knobs on the front
panel: an output level control, a click
sensitivity control and a selector
switch with three positions: Bypass,
Process (click suppression) and Filter.
The latter setting adds treble cut to
the click suppression, as noted above.
Operating principles
So how does the LP Doctor eliminate record clicks and pops? Essentially, each channel has a comparator
which looks for the sudden large
signal excursions which produce a
click or pop.
When the click signal is detected,
the audio signal is muted (switched
off) to prevent the click from passing
through to the output. The signal is
shut off for a short period (about seven
milliseconds) so that it is more or less
unnoticed by the listener.
The problem is, by the time the
comparator has detected the signal
and has muted it, some of the click
will have already passed through to
the output. So only part of the click or
pop will have been muted, resulting
in a shorter click but still just as loud
and annoying.
Specifications
Frequency response of RIAA phono preamplifier.... within ±0.3dB from 20Hz to 20kHz (see Fig.3)
Signal-to-Noise........................................................... -83dB unweighted (20Hz to 20kHz) with respect to 10mV input at 1kHz and
560mV; -88dB A-weighted (-84dB and -89dB in Bypass mode, respectively)
Total Harmonic Distortion (THD)................................ 0.3% at 1kHz and 1V, 3% <at> 10kHz <at>-20dBV (0.008% <at> 1kHz and .01% <at>
10kHz in bypass mode)
Separation between channels.................................... -67dB at 100Hz; -73dB at 1kHz; -58dB at 10kHz
Treble filter................................................................. -3dB at 10kHz, 12dB/octave slope
Maximum input signal................................................ 190mV RMS at 1kHz
Signal delay time....................................................... 1ms
Click muting time....................................................... 7ms
26 Silicon Chip
Inside the LP Doctor. It uses a digital delay in each channel for effective muting of clicks and pops.
Therefore, we need to incorporate
a delay circuit. This allows for the
response time of the comparator and
muting circuit such that the delayed
signal has the full muting applied to
it. In this way, the comparator/muting
circuit can attenuate the whole of the
click or pop instead of just the latter
part of it.
Block diagram
Fig.1 shows the block diagram of
the LP Doctor. For simplicity, only
the left channel is depicted; both
channels are completely separate and
identical.
First, the signal from the magnetic
cartridge is fed to the RIAA (Recording Industry Association of America)
preamplifier.
From there, the amplifier signal
is fed via four separate paths: (1)
straight through to the output via
the bypass position on switch S2; (2)
to the one-millisecond digital delay
unit (IC3) which converts the analog
signal to a digital signal, feeds it into
memory, shuffles it out again with the
required delay and converts it back to
analog; (3) to the click detector which
operates switch IC4 where click muting takes place and (4) to the clipping
detector which drives a front panel
LED to indicate that the signal level
is too high.
Circuit description
The full circuit of the LP Doctor,
apart from its power supply, is shown
in Fig.2. As with the block diagram,
we are only showing the left channel.
The equivalent IC and other device
numbers for the right channel are
shown in brackets.
The phono (magnetic cartridge)
signal is fed directly from the input
socket via inductor L1, a 150Ω resistor
and a 47µF bipolar capacitor to the
non-inverting input, pin 3, of op amp
IC1a. The inductor, series resistor and
shunt 100pF capacitor form a filter to
attenuate RF signals which may be
picked up by the phono leads.
The RIAA equalisation is provided
by the feedback network comprising
16kΩ and 200kΩ resistors and the
.0047µF and .015µF capacitors, connected between pins 1 & 2. This network provides the standard equalisation time constants of 3180µs (50Hz),
318µs (500Hz) and 75µs (2122Hz).
The circuit also includes the IEC
recommendation for rolloff below
20Hz. This is provided by the 1kΩ resistor in series with the 10µF capacitor
and by the 10µF coupling capacitor to
the 10kΩ potentiometer VR1. There
is also rolloff in following stages to
ensure that the signal below 20Hz is
attenuated. This rolloff prevents amplification of record warp and rumble
which occur at sub-audible frequencies but could possibly overload an
amplifier and loudspeakers.
The gain of IC1a is 22.45 (27dB) at
1kHz with a boost of +13.1dB at 100Hz
and a cut of -13.75dB at 10kHz. With
a typical magnetic cartridge output,
the gain provides us with a nominal 100mV of signal. This is further
boosted by op amp IC2a, by a factor
of 11, to produce a nominal signal
level of 1.1V.
The frequency response graph of
Fig.3 shows the overall performance
of the RIAA preamplifier. It was
measured by applying an inverse
RIAA signal to the preamplifier. The
response should be a dead-flat line if
the RIAA preamplifier is ideal. Our
preamplifier is not ideal but it is pretty
close – the curve deviates by no more
than ±0.3dB over the whole frequency
range from 20Hz to 20kHz.
IC6a & IC6b are comparators which
form the clipping detector and they
monitor the signals from IC2a (left
channel) and IC2b (right channel) via
10kΩ resistors and a 10µF coupling
January 2001 27
Fig.4: this is how the click detection works. The top
trace is the averaged signal at pin 5 of IC12a, while
the lower trace is the rectified signal at pin 6. When
a transient occurs, pin 6 goes above pin 5 and the
comparator output goes low to trigger the
monostable timer IC13.
Fig.5: these scope waveforms show only a short
click being detected and muted but they can last a
lot longer than this so we have settled for a
compromise muting time of 7ms.
capacitor. The non-inverting input of
IC6a is held at +1.14V while the inverting input of IC6b is held at -1.14V.
Thus, IC6a monitors positive swings
of the signal while IC6b monitors
negative swings.
So if the audio signal to the comparators exceeds 2.3V peak-to-peak
(800mV RMS), LED1 will light.
16 provides the demodulation of the
digital back to analog after the delay
within the IC. The 22kΩ and 11kΩ
resistors plus the 560pF and 150pF
capacitors at pins 13 & 14 form another low-pass filter which removes
the 500kHz digital switching artifacts
from the now-delayed analog signal.
Digital delay
The output signal from IC3 passes
through a final filter comprising a
270Ω resistor and .001µF capacitor
and is then fed through a 10µF DC
blocking capacitor to the analog
switch, IC4. This is an optically controlled FET which has a low impedance when the internal LED is lit and
a very high impedance when the LED
is off. IC4 is controlled by the click
detection circuitry involving IC11,
IC12 & IC13.
When switch IC4 isclosed, the signal passes through unaffected to pin
3 of op amp IC5a which is connected
as a unity gain buffer.
However, there is a wrinkle here
because the 10kΩ resistor and .01µF
capacitor following IC4 form a sample-and-hold circuit and the capacitor
is constantly following the audio level
as it is charged and discharged via the
10kΩ resistor.
When switch IC4 is turned off, in
response to a click, the voltage at pin
IC3 provides the digital delay and
this is set to provide a time delay of
one millisecond by data fed into its
pins 4, 5 & 6 at the moment of switchon. This initialising data is provided
by ICs 17, 18, 19 & 20. We’ll briefly
describe their operation later in this
article. For the moment, though, all
we need to know is that IC3 (IC7 in
the right channel) are set to provide
a one-millisecond delay.
The input signal from IC2a is coupled into IC3 via some low-pass filter
components comprising the 22kΩ and
11kΩ resistors at pins 23 & 22 of IC3.
This filter rolls off at 12dB/octave
above 36kHz to prevent quantisation
errors in the analog-to-digital conversion process.
The .068µF capacitor series 27Ω
resistor between pins 20 and 21 form
the integrator used in the analog to
digital (delta-sigma type) conversion.
The .068µF capacitor at pins 15 and
28 Silicon Chip
Muting switch
3 of IC5a remains roughly where it
was until the switch closes again.
In this way, the signal is not muted
down to zero but to its average level.
This results in a less-audible muting
effect and it duplicates the muting
action of the best software packages
in removing clicks.
The signal voltage from the sampleand-hold circuit is applied to pin 3
of IC5a via a 10µF bipolar capacitor.
This is included to avoid having the
input bias voltage for IC5a from being
applied to the output side of IC4. If
it was, then an audible click would
be produced each time IC4 switched
on or off.
IC5a’s output is applied direct to
switch S2a and becomes the “Processed” output. It is also fed to op
amp IC5b which provides the “Filter”
output to switch S2a. IC5b provides
a gentle treble cut at 12dB/octave
above 10kHz.
Click detection circuitry
Apart from the delay circuit just
described, the click detection circuit
is really the heart of the LP Doctor.
It takes the signal from IC2a and
further amplifies by 4.7 in IC10a. It
is then AC-coupled via a 1uF bipolar
capacitor to a precision full-wave rectifier comprising op amps IC11a and
IC11b, diodes D6 & D7 and associated
Fig.6: the loading process for the delay codes which are
fed into IC3 (and IC7) at switch-on. Serial data (lower
trace) is transferred on the negative edge of SCK (centre
trace). The REQ line (top trace) must be low before the
following 12 clock pulses. The positive edge of REQ signals
the end of the serial data stream. On the first clock pulse,
the sleep data is input and this must be a low. The following six pulses are the delay codes while next are the
low mute and ID1 and ID2 (identification codes).
resistors.
When the input signal goes positive,
the output of IC11a goes low, biasing
on D7 so that the gain is set by the
10kΩ input resistor R1 and the 10kΩ
feedback resistor R2; thus gain is -1.
This signal is coupled to the inverting input of IC11 via the 5.1kΩ
resistor. Gain for IC11b is set by the
5.1kΩ input resistor and 10kΩ feedback resistor between pins 6 and 7
and is therefore -2. Overall gain of
the input signal for positive signals
is therefore -1 x -2 = + 2.
However, there is another path for
the input signal via the 10kΩ resistor
R3 to pin 6 of IC11b. This signal gives
a negative signal at the output of IC11b
with a gain of -1. Adding the two gains
Fig.7: this is the effect of the delay through IC3 and
IC7. The top trace is the input signal and the lower
trace is the delayed signal.
gives us +1.
For negative input signals the output of IC11a is clamped high, due
to conduction of diode D6 and the
cathode of D6 is held at ground, effectively switching the output of IC11a
out of circuit. Signal then passes via
the 10kΩ resistor R3 to pin 6 of IC11b.
IC11b inverts the signal and provides
gain at -1. Since the input signal is
negative, the signal at pin 7 of IC11b
is positive.
Thus pin 7 of IC11b always goes
positive, for both positive and negative swings of the input signal and so
we have a precision full-wave rectifier.
Trimpots VR2 and VR3 provide
offset trimming for IC11a and IC11b
respectively. These are set so that pin
1 and pin 7 of IC11 are at ground (0V)
when no signal is applied.
Comparator has
floating threshold
The rectified signal from IC11b is
applied via the 27kΩ resistor and potentiometer VR4a to pin 6, the non-inverting input of comparator IC12a. It is
also applied to pin 5 via a filter network
comprising a 4.7kΩ resistor and 1µF
bipolar capacitor, before being applied
to the inverting input of IC12a.
So IC12a has a slowly varying DC
level at pin 5 and the rapidly moving
signal level at pin 6 and it is looking
for a sudden transient, ie, a click or
pop, which will cause its output at
pin 7 to go low.
The Marantz CD
Recorder is an
attractive
alternative to
dubbing your
LPs to CD via a
computer. You
don’t need a
computer or to
learn about new
software and the
result is much the
same, when you
process your LPs
via the LP Doctor.
January 2001 29
The oscilloscope waveforms of Fig.4
show how the click detection works.
The top trace is the averaged signal at
pin 5 of IC12a, while the lower trace
is the rectified signal at pin 6. When a
transient occurs, pin 6 goes above pin
5 and the comparator output goes low
to trigger the monostable timer IC13.
IC13 is a CMOS 555 connected as
a monostable. It works in two ways.
Normally, when pin 7 of IC12a is high
IC13 is untriggered (as pin 2 is high)
and the .0068µF at pins 6 & 7 of IC13
will be discharged and the output at
pin 3 will be low. When pin 7 of IC12a
goes low, it turns on transistor Q1 to
maintain discharge the .0068µF capacitor and low at pin 2 of IC13 triggers the
beginning the timing period.
The result is a seven-millisecond
positive pulse from pin 3 of IC13 and
this turns on transistor Q2 which turns
off IC4. IC13 also drives LED2 which
gives a visible indication of the muting action. This is shown in the scope
waveforms of Fig.5. This shows only a
short click being detected and muted
but they can last a lot longer than this
so we have settled for a compromise
muting time of 7ms.
If IC12a detects a longer transient, it
will hold Q1 on and keep the .0068µF
capacitor discharged for longer and this
will extend the muting period beyond
the nominal seven-millisecond period.
Comparator IC12b is there to provide
power-on muting via IC13 and IC4.
Initially, the 10µF capacitor at pin 3
of IC12b is discharged and pin 1 is
low. When the 10µF capacitor charges
via the 100kΩ resistor, its voltage goes
above pin 2 and so pin 1 of IC12b goes
high, to let IC4 unmute via IC13 and
Q2.
The scope waveforms of Fig.7 show
the effect of the delay through IC3 and
IC7. The top trace is the input signal
and the lower trace is the delayed
signal.
Delay Control
Power
IC3 (and IC7 in the right channel) are
set to the 1ms delay time via their REQ,
SCK and DATA inputs at pins 4, 5 & 6
at switch-on. IC17 to IC20 are used to
provide these clock and data codes and
after they have done this, they have no
more function in the circuit, until it is
next switched on.
They function in exactly the same
way as in the Digital Reverberation circuit described in last month’s issue so
we won’t go into the detail here, except
to say that IC20, the 74HC165 serial
shift register, is responsible (can we
say that about a lowly IC?) for loading
in the eight bits of data at switch-on. It
is clocked by IC17, the 4060 oscillator/
counter while IC18 & IC19 perform
related house-keeping tasks.
The scope waveforms of Fig.6 show
the loading process for the delay codes.
The serial data (lower trace) is transferred on the negative edge of SCK
(centre trace). The REQ line (top trace)
must be low before the following 12
clock pulses. The positive edge of REQ
signals the end of the serial data stream.
On the first clock pulse, the sleep data
is input and this must be a low. The
following six pulses are the delay codes
while next are the low mute and ID1
and ID2 (identification codes).
The power supply circuit is shown
separately on Fig.8. It uses a transformer with two 9V windings connected
in series to give an 18V centre-tapped
supply. This feeds a full-wave rectifier (diodes D1-D4) and the two 470µF
25VW capacitors. The resulting DC
voltage is around ±12V. This feeds
adjustable 3-terminal regulators set to
give ±7.5V supply rails.
The +12V supply is also applied to
a 5V regulator, REG3, via diode D5.
The diode isolates the supply to the
input capacitors of REG3 when power
is switched off. The idea is to maintain
the +5V supply to the delay ICs (IC3 &
IC7) to avoid switch-off thumps.
The mains power switch is bypassed
with a 250VAC-rated .001µF capacitor.
This prevents arcing across the switch
when it is switched off. The mains
earth connects to the circuit earth via
a 0.47µF capacitor to prevent hum in
the signal where there is no connection
to mains earth in any accompanying
amplifier. Should the power amplifier
be earthed, then the capacitor will
minimise any resulting hum loop.
Next month we will complete the
presentation of the LP Doctor with all
the constructional information and
the parts list.
SC
Fig.8: the power supply of the LP Doctor provides +5V and ±7.5V
rails. The 5V rail powers to the two delay chips.
IN
0.001F
250VAC
F1
150mA
A
SLOW
BLOW
2200F
16V
D1 - D4
4 x 1N4004
D5
1N4004
S1
250VAC
0V
470F
25V
SC
LP DOCTOR POWER SUPPLY
30 Silicon Chip
+5V
10F
16V
+7.5V
1k
4.7k
100F
16V
4.7k
100F
16V
470F
25V
IN
2001
OUT
ADJ
0.47F
CHASSIS
2200F
16V
OUT
GND
REG1
LM317
IN
9V
N
E
T1
20VA 9V
0V
240VAC
IN
REG3
LM2940-5
ADJ
OUT
REG2
LM337
10F
16V
10F
16V
1k
-7.5V
|