Fullscreen HTML5Med Res (??MB)HTML4

“Truly revolutionary . . .” The PHILIPS and EL3302 cassette recorder What are the most revolutionary domestic electronic products of the last 50 or so years? The Philips Compact Cassette – and the recorder/player it was specifically designed to fit – is one that stands out. It changed our way of life immeasurably but few people today would understand how “revolutionary” the Philips Compact Cassette really was. By Ian Batty 26 Silicon Chip I’m going out on a limb by calling it revolutionary. It’s a big claim – but bear with me. Up until the early 1960s, there had been tape recording in one form or another, since people took it up after Valdemar Poulsen’s 1898 demonstration. Continuing development led to standardisation on quarter-inch tape running at 15 inches per second (ips) for pro/studio equipment and slower 7½ and 3¾ ips tape speed for domestic tape recorders. Using only one side of the tape in one pass allowed users to turn the tape reels over to get double the record/playback time. Four-track developments allowed stereo and semi-pro four-track operation. Held on reels similar to 8mm film reels, tapes were exposed to contamination and needed to be hand-threaded into the mechanism for use. But even the smallest portable transistorised units were still quite large, offering playing durations under one hour at modest quality. In hindsight, with the burgeoning prosperity of the 1960s, someone Australia’s electronics magazine siliconchip.com.au was bound to turn the audio world upside down with an economical, portable, highquality audio format. And most people would know that the Philips cassette recorder was the result. It was truly innovative but why call it “revolutionary”? Vive la révolution! Specifically, it was crucial to the Polish revolution. The aftermath of World War II saw many countries fall into the orbit of the Soviet Union, Poland among them. But by the late 1970s, civil dissatisfaction was gaining strength in Poland. The unrest in Wroclaw and later in the Gdańsk shipyards gave birth to “Solidarność Walcząca” – Fighting Solidarity – focusing organised resistance against the ruling Communist Party. But how could the revolutionaries communicate with the general population in a dictatorial, one-party state? By telephone? Too risky, as you could be intercepted and arrested. Newspapers? Forget that time-honoured medium, as well as its newer cousin, radio – as all media were effectively under State control. You might post (or carry) printed reports and speeches but they lack the immediacy of a rally and the power of a crowd responding to inspirational speakers, laying out their criticisms and remedies. This is where the humble Compact Cassette was the ideal tool. Portable recorders allowed organisers to capture the excitement of mass meetings, the stirring voices of Lech Wałęsa and his fellow Solidarity workers. Cassette copying could be done with just a few machines and some simple cabling. And the cassettes themselves were small and unobtrusive, easily carried in a bag or a coat pocket. As you would expect, the Government didn’t just roll over. Almost a decade of civil strife, including martial law and extra-judicial killings, would pass before the Polish people were able to vote freely for a democratic government. So the humble cassette tape and recorder had helped unite and inspire a nation hungry for responsible government. Did you know . . . the Berlin Radio Show in August, 1963. We’re going to gloss over the EL3300, The compact cassette/recorder and the following model, the EL3301 was never intended for music. It was (introduced in 1967; the first to introenvisaged as a dictation machi ne, duce accidental recording protection), hence the stop/start switch on the to concentrate on the model that most microphone! experts regard as “setting the stage” for the compact cassette’s massive success, the EL3302. This was first manuinaccessible test points. And we want to profactured in 1968. vide external power. And be able to listen back on earphones. Groundbreaking technology And please let’s not have a palm-sized So just how revolutionary was it, electronipatchboard with a socket for this, a socket for cally? It was pretty ground-breaking. that, another for something you just thought Before it came along, if we took the old adof. It had to be kept simple. That meant simple age of “a kilohertz per inch per second”, we controls, as well as a separate record level and would accept a reel-to-reel tape system givplayback volume, and a recording level meter ing us a 15kHz response at 15 ips (38cm/s). that would double as a battery meter. We might even accept a 7.5 ips machine Furthermore, it was stereo-mono comfor interviews, or a 3.75 ips “cheapie” for telpatible. ephone-quality speech or dictation, topping Previously, we had the crazy reel-to-reel out at around 3.7kHz. situation where you could not play stereo But a response of under 2kHz for anything? tapes on a mono machine and you could acNot good. cidentally erase the original recording of the So the first challenge was to get any kind Titanic’s sinking. of quality at the uncommonly slow speed of OK, there never was such a recording, but 1.875 ips (4.76cm/s). Akai had been able to you get the idea. do this with their X4/X5 open-reel models, Playing time? With LP records rarely but only with a sophisticated cross-field bias reaching 30 minutes per side, a “60 minute” system which could not be used in a cascassette would be a good start. The cassette sette format. tape manufacturers would take it from there. There were many other challenges. If it “Open Source” 1960s style was going to be a battery-powered machine, Philips needed to ensure constant recordHaving invented the compact cassette, ing and playback speed as the batteries disPhilips wanted rapid market uptake. Faced charged and it would need to provide constant with the problems of any single-source manspeed with changes in ambient temperature. ufacturer trying to scale up a new product, We’d also like to see every transistor used after some negotiations (particularly with in both recording and replay, with no wasteSony) they decided to offer the design free ful dedicated erase/bias oscillator, as used of royalties to any other manufacturer, so in machines of that era. long as the mechanical design was adhered It would need to drive an internal speaker to, and the relevant logos and trademarks but an external connection would allow it to were applied. show off a bit. And we also want to record The rest is history: manufacturers large from a microphone (easy enough) and from and small flocked to the table and estabhigh-level sources such as gramophones and lished an audio standard that lasted well radio tuners. into the 1980s. One more thing – thanks for the adjustment Continuous improvement in electronics on the recording bias, but let’s not force the and tape media were augmented by noisepoor techies to hook up elaborate test jigs to reduction systems such as that by Dolby Back to the 1960s But we are getting ahead of ourselves. We must turn back the clock to the early 1960s when the first Compact Cassette and the matching recorder, the EL3300, was developed by Philips in their Hasselt (Belgium) laboratory. Prototypes of both the Compact Cassette and the EL3300 were first demonstrated at siliconchip.com.au Inside the compact casette: maintaining some fidelity at the very slow speed (4.76 cm/s) on very narrow tape (3.8mm) was a real technology breakthrough, as was recording in both directions in mono or stereo, each compatible with the other. Australia’s electronics magazine July 2018  27 1962 Led by Lou Ottens at their Hasselt, (Belgium) plant, Philips develop the Compact Cassette format.   1963 Laboratories, to deliver results bettering vinyl discs. Computer software, too The cassette tape format was even adopted to store computer programs and data using the famous Kansas City format. Remember that extra DIN socket beside the keyboard port on the first IBM Personal Computer? Yep, that was a cassette port. Commodore computers even supplied branded tape drives for their VIC20/C64 series, as did other home computer manufacturers. The tape mechanism The EL3302 uses a sliding deck mechanism that carries the two tape heads and the pinch roller, engaging the cassette during recording and playback. The capstan is fixed to the main chassis. The cassettes are vertically registered by four chassis-mounted pins, with the back pressed down by a leaf spring. The two front pins, topped by conical guide cones, allow the cassette to snap lightly down at the front. The cassette is pressed lightly froward against the front pins (for complete registration) by the rear leaf spring. Conventional (ie, reel-to-reel) tape drives set the driving spindle (capstan) against the tape’s oxide side, with the pinch roller against the back. Allowing the metal capstan to contact the sensitive oxide layer gives much less tape deterioration than would happen with a rubber pinch roller contact. This works fine for a reel-to-reel system, where the system could be “oxide out” or use the universal modern plan of “oxide in”. But the compact cassette needed to present its oxide to the heads outside the cassette housing, and making the capstan bear on the (outer) oxide side would have demanded fitting each cassette with its own internal pinch roller. Philips reversed the usual plan, placing the capstan in contact with the tape back (inside the cassette housing) and the pinch roller outside, in contact with the oxide layer. While this works fine, it does allow shed oxide to accumulate on the pinch roller. Oxide accumulation on the roller (or any 28 Silicon Chip 1964 Intended for dictation, The EL3300 went on sale in the Compact Cassette and Europe and the UK in 1964 EL3300 recorder were first and in the US (under the introduced at the August Norelco brand) in November 1963 Berlin Radio Show. of the same year.   1966 Under pressure from (mainly) Sony, the Compact Cassette format was made royalty-free to other manufacturers sticky matter) can grab the tape and bunch it up around the pinch roller. So regular inspection and cleaning are advisable. The tape drive must apply a small “holdback” torque to the supply reel to prevent slack tape between the supply reel and heads. So the transport design applies a few grams of tension to keep the tape taut. Intimate contact between the tape and record/play head is critical to properly record and playback, and each cassette has a spring-loaded pressure pad for this purpose. Oxide accumulation on the pressure pad can cause a squealing sound in record or playback operation. The erase head has no pressure pad; the tape naturally wraps over its curved surface, and its powerful magnetic field is sufficient to erase the tape without needing perfect contact. The pinch roller is slightly wider than the tape, allowing its top and bottom margins to contact the capstan and pick up positive drive. After leaving the capstan/pinch roller station, it’s vital that the tape is gathered up to prevent it spooling loosely out and jamming. Failure of take-up tension is probably the most common cause of tangled/jammed cassettes. Take-up tension is applied to the take-up spindle via a felt-pad clutch driven from the flywheel. The DC motor, controlled by a polarityreversing multi-pole leaf switch, drives the flywheel via the main belt. The two cassette spindles (supply and take-up) are driven by the secondary shuttling (fast forward and rewind) mechanism. For shuttling, the heads and pinch roller remain in the retracted position, with full drive being applied to the take-up or supply spindles as determined by the position of the 1968        The 1967 Philips EL-3302, with improved performance, including better battery life and motor speed control from its 5 x AA cells (7.5V) operation handle. During playback and recording, the shuttling mechanism is disengaged from the main flywheel but lightly loads the supply spindle to ensure holdback tension between the supply reel and the capstan. The deck mechanism slides forward, inserting the two heads and the pinch roller into the cassette. At the same time, power applied to the set starts the motor’s drive to the capstan and to the take-up spindle. For playback, the play/record switch sits in its normal (play) position. For recording, the play/record switch is actuated, but only if a thin spring leaf is depressed by the record button. This is permitted if the recording tab on the rear of the cassette body has not been broken out; as purchased, the tab’s existence allows a cassette to be recorded on. Pre-recorded cassettes had the tab missing. If you subsequently wanted to record over it, the standard workaround was to put a piece of tape over the missing tab. Recording emphasis and equalisation The tape medium does not respond equally to all audio frequencies, yet we expect any record-play system to reproduce the original sound spectrum faithfully. So the designers needed to compensate for the tape medium’s peculiarities. Let’s look at the recording process first. For recording, the critical measure is the actual variations in magnetic flux “printed” onto the tape’s active layer. Whether it’s an oxide or a metallic coating, it’s easy to get a flux proportional to input signal up to the audio mid-range. The actual frequency varies with tape speed: for 38cm/s, flux is conThere were four different types of cassette tapes over the years: Type 1 – iron oxide, two write-protection notches (bottom) Type 2 – chrome/cobalt, two protection notches (middle) Type 3 – ferrichrome (not shown) Type 4 – metal, two more notches in the centre of the cassette (top). Each successive type gave improved perfomance. Source: vintagecassettes.com/history/history.htm Australia’s electronics magazine siliconchip.com.au        1973    1979   1970 Nakamichi produced cassette decks from the early ’70s and quickly became the choice of “true” audiophiles. Their top deck retailed for $US6000. . . in 1978! The Sony Walkman, launched July 1st 1979, brought mainstream appeal “on the move” to the Compact Cassette – truly revolutionary! stant to around 4.5kHz: for 4.75cm/s it’s about 1.3kHz. Flux on tape It’s desirable to correct this fall-off during recording as shown in Fig.1, as this preserves the desired level of flux on the tape at a high level, rather than letting it fall towards the system’s natural noise floor. This is an equalisation process, since it’s applied to correct system deficiencies, and is not counteracted during replay. Early mains-operated tape recorders were sensitive to mains hum, so the National Association of Broadcasters (NAB) issued a standard that boosted low frequencies around 50Hz. The specification for a time constant of 3180µs equates to around 50Hz, and this time constant specification allows easy design of a single RC feedback network for pre-emphasis. As this is pre-emphasis, its boosting of lowfrequency content will be removed by complementary de-emphasis on playback. Ultimately, there were two high-frequency equalisation curves as can be seen in Fig.2 along with the matching playback curve in Fig.4: 120µs (1.32kHz) for conventional ferric oxide tape and 70µs (2.26kHz) for chromium dioxide tape, which came along much later. A matter of bias There’s also a problem with the linearity of any magnetic circuit and here we must discuss the relationship between magnetisation (B) and magnetic flux (H). The typical B-H curve shows how recorded flux fails to match the magnetising current at low levels. See Fig.3. Notice that the path a-b only ever happens once for unmagnetised tape: every subsequent excursion of the magnetising field, H, will produce a flux, B, somewhere along b→c→d→e→g. All types of cassette players were produced by various manufacturers – this “My First Sony” aimed squarely at the children’s market. The result of this gross non-linearity is very similar to severe crossover distortion in a push-pull Class B amplifier. The earliest method to combat this was to use DC bias. This shifted the recording current up one half of the B-H curve but gave limited dynamic range and was very noisy. The solution, still in use, was to use highfrequency bias. This effectively blankets the tape with ultrasonic signal of greater amplitude than the audio signal being recorded (the EL3302 uses a bias signal of ~40kHz). The cumulative effect of the ultrasonic bias with the audio signal is a B-H curve that’s linear up to the point of magnetic saturation. Once the signal has been recorded on tape, it must be played back. In playback, the moving tape’s magnetic flux patterns cross the replay head’s pole pieces. Now, low-frequency magnetic patterns on our tape will be passing the head fairly slowly, giving slow flux changes and thus a low output voltage. But high-frequency patterns will be passing much more quickly, giving a high output voltage. You get a doubling of voltage with a doubling of frequency; more specifically, 6dB/octave or 20dB/decade. Even with a perfect recording system, the playback signal will need to be corrected so that the original audio signal’s spectral content is faithfully reproduced. Notice that this 6dB/octave rise did not exist in the recording phase, so its correction is a new application of equalisation. This involves de-emphasis as well as correcting the low-frequency pre-emphasis added during recording to reduce any 50Hz hum in the overall system. Circuit Description Now look at the circuit of the Philips Commodore Computers (remember them?) adopted the Compact Cassette format – and a dedicated recorder, the 1530 Datasette – as the storage medium for their Vic-20 computer, announced in 1980. It preceded floppy disks by some time but took (sometimes) tens of (impatient) minutes to load even quite simple programs. siliconchip.com.au 1997    2017 Cassette-only players While there are still some cassette-only players made, have morphed into all-indigital players sounded the one music systems, such as this modern AM/FM/CD/ death knell for most: the mpman was the first in 1997. Cassette unit from Philips. Australia’s electronics magazine Inside the EL3302, showing the transport and heads. To initiate recording you would hold down the record button and slide the fourfunction button towards the cassette. EL3302 (Fig.5). I’ve omitted circuit DC and signal voltages for brevity, but you can find the Dutch service manual, with full analysis, along with exploded diagrams of the mechanism, clear circuit diagram and board layouts plus electrical and mechanical adjustments at: https://elektrotanya.com (you will need to register via an electrical theory test). Switching between record and playback is handled by a multi-pole linear switch M1, with playback contacts marked as “I” and record contacts marked as “II”. The switch runs almost half the length of the main circuit board. Note that all the transistors are Philips germanium types while the diodes (all BA114) are silicon. Let’s start with the easy part, the output amplifier. It’s a conventional complementary-symmetry design, using the germanium AC127/AC128 pair to drive the speaker. Biasing is handled by D3, a BA114 silicon diode July 2018  29 Fig.5: not the first Philips cassette recorder (that honour belongs to the EL-3300), the 1968 EL3302 had a number of refinements to improve performance, and is regarded as the machine which brought the Compact Cassette format – and portable music – to the masses. The bizarre aspect is that the EL-33XX series was never intended to be used as a portable music machine: it was designed for business dictation! that gives a pretty constant 0.6V drop but responds to temperature increases by reducing its forward voltage. This means that the output transistors will get the lower bias needed at higher temperatures and will be protected from thermal runaway. The AC127/128 pair only need about 120mV each and the voltage divider comprising resistors R38 & R39 neatly provides this. The driver transistor Q5 (a lowpower, high-gain AC126) couples directly to the output pair. Its emitter goes via R37 to ground, and there is almost no DC voltage drop across the resistor. Q5 has bias applied to its base from the emitters of the output pair, via R42 and R35, forming a voltage divider with R33. But for this to happen, we need the top output transistor, Q6, to turn on. Since Q6 gets bias from the battery via R41, it will turn on strongly and pull its emitter up close to the supply voltage. This ensures that Q5 will get base bias via R42/R35, putting it into conduction. Q5’s collector current will draw the D3/R38/R39 bias network down, thus reducing Q6’s base voltage. Since this will also cause Q6’s emitter voltage to fall, the circuit experiences negative voltage feedback, stabilising the circuit with the Q6/Q7 junction at half supply, around 3.7V. There’s a capacitor, C23, in the bias circuit to ensure stability. This biasing arrangement applies both in recording and playback. During playback, the amplifier drives the speaker so you can hear the program while in record mode, it provides the ultrasonic bias and erase signals at 40kHz. So let’s look at playback mode first. Signal is applied from preamp output amplifier Q4 via R30 and C21 to Q5’s base. Switching at Fig.1: typical roll-off that would occur when recording to a compact cassette tape. 30 Silicon Chip the emitter puts C23 into circuit, bypassing emitter resistor R37. This allows Q5 to run at full gain but the overall circuit has negative feedback applied from the output emitters via R36 and C22 in series with the Q5’s bias network and Q5’s (lower) input impedance. Audio output is conveyed via switch contacts to the internal 8-ohm speaker or (if plugged in) to an external speaker. Since Q6 must draw some 10~20mA of peak base current, R41 is bootstrapped from the active terminal of the speaker. Returning the speaker’s “cold” terminal to the battery supply means that its active terminal ranges (on the output’s positive halfcycle) from around 7.5V up to some 11V at full output, thus providing adequate base current for Q6. In record mode, the output amplifier is con- Fig.2: the two high-frequency equalisation curves used during recording at 120µs (Type 1) and 70µs (Type 2-4). Australia’s electronics magazine siliconchip.com.au figured to operate as the erase/bias oscillator, running at some 40kHz. This will need (i) a resonant circuit tuned to 40kHz and (ii) positive feedback from output to input. The resonant circuit is easy. The inductance of erase head K2 is paralleled by the capacitor combination C27/C28/C29, with C28 & C29 for impedance matching. The tuned circuit then feeds back to the emitter of Q5 via R43 to the junction of Q5’s emitter and R37 (now unbypassed, since the I switch is open). The I switch connecting to the speaker is also disconnected to prevent speaker loading. For oscillation we need (i) 0° phase angle and (ii) gain >1.0 around the loop. Feedback goes to Q5’s emitter, and its base is grounded by switch II connecting the base to C23. The output stage operates as emitter-followers, so we have our 0° phase around the Fig.3: the B-H curve shows how recorded flux fails to match the magnetising current at low levels. siliconchip.com.au loop from Q5 collector to emitter. A commonbase stage has voltage gains equal to (or better than) a common-emitter stage, so the entire circuit will have a loop gain of greater than one and the circuit will oscillate at around 40kHz. The erase head, being in the oscillator circuit, receives the full drive signal and is able to erase any signal on the tape passing it. The recording head needs a smaller amount of the 40kHz signal for bias. This is picked off via C20 and R53, with preset R53 adjusted for the optimum bias level. Just before we leave this circuit, there’s R5 (22W) in series with the record/playback head, and connecting to 6-pin power socket BU2. In record mode, a small amount of bias voltage will appear across series resistor R5, allowing correct bias adjustment without the need for connections into the tightly-packed circuit board. This is done by connecting a millivoltme- Fig.4: the playback equalisation curve. Australia’s electronics magazine ter to BU2 pin 6 and setting R53 for around 25mV. Still in record mode, the preamp section uses four transistors to amplify the microphone signal of about 0.2mV and to apply pre-emphasis to the audio signal. It then drives the record section of the record/play head to “write” magnetic patterns on the cassette tape. Input amplifier Q1, a low-noise AC125, operates as a conventional combination-biased, common-emitter amplifier. It’s a “flat” stage with no shaping of its frequency response. The main section involving Q2 & Q3, also AC125s, uses a similar configuration but has either of two negative feedback paths in action, one for playback, one for recording. During playback, Q2 & Q3 get Q1’s signal directly via C3 and C5. The amplified signal appears at both the collector (across R18) and emitter (across R20) of Q3. Q3’s emitter signal is switched into the series network of C11 & R13, and sent (as negative feedback) to the base of Q2. This network causes a drop in gain with frequency. It’s a classic -6dB/octave RC feedback loop that equalises the replay head’s natural 6dB/octave output rise. Q3’s output goes, via further switching to R52, the playback volume control. From R52, the audio goes via R24 to Q4, an emitter follower which has a low output impedJuly 2018  31 On the main PCB, due to a lack of space, most components are mounted upright. It plays and records in mono only, not stereo. ance; especially necessary in record mode. Q4’s emitter signal goes via further switching, to the base of audio driver Q5 and thence via Q6/Q7 output stage to the speaker. Looking back to Q3, its collector output signal is also connected back to BU1, the DIN microphone/high-level input socket, to supply playback audio to an external amplifier. The signal also passes via R22 to the battery/utility connector BU2, to drive highimpedance headphones independent of the speaker and volume control. In record mode, Q1 gets either the microphone signal directly from socket BU1 pins 1 and 4, or an attenuated high-level signal from pins 3 and 5, via R1/R2. As in playback mode, preamp Q1 has a flat response. Record level control R51 is switched into circuit, allowing correct adjustment for recording. Like Q1, Q2 now operates with no feedback, giving maximum gain across the audio bandwidth. Q2’s signal is applied to Q3’s base via C6. It’s here that feedback is applied while recording. Q3’s output is switched directly to Q4’s base, eliminating the playback volume control. Q4’s emitter connects to an equalising network (R25/C16/R21/C14). At low frequencies, C16 & C14 have no effect, allow- ing full negative feedback from Q4’s emitter back to Q3’s base. As the frequency increases, the reactances of C16 & C14 decrease, feedback decreases and gain rises at higher frequencies. This network creates two break points to give a 12dB/octave rise in head current (and thus recorded flux) that tops out around 10kHz. This gives high-frequency equalisation to compensate for recording losses at the high of the audio band. Record amplification terminates with Q4. As well as applying feedback to Q3, Q4’s output feeds the record winding on the record/ play head via R31. This resistor’s value is high compared to the tape head’s reactance at low frequencies, so it forms a substantially constant-current drive for recording. This eliminates the need to compensate for the tape head’s inductive reactance (and thus recording current) varying with frequency. The final output branch goes, via C18/R28, to Q8, a diode-connected AC127. This rectifies the audio signal and drives the meter to show the correct recording level. Notice that, in playback, it connects to the battery supply via R34 to show the battery condition. Now for the cleverest part of this little gem, the requirement for a constant tape speed re- gardless of battery voltage and temperature. Previously, a good old governor would be used, involving a small centrifugal contact on the motor armature. As the motor reached the correct speed, the contacts would open. With the supply broken, the motor would slow slightly, allowing the contacts to close and supply power again. In practice, the speed fluctuation was quite small and could easily be damped using a rubber belt drive to a low-speed flywheel. It’s really a centrifugal version of the Tirrill (vibrating-contact) regulators used with motor car generators and early alternators. Like the Tirrill regulator, this is electrically noisy and is prone to erratic operation due to contact wear and corrosion. This method was used to obtain a constant speed for battery-driven record players of the day. An electronic speed regulator A DC motor spins due to interaction between its armature’s magnetic field and the stationary field magnet. But the armature windings are continuously passing through the field magnet’s field, so the armature winding develops a back-EMF that acts against the applied supply voltage and thus reduces the motor’s current demand. The two EMFs balance according to load, with the back-EMF decreasing under load and allowing the motor to draw extra current. An ideal motor would maintain constant speed. Armature resistance compromises a motor’s EMF-balancing process, thus practical motors slow with load. So, why not design a motor controller that can account for the armature resistance? It wouldn’t be as precise as adding a tachometer winding reporting speed feedback to a constant-speed electronic servo but it would work pretty well. Testing the EL3302’s Frequency Response Testing frequency response in flat passband) you’d get bars of constant real time (such as an amplifier) is height across the audio spectrum. So, a bit tricky. what I did was simply to record audioYou need to set the audio genfrequency pink noise for a few minutes, erator to, say, 20Hz and measure then play it back into the spectrum anathe output. Then do this for 50Hz, lyser software to determine the EL3302’s 100Hz etc, all the way up to at frequency response immediately. least 20kHz. Record/playback response For a record/playback system, you’d need to record, say, 15 secMy spectrum analyser of choice is Real onds for each spot frequency, then Time Analyser (True RTA), which an audio play the tape back and do your generator (sine/square/white noise/pink measurements, maybe rewinding EL3302 Record-Playback response using TrueRTA noise), an audio digital oscilloscope and if you missed a reading. a spectrum analyser. Spectrum Analysis software and pink noise. Spectrum analysis software Distortion analysis software is also makes this much easier. A pink noise tave (or part thereof) rather than the rising available, but I find it easy enough to use source features a high-frequency roll-off energy content of white noise. Put through a a signal generator and my noise and disthat gives constant energy levels per oc- spectrum analyser (with equipment having a tortion meter. combines 32 Silicon Chip Australia’s electronics magazine siliconchip.com.au The underside of the Philips EL3302 shows that all the copper tracks have been tinned. The main PCB is at the bottom of the photograph while the motor controller is at the upper left. Helping to put you in Control WE HAVE MOVED 44 Frankston Gardens Drive Carrum Downs VIC 3201 PH (03) 9708 2390 FX (03) 9708 2392 12V Programmable Logic Relay The motor controller uses Q9 & Q10, in series with the motor to ground, with the motor’s “top” connection going to the battery supply. Transistor Q9 operates as a comparator. Its bias divider (R45/R54/R55/S3/R49) is strung between the battery supply and the collector of regulator Q10. Q9’s emitter voltage is stabilised by series diodes D1 & D2 to about 1.2V above Q10’s collector voltage. Since Q10’s collector is the “sink” for the motor’s circuit connection, Q9’s base-emitter bias responds to the voltage drop across the motor. That is, Q9 & Q10 regulate the motor voltage according to the setting on preset pot R54. So far, we only have an adjustable electronic regulator that would keep the motor voltage constant as the battery supply ran down. We need to add load regulation to keep the motor speed constant as the motor’s mechanical load varies. Paralleled resistors R47 & R48 perform this function. If the motor current rises, the voltage across R47 & R48 will increase. This will reduce the voltage at the junction of R47 & R48 with Q10’s collector, increasing the voltage across the base bias divider to Q9. Since the emitter voltage is derived from the top of R47 & R48, the overall bias will increase and the Q9/Q10 combination will draw more current, restoring the motor’s speed to the set point. It’s a positive feedback circuit but the amount of feedback is finely balanced to counteract the motor’s natural drop in speed with increasing load. And S3? It’s a small coil of copper wire. The slogan “Nakamichi Spoken Here” was on a sticker displayed on the windows of the best audio retailers in the 1970s and 80s. It became one of the more esoteric advertising slogans, spoken in almost hushed, reverent tones! siliconchip.com.au But rather than acting as an inductor the controller uses this winding’s temperature coefficient of some +0.4% per °C to compensate the regulator against varying ambient temperature. So is this the first practical electronic motor speed regulator? Probably not, but it would have been the first to be used on such a wide scale in a consumer electronic product. All told, one can only admire the clever design aspects of this ground-breaking product. If you add up all the elegant, clever design elements, include its launching of the personal audio industry, pop in the EL3302’s part in the demise of a dictatorship, and I think I can well and truly justify that “revolutionary” title I talked about at the start of this feature. Getting it going Apart from a missing badge (top right on the speaker grille) and a worn-out carry case, my unit was in good external condition. Inside, both of the drive belts had decomposed into a sticky black goo. It’s a common fault with tape drives but I was able to get a replacement set online. The black goo is hard to remove but I found turpentine useful. Fortunately, the rubber rims on the wheels and spindles were still in good condition. Electrically, it was fine apart from noisy pots. Fiddling with the bias setting gave no better results than original specifications. How good is it? Good enough to start a revolution! The manufacturer specifies ±6dB for the frequency response, and my EL3302 achieved this over 95Hz~12.9kHz, with the more common spec of -6dB giving a result of 190Hz~8kHz. THD at 1kHz, full level was around 3% and 1.8% at 10dB down. The signal-to-noise ratio was around -52dB at 1kHz. These figures were achieved with a ...Continued on Page 103 Australia’s electronics magazine TECO SG2 Series PLR, 12VDC Powered, 6 DC Inputs, 2 Analog Inputs, 4 Relay Outputs, Keypad / Display, Expandable (Max. 34) I/O. Free Software. SKU: TEC- 004 Price: $149.95 ea + GST BACnet MS/TP IO-Module BACnet I/O Module with DIN-rail Enclosure. Ideal for building monitoring. 2 x NTC10/resistive/voltfree digital inputs, 4 x analogue 0-10V outputs, 2 x dedicated volt-free digital inputs. 2 x 24Vac triac outputs. 24Vac power supply. SKU: SXS-106 Price: $195.00 ea + GST Modbus Universal AI Module The A-1019 Remote Modbus module provides 4 isolated digital inputs, 8 Universal Analog Inputs. Analog input Type:0/4~20mA,J,K,T,E,R,S,B Thermocouple and Thermistor. RS-485 interface supports a simple ASCII protocol and Modbus RTU. SKU: YTD-419 Price: $192.00 ea + GST Closed Loop Stepper Drive Leadshine CS-D508 is a closed loop stepper drive, suitable with 2-phase stepper motor in frame size NEMA 17, 23, or 24 with 1000-line incremental encoder. SKU: SMC-160 Price: $169.00 ea + GST Closed Loop 2 Phase Motor With a 1000 line encoder. This 3.1Nm stepper motor can be used with the CS-D508 drive to detect missing steps and improved performance. SKU: MOT-163 Price: $159.95 ea + GST 5V TTL to RS-485 Converter TTL-485-5P is a bidirectional port powered RS-485 to 5V TTL converter with built in surge protection. SKU: TOD-030 Price: $59.95 ea + GST For Wholesale prices Contact Ocean Controls Ph: (03) 9782 5882 oceancontrols.com.au Prices are subjected to change without notice. July 2018  33 MARKET CENTRE Cash in your surplus gear. Advertise it here in SILICON CHIP FOR SALE PCB PRODUCTION KIT ASSEMBLY & REPAIR tronixlabs.com.au – Australia's best value for supported hobbyist electronics from Adafruit, SparkFun, Arduino, Freetronics, Raspberry Pi – along with kits, components and much more – with same-day shipping. Trouble buying old components? Need to re-spin an obsolete PCB? We do PCB layouts from files, drawings or samples. Contact Steve at sgobrien8<at>gmail. com or phone 0401 157 285. Get your old PCBs updated and keep production going! DAVE THOMPSON (the Serviceman from SILICON CHIP) is available to help you with kit assembly, project troubleshooting, general electronics and custom design work. No job too small. Based in Christchurch, NZ but service available Australia/NZ wide. Email dave<at>davethompson.co.nz PCBs MADE, ONE OR MANY. Any format, hobbyists welcome. Sesame Electronics Phone 0434 781 191. nev-sesame<at>outlook.com www.sesame.com.au VINTAGE RADIO REPAIRS: electrical mechanical fitter with 36 years ex­ perience and extensive knowledge of valve and transistor radios. Professional and reliable repairs. All workmanship guaranteed. $17 inspection fee plus charges for parts and labour as required. Labour fees $38 p/h. Pensioner discounts available on application. Contact Alan, VK2FALW on 0425 122 415 or email bigalradioshack<at>gmail. com LEDs, BRAND NAME and generic LEDs. Heatsinks, fans, LED drivers, power supplies, LED ribbon, kits, components, hardware, EL wire. www.ledsales.com.au PCB MANUFACTURE: single to multi­ layer. Bare board tested. One-offs to any quantity. 48 hour service. Artwork design. Excellent prices. Check out our specials: www.ldelectronics.com.au KEITH RIPPON KIT ASSEMBLY & REPAIR: * Australia & New Zealand; * Small production runs. Phone Keith 0409 662 794. keith.rippon<at>gmail.com ADVERTISING IN MARKET CENTRE Classified Ad Rates: $32.00 for up to 20 words (punctuation not charged) plus $1.20 for each additional word. Display ads in Market Centre (minimum 2cm deep, maximum 10cm deep): $82.50 per column centimetre per insertion. All prices include GST. Closing date: 5 weeks prior to month of sale. To book, email the text to silicon<at>siliconchip.com.au and include your name, address & credit card details, or phone Glyn (02) 9939 3295 or 0431 792 293. The Philips Compact Cassette . . . continued from page 33 selected “normal” 120µs (ie, a standard ferric oxide) tape. I found some “junk box” tapes to be pretty awful. The record level meter is reliable, with an acceptable 3% distortion level corresponding to the centre of the red zone. Speed constancy is specified in two ways: wow (slow variations up to 5Hz), and flutter (variations from 5Hz to 30Hz). Wow measured at 0.3%, flutter at 0.4%. I expected better and suspect variations in holdback tension as the main cause. There was also a definite “flanging” effect (for anyone who remembers “Itchycoo Park”) siliconchip.com.au that’s consistent with tape slewing across the playback head. Playback speed was constant down to a supply voltage of 4.7V. EL3302 versions There are the preceding EL3300/ 3301, distinguished mainly by a white plastic operation lever, and the following EL3303. Several variants of the EL3302 were produced around the world. The basic mechanism was widely re-badged by European (Telefunken, Siera), US (Norelco, Mercury, Wollensak) and Japanese (Panasonic) manufacturers, among others. Australia’s electronics magazine Further reading For the EL3302, see: www.petervis.com/ Cassette_Tape_Recorders/ and look for the EL3302 – as well as the user manual, Peter has an extensive description complete with great photos. For general references, see: en.wikipedia. org/wiki/Compact_Cassette For a more complete discussion, see: siliconchip.com.au/link/aaj2 On bias, (a quick summary), see: siliconchip.com.au/link/aaj3 For a detailed discussion of bias, see: www.hccc.org.uk/acbias.html SC July 2018  103