Purchase Self on Audio - 2nd Edition. Print Book & E-Book. ISBN ,
Table of contents
Line inputs of the pre-amplifier are suitable for playback purposes. This is exactly compensated for outside the feedback loop by the low-pass filter R1 C1 , which also helps to reject high frequencies above the audio band. For convenience I have referred to the next stage of the circuit as the normalization amplifier because signals leaving this should be at the nominal operating level of 0 dBm by manipulation of the input gain controls.
Separate controls are provided for each channel to allow stereo balance. A later ganged control is used for volume setting and causes no operational inconvenience. Two line inputs are also provided; line low requiring 30 mV and line high mV to give 0 dBm from the normalization stage with the input gain control fully advanced. When these inputs are selected, the feedback networks are altered to adjust the gain and give a flat frequency response. Ultrasonic filters are incorporated to ensure stability and aid r. Capacitor C3 in the feedback arm reduces the gain to unity at d.
If a fault causes the amplifier output to saturate positively the capacitor is protected by a diode which has no effect on the distortion performance. The circuitry of the normalization amplifier is complicated because its performance is required to be extremely high. The harmonic distortion is far below 0.
This large amount of preamplifier headroom allows gross preamplifier overload before clipping. The input stage of the amplifier is a differential pair with a constant-current source for good common-mode rejection. The operating currents are optimized for Advanced preamplifier design 7 good noise performance, and the output is buffered by an emitter-follower. The main voltage amplifier, Tr9 has a constant-current collector load so that high voltage gain at low distortion can be obtained. This performance is only possible if the stage has very little loading so it is buffered by the active-load emitter-follower.
The various current sources are biased by a l. Hence, this method provides exceptionally stable d. After the normalization stage the signal is applied to a tone-control circuit based on the Baxandall network. The main limitation of the Baxandall system is that the turnover frequency of the treble control is fixed. In contrast, the bass control has a turnover frequency that decreases as the control nears the flat position. This allows a small amount of boost at the low end of the audio spectrum to correct for transducer shortcomings. The equivalent adjustment at the high end of the treble spectrum is not possible because boost occurs fairly uniformly above the turnover frequency for treble control settings close to flat.
In this circuit the treble turnover frequency has been given three switched values which have proved useful in practice. Switch 2 selects the capacitors that determine the turnover point. In addition, the coupling capacitor C4 has a significant impedance at 10 Hz so that the maximum bass boost curve not only shelves but begins to fall.
The tone-control amplifier uses the same low distortion configuration as the normalization stage, but it is used in a virtual-earth mode. Resistor R4 has been increased to 5. This modification makes it much easier to compensate for stability in the unity-gain condition that occurs when treble-cut is applied. Level detection circuitry From the tone-control section the signal is fed to the final volume control via the muting reed-relay.
Note that this arrangement allows the volume control to load the input of the external power amplifier even when the relay contacts are open, thus minimising noise. Although three separate circuits are shown, these may be omitted as required. Each channel is provided with two peak-detection systems, one lights a green l. Each channel is also provided with a VU meter driver circuit. Transistor Tr22 forms a simple amplifying stage which also acts as a buffer.
Voltage feedback is used to ensure a low-impedance drive for the meter circuitry. The first peak detector is formed by IC1 and its associated components. When the voltage at pin 2 goes negative of its quiescent level by 1 V, the timer is triggered and the l. The relatively heavy l. The clipping detector continuously monitors the difference in voltage between the tone-control amplifier output and both supply rails.
If the instantaneous voltage approaches either rail, this information is held in a peak-storage system. This allows C5 to charge and turn on Tr26 , and Tr27 and hence the l. In this way both positive and negative approaches to clipping are indicated. This comprehensive level indication does of course add Advanced preamplifier design 9 significantly to the task of building and testing the preamplifier.
If desired, any or all of the three sections may be omitted. Noise gate The final section controls the muting reed-relay. Pin 2 on IC4 is, however, briefly held low by C6 , and the is therefore immediately triggered to send pin 3 high. This saturates Tr28 which prevents Tr29 from turning on. At the end of the time delay, pin 3 goes low and relay driver Tr29 is no longer disabled Figure 5.
Kundrecensioner
The noise gate uses two amplifiers with gains of about These sample both channels at the output of the normalization stage and the inputs are clamped with diodes so that the normalization amplifiers may use their full voltage swing capability without damaging the s. Due to their high gain, under normal signal conditions the op-amp outputs move continuously between positive and negative saturation which keeps the storage capacitor C7 fully charged. In the silent passages between 1. The noise gate is provided with an override switch for use with line input signals.
The delay switch-on overrides all of the circuitry. Amplifier IC2 is repeated for a stereo system. To facilitate this the response of the amplifiers is deliberately extended below the audio band. When the stylus leaves the record surface and the 1. At this point the switches and its output goes low to cut off the base drive to Tr29 , and switch off the relay. When the stylus is replaced on a record, the process takes place in reverse, the main difference being that C7 charges at once due to the low forward impedance of D2.
To prevent the relay sporadically operating when the preamplifier is handling signals presented through the line inputs, an extra wafer on the source-select switch is arranged to override the rumble-sensing circuit, and provide permanent unmute. In addition, S3 provides a manual override for testing and comparison purposes.
The power supply is shown in Figure 6. Each regulator IC requires about 7 cm2 of heat sink area. Physical layout of the preamplifier is no more critical than that of any other piece of audio equipment. In general it is wise to use a layout that places the disc input amplifier as close as possible to its input socket, and as far as possible from the mains transformer. Advanced preamplifier design 11 used between the disc input stage and its input socket, and between the final volume control and the output socket. The earthing requirements are straightforward and the circuit common 0 V rail is led from the input sockets through the signal path to the output volume control, and finally to the 0 V terminal of the power supply.
This arrangement minimises the possibility of spurious e. The only problem likely to be encountered is the formation of an earth loop when the preamplifier is connected to a power amplifier. Therefore, it may be satisfactory in a permanent installation to have the preamplifier circuitry connected to mains earth only through the signal lead to the power amplifier. The preamplifier case must of course be connected to the mains earth for safety reasons. It is preferable to define the potential of the preamplifier even if the power amplifier is disconnected. Testing is relatively straightforward, providing the preamplifier is constructed and checked stage by stage.
Dynamic parameters such as THD are not accurately measurable without expensive test gear, but it has been found in the course of experimentation that if the d. The non-signal circuitry should be relatively simple to fault-find. No problems should be encountered with the noise gate section which has proved to be very reliable throughout a protracted period of Component notes All unmarked diodes are 1N or equivalent. The muting reed relay should be a two pole make type with an 18 V coil. If a different coil voltage is used, the value of the dropper resistor should be adjusted. The VU meter should have a 1 mA movement.
If an internal diode and series resistor are fitted, the external components should be omitted. Switch 1 source select is a five pole 3 way. Switch 2 treble frequency is a four pole 3 way. The only preamplifier adjustment is for the VU meter calibration. This should be set to IV r. For normal operation the input gain controls should be set so that the meter indications do not exceed 0 VU, to preserve a safety margin in the later stages.
This completes the preamplifier design. The Audio Handbook, Newnes-Butterworth, It was my own reaction to the relative complexity of the Advanced Preamplifier just described; I set out to produce a preamplifier that was conventional, to see just how good it could be. At that time the available op-amps were looked at with entirely justified suspicion; they were relatively noisy and prone to crossover distortion in their output stages.
Crossover might be inescapable in a power amplifier, but it was definitely not a good thing to have in a preamp. Hence the use of discrete Class-A circuitry throughout. The op-amp was just becoming available at the time, but was ferociously expensive. The basic philosophy was the use of simple two or threetransistor stages, enhanced with current-source outputs when required, running from a rather high rail voltage to increase headroom and reduce distortion at a given signal level.
Linear Audio | your tech audio resource
A single supply rail was used, without regulation, but with an extra RC filter to reduce ripple to about 50 mV. This minimal-cost arrangement gave hum and noise figures as good as those from the dual-rail, IC regulator method.
Where it fell down was that there was of course no DC regulation, so when the rather low noise and distortion were being measured, the audio analyser residual signal heaved up and down like a rough sea, making measurements rather tricky, even when a high-pass filter was employed. Note the transistor equivalent of the White cathode-follower at the disc-stage output, giving push-pull Class-A operation with beautiful simplicity. Not so conventional after all. This new design sacrifices very little of that performance and uses a small number of low-cost transistors to significantly reduce the cost.
A novel active gain control makes best use of the dynamic range and removes the problem of volume control placement. This preamplifier offers a similar performance to that of the advanced preamplifier published previously,1 but with a simpler design that reduces the parts count and hence cost. In normal use, the signal levels are kept around 50 mV by exchanging the normal potentiometer volume control, which acts as an attenuation control, for an active gain control.
The distortion performance is also improved because unwanted gain will be used to give higher negative feedback and thus greater linearity. The active gain control uses a shunt feedback circuit where the volume control varies the resistance of a feedback arm as shown in Figure 1. This is followed by a third-order high-pass filter which removes subsonic signals while they are still at a low level.
Both bass boost and treble cut portions of the RIAA equalisation take place in the first stage. This is followed by a Baxandall tone control which has unity gain at 1 kHz. The use of an active gain control eliminates the problems associated with a normal volume control. If all of the gain is placed before the control, the supply voltage limits the overload margin. If some gain occurs after the volume control, then the signal-to-noise ratio is degraded because noise generated in the later stages does not undergo attenuation.
The use of two controls, one early and one late in the signal chain, is one method of avoiding this compromise1 but a true gain control is considered to be a more elegant solution. Because a low-cost, single unregulated power supply is used with firstorder RC smoothing to reduce ripple, all sections of the preamplifier are designed with high ripple-rejection performance. Disc input stage The most difficult stage to design in a preamplifier is the disc input, and the problems are compounded if, as in this case, the gain of the stage is low to allow a high overload margin.
The signal voltages shown are for maximum gain at 1 kHz. This situation becomes worse at higher frequencies when the reactance of the equalisation components falls. Therefore, as a large voltage swing at the output is desirable, a large amount of current must be able to flow into and out of the feedback network at high frequencies. A second, and related problem, is that if the gain at 1 kHz is low, the gain at 20 kHz must be Therefore, it becomes more difficult to set the top end of the RIAA curve accurately. It should be noted that if the correct turn-over frequency is chosen for the final low-pass network, the RIAA amplitude and phase curves are obtained exactly.
Because the output signal level will normally be around mV the THD level will be much lower. Noise Disc input better than 68 dB below 5 mV r. Another consequence of the fall in closed-loop gain at high frequencies is that the compensation for Nyquist stability is more difficult, and in this design it was necessary to add a conventional RC step-network to the normal dominant-pole compensation. The dominant-pole capacitor is kept as small as possible to preserve the slew-rate capability of the stage. The basic disc input stage is shown in Figure 2.
In this series-feedback configuration almost all of the voltage gain is provided by the second transistor, which has a bootstrapped collector load for high open-loop gain and linearity. The final transistor is an emitter-follower for unitygain voltage buffering. This configuration allows the use of a p-n-p input transistor for optimum noise performance, but it also means that the collector current must flow through the feedback resistance RF.
This places another constraint on the design of the feedback network because an excessive voltage drop must be avoided. As the disc input amplifier must be capable of sourcing or sinking large peak values of current to drive the capacitive feedback loop at high frequencies, the conventional emitter-following output circuit in Figure 2 is not suitable because the sink current causes a voltage drop in RE.
Replacing RE with a constant-current source is more effective because the maximum sink current becomes equal to the standing current of the stage. However, this would still limit the output of the disc stage at high audio frequencies due to an inability to sink sufficient current. For this reason, the push-pull class A configuration in Figure 3 was chosen. The bottom transistor is a current-source which is modulated in anti-phase to the top emitter-follower, via the current-sensing resistor RA and a capacitor.
This can also be considered as a negativefeedback loop that attempts to keep the current in RA constant. Due to the anti-phase drive of the lower transistor, this stage can sink a peak current of twice the standing current, and therefore give twice the output swing at high frequencies.
High-performance preamplifier 19 A practical circuit of the disc input amplifier and its associated subsonic filter is shown in Figure 4. All of the d. This chain is heavily decoupled by C2 to prevent supply-rail ripple entering this sensitive part of the circuit. Note that Tr3 and Tr5 are isolated from the bias voltage by R14 and R22 to simplify any fault-finding. Although the frequency response shows a loss of only 1. The unity-voltage gain element of the filter is formed by Tr5 and Tr6 arranged as an emitter-follower with a current-source load.
This configuration was chosen for its excellent linearity. An output of about 50 mV is available for tape recording although the exact voltage will depend on the cartridge sensitivity. Resistor R24 prevents damage to Tr6 if the tape output is shorted to earth, and resistor R25 maintains the output of the disc stage at 0 V d.
The total harmonic distortion from input to tape output at 6 V r. The distortion during use will be even lower. The disc input will accept more than 1 V r. It is felt that the improvement these figures show over conventional methods justifies the complication of a low-gain disc input stage.
The signal-to-noise ratio for a 5 mV r. The remaining part of the preamplifier comprises an active gain-control and the tone-control stage. The input switching is simple and requires only one switch section per channel. Also, any line input of suitable sensitivity can be used as a tape monitor return.
The shunt-feedback configuration of the active gain control enables each line input to have its sensitivity defined by the value of a single series input resistor. This gain is only used in the disc mode. The most sensitive line input is rated at mV for a mV output and the least sensitive input has unity gain. Any sensitivity between these two limits may be provided by using the appropriate series resistor value. If a higher voltage transformer is used, R57 should be increased accordingly.
The gain control comprises Tr 7 and Tr8 arranged as a cascode voltage amplifier with a bootstrapped collector load and Tr 9 as a conventional emitter-follower. The linearity of this circuit is increased by a current injected into Tr7 through R Resistor R40 prevents high-frequency instability when the volume control is set to zero gain. The tone-control is a conventional Baxandall circuit, with Tr10 providing a high voltage-gain by its bootstrapped collector load. Transistor Tr11 is another emitter-follower which buffers the high impedance at the collector of Tr The output is taken through R52 , which protects the output against short-circuits.
Because the output impedance is low, long cables may be used without loss of high frequencies. The power supply is shown in Figure 5. Construction Normal precautions should be taken to keep a. The leads to R54 should be kept short to prevent hum pick-up on the virtual-earth point of the gain control. Typical voltages for various parts of the circuit are shown in Figure 4. Several modifications can be made to the preamplifier to suit individual requirements. Firstly, the treble turnover frequency of the tone-control section can be increased from 2 kHz as shown in Figure 4, to 5 kHz, for example, by reducing C25 to pF.
For variable turnover frequencies C25 can be made switchable. Some purists may feel that the provision of a tone-control is unnecessary, and even undesirable. In this case, the output High-performance preamplifier 23 should be taken from the junction of C22 and R54 , but R52 and R53 should be retained at the output. In the circuit of Figure 4, no balance control is included. This function was performed in the prototype by a dual-concentric volume control. If, however, a conventional balance network is required this can be added at the output of the preamplifier although the low output impedance will be sacrificed.
It became clear that even if ultimate performance was the goal, it was no longer economical or sensible to assemble eight or more transistors into a home-made discrete op-amp. The adoption of op-amps meant a return to dual regulated power supplies, but time and progress had made this option less costly than it had been when the Advanced Preamplifier was designed. I found that getting the RIAA equalization accurate beyond a certain point by cut-and-try methods was virtually impossible, and very quickly exhausted my limited supplies of patience, so I fired up the BBC Model-B computers only took a second or so to boot in those days — we have come a long way since then and wrote some software to optimize the RIAA component selection, evaluated dozens of different volume-control laws, and minimised noise at every point; it was very well received, so it was worth it.
As with several of the designs described here, I still use the prototype on a regular basis. Until relatively recently, any audio preamplifier with pretensions to aboveaverage quality had to be built from discrete transistors rather than integrated circuits. In an article some years ago,1 I attempted to show that it was still feasible to better the performance of such devices by using simple two or three-transistor configurations.
Precision preamplifier 25 The appearance of the low-noise op-amp at a reasonable price, has changed this. It is now difficult or impossible to design a discrete stage that has the performance of the without quite unacceptable complexity. The major exception to this statement is the design of low-impedance low-noise stages such as electronically-balanced microphone inputs or moving-coil head amplifiers, where special devices are used at the input end. This version is internally compensated for gains of three or more, but requires a small external capacitor 5—15 pF for unitygain stability.
The is a very convenient package of two s in one 8-pin device with internal unity-gain compensation, as there are no spare pins. Noise performance obviously depends partly on external factors, such as source resistance and measurement bandwidth, but as an example consider the moving-magnet disc input stage shown in Figure 3. Architecture As explained in a previous article,1 the most difficult compromise in preamplifier design is the distribution of the required gain usually at least 40 dB before and after the volume control.
The more gain before the volume control, the lower the headroom available to handle unexpectedly large signals. The more gain after, the more the noise performance deteriorates at low volume settings. Another constraint is that it is desirable to get the signal level up to about mV r.
The only really practical way to get the best of both worlds is to use an active gain-control stage — an amplifier that can be smoothly varied in gain from effectively zero up to the required maximum.
A figure such as this is quite adequate, and surpasses most commercial equipment. One must next decide how large an output is needed at maximum volume for the 5 mV nominal input. This decision is made easier because using an active gain-control frees us from the fear of having excessive gain permanently amplifying its own noise after the volume control. Raising the mV to this level requires the active gain stage to have another 26 dB of gain available; see the block diagram in Figure 1.
The final step in fixing the preamp, architecture is to place the tonecontrol in the optimum position in the chain. Like most Baxandall stages, this requires a low-impedance drive if the response curves are to be predictable, and so placing it after the active gain-control block which has the usual very low output impedance looks superficially attractive. However, further examination shows that a the active-gain stage also requires a low-impedance drive, so we are not saving a buffer stage after all, and b since it uses shunt feedback the tone-control stage is rather noisier than the others,2 and should therefore be placed before the gain control so that its noise can be attenuated along with the signal at normal volume Aux.
Balance control Figure 1 Block diagram. Tone-control placed before gain-control block to reduce noise from tone-control. That due to Baxandall, chosen for this design, is at d. The tone-control is preceded by a unity-gain buffer stage with low output impedance and a very high input impedance, so that the load placed on line input devices does not vary significantly when the tapemonitor switch is operated. This brings us to the block diagram in Figure 1. Figure 3 shows the circuit diagram of the complete preamplifier. The components around A1 and A2 make up the moving-magnet disc stage and its associated subsonic filter.
Disc preamplifier stage A1 uses a quite conventional series feedback arrangement to define the gain and provide RIAA equalisation. The reality behind this rather woolly phrase is that the series configuration cannot give the continuously descending frequency response in the ultrasonic region that the RIAA specification seems to imply, because its minimum gain is unity. Hence sooner or later, as the frequency increases, the gain levels out at unity instead of dropping down towards zero at 6 dB per octave.
As described in Refs. Decoupling capacitors for i. This function is performed in Figure 3 by R8 and C11 , which also filter out unwanted ultrasonic rubbish from the cartridge. It was intended from the outset to make the RIAA network as accurate as possible, but since the measuring system used Sound Technology A has a nominal accuracy of 0.
Designing RIAA networks to this order of accuracy is not a trivial task with this configuration, due to interaction between the time-constants, and attempting it empirically proved most unrewarding. However, Lipshitz, in an exhaustive analysis of the problem, using heroic algebra in quantities not often seen, gives exact but complicated design equations. The Lipshitz equations were manipulated on an Acorn Atom microcomputer until the desired values emerged.
These proved on measurement to be within the 0. Design aims were that the gain at 1 kHz should be 26 dB, and that the value of R3 should be as small as feasible to minimize its noise contribution. These two factors mean that the RIAA network has a lower impedance than usual, and here the load-driving ability of the is helpful in allowing a full output voltage swing, and hence a good overload margin. There is a good reason why the RIAA capacitors are made up of several in parallel, when it appears that two larger ones would allow a close approach to the correct value.
It is pointless to design an accurate RIAA network if the close-tolerance capacitors cannot be easily obtained, and in general they cannot. The exception to this is the well-known Suflex range, usually sold at 2. These are cheap and easy to get, the only snag being that 10 nF seems to be the largest value widely available, and so some paralleling is required. This is however a good deal cheaper and easier than any other way of obtaining the desired close-tolerance capacitance.
Metal-oxide resistors are used in the RIAA network and in some other critical places. It is intended to provide some discrimination against subsonic rumbles originating from record warps, etc. Nonetheless the time-constant has been included, in order to keep the bottom octave of the RIAA accurate. The time-constant is not provided by R3 C3 which is no doubt what the IEC intended but by the subsonic filter itself, a rather over-damped third-order Butterworth type designed so that its slow initial roll-off simulates the Implementing the IEC roll-off by Precision preamplifier 31 reducing C3 is not good enough for an accurate design due to the large tolerances of electrolytic capacitors.
Capacitor C1 defines the input capacitance and provides some r. A compromise value was chosen, and this may be freely modified to suit particular cartridges. The suffix A denotes selection for low noise by the manufacturer. Subsonic filter As described above, this stage not only rejects the subsonic garbage that is produced in copious amounts by even the flattest disc, but also implements the IEC roll-off.
Below 16 Hz the slope increases rapidly, the attenuation typically increasing by 10 dB before 10 Hz is reached. The filter therefore gives good protection against subsonic rumbles, that tend to peak in the 4—5 Hz region. The tape output is taken from the subsonic filter, with R12 ensuring that long capacitative cables do not cause h. High-impedance buffer This buffer stage is required because the following tone-control stage demands a low-impedance drive, to ensure that operating the tape monitor switch S2 does not affect the tape-output level.
If the input selector switch S1 was set to accept an input from a medium impedance source say 5 k , and the buffer had a relatively low input impedance say 15 k , then every time the tape-monitor switch was operated there would be a step change in level due to the change of loading on the source. This is so effective that the input impedance is defined only by R Unlike discrete-transistor equivalents, this stage retains its good distortion performance even when fed from a high source resistance, e. Tone-control stage Purists may throw up their hands in horror at the inclusion of this, but it remains a very useful facility to have.
The stage is based on the conventional Baxandall network with two slight differences. Firstly the network operates at a lower impedance level than is usual, to keep the noise as low as possible. The common values of k for the bass control and 22 k for the treble control give a noise figure about 2.
Even with the values shown, the tone stage is about 6 dB noisier than the buffer that precedes it. Both potentiometers are 10 k linear, which allows all the preamplifier controls to be the same value, making getting them a little easier. The low network impedance also reduces the likelihood of capacitative interchannel crosstalk. Secondly, the tone-control stage incorporates a vernier balance facility. This is also designed as an active gain-control, with the same benefit of avoiding even small compromises on noise and headroom.
The balance control works by varying the amount of negative feedback to the Baxandall network, and therefore some careful design is needed to ensure that the source resistance of the balance section remains substantially constant as the control is altered, or the frequency response may become uneven. If you need a greater range than this, perhaps you should consider siting your speakers properly.
Active gain-control stage An active gain-control stage must fulfil several requirements. Firstly, the gain must be smoothly variable from maximum down to effectively zero. Secondly, the law relating control rotation and gain should be a reasonable approximation to logarithmic, for ease of use. Finally, the use of an active Precision preamplifier 33 stage allows various methods to be used to obtain a better stereo channel balance than the usual log. All the configurations shown in Figure 2 meet the first condition, and to a large extent, the second.
Figures 2 a and 2 b use linear controls and generate a quasi-logarithmic law by varying both the input and feedback arms of a shunt-feedback stage. The arrangement of Figure 2 c , as used in the previous article, offers simplicity but relies entirely on the accuracy of a log.
While 2 a and 2 b avoid the tolerances inherent in the fabrication of a log. This leads to imbalance at high gain settings. Peter Baxandall solved the problem very elegantly,1 by the configuration in 2 d. A buffer is required to drive Ra from the pot. It can be readily shown by simple algebra that the control track resistance now has no effect on the gain law, and hence the channel balance of such a system depends only on the mechanical alignment of the two halves of a dual linear pot.
A practical version of this is shown in Figure 3. Capacitor C25 ensures h. In short, this is a good way of wringing the maximum performance from inexpensive controls, and all credit must go to Mr Baxandall for the concept. At the time of writing there is no consensus as to whether the absolute polarity of the audio signal is subjectively important. In case it is, all the preamplifier inputs and outputs are in phase, as the inversion in the tone stage is reversed again by the active-gain stage.
Power supply The power supply is completely conventional, using complementary i. Since the total current drain both channels is less than 50 mA, they only require small heatsinks. A toriodal mains transformer is recommended for its low external field, but it should still be placed as far as possible from the disc input end of the preamplifier. Distance is cheaper and usually more effective than Mu-Metal. In my view, however, the headroom already available is ample. Construction The preamplifier may be built using either op-amps or the dual type.
The latter are more convenient requiring no external compensation and usually cheaper per op-amp, but can be difficult to obtain. To compensate each for unit gain, necessary for each one, connect 15 pF between pins 5 and 8. Note that the rail decoupling capacitors should be placed as close as possible to the op-amp packages — this is one case in which it really does matter, as otherwise this i.
It must also be borne in mind that both the and have their inputs tied together with back-to-back parallel diodes, presumably for voltage protection, and this can make fault-finding with a voltmeter very confusing. Precision preamplifier 35 Only 2. Each of these resistors sets a critical parameter, such as RIAA equalization or channel balance, and no improvement, audible or otherwise, will result from using metal oxide in other positions. Several preamplifier prototypes were built on Veroboard, the two channels in separate but parallel sections.
The ground was run through in a straight line from input to output. It must be appreciated that the crosstalk performance depends almost entirely on keeping the two channels physically separated.
Self on Audio, Second Edition
Some enthusiasts will be anxious to a use gold-plated connectors; b by-pass all electrolytics with non-polarized types; or c remove all coupling capacitors altogether, in the pursuit of an undefinable musicality. Options a and b are pointless and expensive, and c while cheap, may be dangerous to the health of your loudspeakers. Anyone wishing to dispute these points should arm themselves with objective evidence and a stamped, addressed envelope. The collected audio design articles of Douglas Self, Third Edition is the most comprehensive collection of significant articles in the technical audio press.
This third edition features 45 articles that first appeared in Elektor, Linear Audio, and Electronics World. Including expanded prefaces for each article, the author provides background information and circuit commentary. The articles cover both discrete and opamp preamplifier design, mixing console design, and power amplifier design. The preamplifier designs are illuminated by the very latest research on low noise and RIAA equalization.
The famous series of articles on power amplifier distortion is included, with an extensive commentary reflecting the latest research on compensation and ultra-low distortion techniques. This book addresses the widened scope of technology that has become available to the audio designer over the last 35 years. Software Defined Networks Timothy J. Video Compression Handbook Andy Beach.
Here Comes Everybody Clay Shirky. The Master Switch Tim Wu. Molecular Communication Andrew W. Recursive Digital Filters Stefan Hollos. Where the Animals Go James Cheshire. Software Networks Guy Pujolle. Magic and Loss Virginia Heffernan. Media Technologies Tarleton Gillespie. Art of Walt Disney: Reinventing Hollywood David Bordwell. Communicating and Mobile Systems Robin Milner.
The Filter Bubble Eli Pariser.