GETTING THE RIGHT Eq FOR 78S
A WORTHWHILE PROJECT FOR THE HOBBYIST 78rpm COLLECTOR TOP: FRONT VIEW BOTTOM: FLIPPED UP AND OVER 2 2
3 RECORD EQ Part 2 PLAYING 78rpm RECORDS By CLAUS BYRITH HE SECOND PART of this article deals with the construction of a simple, T yet effective correction amplifier from just three transistors and a small handful of resistors and capacitors. Today correction amplifiers for record playing are almost always built with OpAmps in IC packages. It is easy to design an extremely well-performing amplifier with a frequency response of almost any shape, but building it without a dedicated printed circuit board is another matter. Integrated circuits are minute and they have lots of legs, so mounting and connecting them take quite some experience. A unit providing accurate compensation for 78rpm records in a modern sound system, correcting the output from a magnetic cartridge for playback through an uncorrected amplifier input such as AUXILIARY or TUNER Components 16 resistors (R1 R16 in Circuit Diagram, Fig 2): all ½W metal film R1 100 kohm R5 120 kohm R9 680 Ohm R13 4.7 kohm R2 4.7 kohm R6 270 kohm R10 330 Ohm R14 1.5 kohm R3 2.7 kohm R7 56 kohm R11 1 MOhm R15 100 Ohm R4 3.9 kohm R8 2.2 kohm R12 22 kohm R16 100 kohm 8 capacitors (C1 C8 in Fig 2) C1 2.2 µf, 35V, electrolytic condenser C5 1.5 nf, 63V+, foil condenser C2 1.5 µf, 63V, foil condenser C6 2.2 nf, 63V+, foil condenser C3 220 µf, 16V, electrolytic condenser C7 220 µf, 35V, electrolytic condenser C4 10 nf, 63V+, foil condenser C8 47 µf, 35V, electrolytic condenser 3 transistors (Tr1 Tr3 in Fig 2): BC 239 or BC 550 or equivalent 4 RCA sockets for input and output Socket for supply voltage Single-pole toggle switch with three fixed positions for treble turnover 24V DC regulated power supply (plug pack/ac adapter), smallest available*, with plug to fit the supply socket (above) 2 copper-clad epoxy or glass fibre plates, approx 9 10cm, for mounting and screening respectively, plus a little extra for islands
Fig 1 A TRADITIONAL CORRECTION CIRCUIT For decades the correction part of a run-of-the-mill amplifier, and for that matter also of more expensive amplifiers, was designed as shown in Fig 1. Component values could differ, but the principle would remain the same: two transistors coupled as an amplifier, and the frequency shaping network connected as a negative feedback path between the output and the emitter of the first transistor. The diagram shown, features correction networks for RIAA LP records and for replay direct from a tape head, as well as a single resistor providing a linear response for a microphone, used also for the now almost obsolete ceramic pickup. From hi-fi aficionados you will inevitably hear strong objections to this solution: firstly, when this circuit corrects for the RIAA curve, the level relation between 20 Hz and 20 khz is 40 db, or 100:1, which implies that if any feedback is left to prevent distortion at 20 Hz, the feedback will exceed 40 db at 20 khz. Such a high a level of feedback and so great a variation can easily lead to instability when the amplifier handles complex signals. Secondly, at high frequencies the feedback network so loads the output as to cause slew-rate distortion, which impairs transient response. These objections are often justified. However, if well designed, and with carefully chosen working points, this type of correction amplifier can work well, as did the pickup input stage of the famous Quad 33 preamplifier, designed on these lines. 4 4
My solution to the problem is akin to this venerable design. It has the obvious virtue of using only components that are reasonably sized and therefore easy to handle; moreover, they are readily available everywhere. But most importantly, the problems mentioned above are largely eliminated. The circuit is shown in Fig 2. What we see around the first two transistors is pretty much the same as in Fig 1 except that the feedback network is different. Moreover, the resistor connecting the collector of the second transistor to the positive supply voltage is significantly smaller, which leads to a lower, yet still sufficient amplification and, more important, a much improved load capability, so that the network cannot at any point overload the output. The correction network is designed to raise the signal at low frequencies with the 3dB point close to 250 Hz. No correction is provided at high frequencies, so the load on the second transistor does not increase with increasing frequency. The maximal rise is around 16 db or 6:1, a perfectly safe ratio for any type of complex signal, and since the load does not increase at high frequencies, slew distortion is absent. The output of the second transistor is directed through two resistors to an emitter -follower with no amplification, but having a very high input- and a very low output-resistance. The current in this stage is sufficient to enable it to drive the capacitance of even a long connecting cable to the main amplifier. From the joining point of the two resistors leading to the emitter-follower, a capacitor can be switched to ground, which provides a passive treble attenuation. The switch can be open (no attenuation), or either of the two capacitors may be switched in. This yields two different 3dB points, one around 5 khz suitable for Decca ffrr records, and another around 3 khz consistent with many post-war American records. It is of course possible to use a switch with more positions and get additional 3dB points to choose from, but in my experience these two, in conjunction with the treble control of the main amplifier, give sufficient flexibility. The bass control of the main amplifier can likewise adapt the bass correction to suit practically all types of programs. The amplifier must have a DC power supply of at least 18 volts, but 24 volts is preferable, improving the overload capability. The supply voltage may be conveniently obtained from a small regulated 24V switch-mode plug pack (AC adapter). As it draws only about 13 ma of current, even the smallest type will do. But even a small current requires good quality regulation to prevent mains interference (hum). Such devices are available at a modest price at any electronics store; they normally come with an output cable fitted with a small concentric plug which of course needs a corresponding socket on the front panel of the finished unit! The amplifier can of course also be fed from batteries: three 9V batteries connected in series will do nicely. But even good batteries scarcely last more than 20 hours and if one forgets to turn off the supply, they will be dead next morning. On the other hand, 20 hours means more than 150 records, so if you are disciplined and remember to switch off your choice!* 5
Fig 2 THE CIRCUIT DIAGRAM 6 6
Fig 3 BC 550 TRANSISTOR This simple amplifier complies with all the requirements listed in Part 1; in addition to a correct Eq, an important feature is the overload capability. The amplification is slightly above 30 db. The distortion is 0.1% at 5 volts output, corresponding to 165mV at the input. Since the normal maximum output from the pickup is 7-8 mv, this yields an overload capability of 20:1, or 26 db. So, is this now a state-of-the-art correction amplifier? The answer is no. With the most advanced operational amplifiers, distortion can be at least 10 times lower and the noise can also be significantly lower. But let us not forget the intended use of this amplifier. Even the very best 78rpm records have distortion figures far greater than the amplifier. At least 3% must be expected at high output levels, and a signal-to-noise ratio of 40 db is very uncommon, found only in a few late and very successful pressings. Therefore, not only is this amplifier good enough, it is in all respects at least 20 times better than the programs it is intended to process. The amplifier is consequently well worth building; it will provide years of pleasant listening to your 78rpms. The components for the amplifier are all standard parts. The resistors are all ½W metal film types. * When the prototype is supplied with 24 volts, the voltage on the collector of the second transistor is 14.2V. The emitter is at 3.9V and the voltage on the emitter of the first transistor the voltage is 0.6V. It is difficult to measure the voltage on the collector of the first transistor because of the high value of the resistor to the supply line. The internal resistance of the voltmeter used can affect the measurement substantially. The voltage on the third transistor s emitter is always about 0.7 volts lower than the voltage on the collector of the second transistor. The voltage across the 100 Ohm resistor from the power supply to the supply line of the amplifier is 1.3V, indicating that the current draw is 13 ma. The voltages depend to some extent on the type of transistors employed so deviations up to 20% are not uncommon. The prototype used BC 550C transistors. 7
For the electrolytics C1, C7 and C8, I have used 35-volt types and for C3 a 16- volt type. The foil condensers C2, 4, 5, 6, can be any type from 63V and up. For C2 an electrolytic could have been used but as their capacitance is not very precise, a foil condenser is chosen. They take up more room, but in 63 volts, compact types are available. Almost any small signal NPN transistors can be used, provided they are of reasonably high amplification and low noise. I used BC 239 for the first prototype because I happened to have a bunch of them; BC 549 and BC 550 are good alternatives, and many other types may be used. The types with the base lead in the middle are the easiest to mount, so if possible chose types with leads as shown in Fig 3. Faulty connection will not ruin the transistors, but they will of course fail to work. The amplifier is built on a piece of copper-clad glass fibre normally used for printed-circuit boards after etching the circuit. Here it is not etched, but small squares are cut from the plate and glued on the copper side, and these islands are used to anchor all components. The copper on the main plate is used for all ground connections. Many layouts are possible nothing is critical my only advice is: do not make the amplifier too small as the soldering and positioning are less convenient if the components are crammed together. Two sockets in parallel (mono connection) for the pickup are provided, as are two parallelled outputs for connecting a standard cable to both channels of the main amplifier. Since this article is aimed at readers unfamiliar with electronic design, a detailed discussion of the circuit seems misplaced. Thus, the function of only a few key components will be explained: more experienced" readers should find the whole concept readily understandable. The 100kOkm resistor from input to ground is in parallel with the input resistance of the first transistor and is responsible for the resulting load of the pickup, which is required to be approximately 47 kohm. The 4.7kOhm resistor to the base of the first transistor helps block strong radio frequency interference from nearby radio or TV transmitters. C4 is responsible for the position of the bass turnover. Increasing the value moves the turnover point downwards and vice versa. The 1MOhm resistor across C4 in conjunction with C2 limits the rise to a maximum of 16 db at 35 Hz. Below this frequency the response flattens to an ultimate minor fall of approx. 2 db at 10 Hz. C5 and C6 are in charge of the treble attenuation. Increasing their value lowers the turnover points. The resistor from output to ground ensures that there is no DC potential on the output, which could cause a loud bang when the device is connected to a main amplifier with the volume control turned up. Finally, the 100-Ohm resistor in the supply line in conjunction with C7 helps shield the amplifier from residual noise in the supply voltage. When batteries 8 8
are used, C7 keeps the AC resistance of the supply low even if the batteries are at the brink of death. The only components you might want to change are the capacitors that define the turnover points (i.e. C4, C5, C6). Changing them does not affect the overall function of the circuit, but solely influences the frequencies. Other components can of course also be altered if special needs arise, but nearly all of them interact, so changing one might well require changing others too you have been warned! The results Amplification: Frequency response: Maximum input level: Maximum output level at all frequencies: Noise on output measured 20 20kHz: Construction 30 db Within 2 db from the standard curve (Fig 1 in the previous article) without engaging treble attenuation 165 mv for frequencies above bass-turnover 5V with 24 volts supply voltage 1mV with BC 550C transistors The main plate is a copper-clad epoxy or glass fibre plate (in the prototype 9 10 cm). The islands are approximately 6 6 mm, except for the supply island, which is a 5mm strip, and the island where the emitter of the first transistor and the correction network meet. This latter island is 6 12 mm (see the photo from the component side). The islands can be cut by a fine-blade jigsaw. I normally glue them with pressure-sensitive contact glue which is clean, fast and seems to last. The RCA phono sockets are mounted in holes in the main mounting plate. As they are screwed onto the copper on the plate, the ground connections are already there, so no ground leads from the plugs are necessary. The same applies to most types of sockets for the supply; however, if the socket is insulated, do not forget a ground connection. Normally the centre of the coaxial plugs on the supply plug packs is the positive, and so it should be. If the copper islands are placed sensibly, the wiring can be made almost exclusively using the leads to the components. I have in my prototype only twice resorted to the use of a connecting wire, 1) from the collector of the second transistor to the start of the two correcting networks and 2) from the treble correcting network to the base of the third transistor. The switch for changing treble turnover is a single-pole toggle switch with three fixed positions. Fixed position means that the lever remains in the chosen position, as opposed to types where the lever returns to the centre position when released. 9
None of the components is particularly sensitive to heat, but capacitors may be harmed by prolonged heating with the soldering iron. The finished unit may be mounted above a second fibre plate of similar dimenstions in such a way that the copper is connected to the copper on the main plate. This ensures that the correction amplifier is screened sufficiently. The amplifier in use The pickup should now be connected to the input sockets and, if a ground lead is present, this must be connected to the copper-ground on the main plate. A screw may be mounted for that purpose. The outputs should be connected to the radio or aux input on the main amplifier, and when 24V DC supply is turned on, the Eq amplifier is ready for service. Set the treble switch to the middle position (neutral), and play one of your favourite records with the tone controls on the main amplifier set to linear. Then adjust the treble attenuator to the position that gives the most natural sound. After that, fine-tune the tone controls to produce the result you want. Sometimes a compromise between the best treble reproduction and an acceptable noise level must be found, which means a more attenuated treble than you really want. A sharp cut-off low-pass filter was often found on the amplifiers of the past, but they became obsolete when the vinyl LP took over. Such a filter might prove useful; it is easily constructed, but falls, however, outside the scope for this article. As I wrote in the first part of this article sharp filters are prone to introduce an unpleasant coloration around the cut-off frequency, so they should only be used when absolutely necessary. It often happens that some attenuation improves reproduction even if no treble emphasis was used during recording. Thus, although HMV never used a deliberate treble lift during recording; nevertheless many HMV records seem to benefit from moderate attenuation in playback. There are two possible reasons: firstly, the response of many microphones of the period had a pronounced peak around 5 khz; secondly, it is no secret that some recording engineers fitted strange objects to their microphones with rubber bands in an attempt to alter the response. A lift up to about 5 khz was considered desirable, taking account of the poor treble response of the average radio-set at this time. My experience is that an Eq amplifier such as the one presented here performs very well indeed on most 78rpm material when used in conjunction with a decent main amplifier with tone controls and a suitable speaker system. It may be extended with a sharp cut-off filter (called a scratch filter in those remote days) which can be made in a simple version from a single transistor, three resistors and two capacitors. Notwithstanding the acknowledged shortcomings associated with any sharp cut-off filter, this could be a worth adding if your speakers do not provide the filtering themselves. We might return to that later. 10 10
I wish builders of this corrector the best of luck. Should anything still need some extra explanation, my email address is clausbyrith@gmail.com. I shall try to answer all questions as quickly as I can. 11
12 12