Fullscreen HTML5Med Res (??MB)HTML4

AUDIO OUT AUDIO OUT L R By Jake Rothman Back to the buffers – Part 2 L ast month, we started looking at the design of highquality discrete buffer circuits. We conclude this month with a complete design. *Add these parts for 48V single-rail operation R4* 33kΩ Fat electrolytics + R12 47Ω 16.2mA 0.92mA C4 1nF 4.4V R7 4.7kΩ R8 390Ω TR2 BC556B +25V + C9 ZD1 3.7V 3.9V R5 10kΩ 47µF 35V 0V + C3* R3* Referring back to Fig.10 last month, if R9 2.2µF 4.8mA 47kΩ 15Ω + 25V the coupling capacitor from the sense 44V C6 C1 0V 2.2µF resistor is too small the distortion rises 470nF 50V TR1 C8 VIN rapidly at low frequencies. So, when BC550C 100µF R11 R2 a standard bipolar transistor is used 25V 47Ω 620Ω –0.8V R1 C2 VO for a current source then the capacitor 100kΩ 100pF 10.4mA (C6, Fig.14) must be an electrolytic R14 R6 TR3 100kΩ 47kΩ type. Luckily, this capacitor has almost 0V MPSA29 LED1 the supply voltage across it, which Red high-efficiency ensures full polarisation and gives C5 + 1.13V low distortion. 10µF 1.7V 10V R10 R13 Wet electrolytics are bulky and un0.58V 560Ω 47Ω –25V reliable over time and I recommend using tantalum devices for superior C10 47µF reliability. However, high CV prod35V uct (often listed as µF×rated voltage) 0V tantalum capacitors are expensive. The value can be reduced to a cheap Fig.14. Reducing the size of C8 by using a Darlington current load. This is the final circuit, 100nF component if the input impedwith THD of 0.00055%. ance of the modulation point on the current load is raised to 470kΩ or over expensive. The common BF244A is rated low-voltage rating of standard JFETs is by using a JFET. The trouble here is that at 30V, and the 2SK30 and U1898 will not a problem if ±15V rails are used, as high-voltage (>30V) JFETs can also be tolerate a 48V power rail. However, the shown in Fig.13. The distortion is a little + C2 100pF 6.2V 1.3mA VIN C1 100pF R1 150kΩ 0V TR1 BC546B +15V* 5.56V *Power rails can be increased to ±25V (All V,I measurements shown for ±15V) C3 100nF R6 33Ω –1.34V 9.5mA 3.14V VO Maximum output ±11Vpk-pk TR3 U1898 R4 330Ω C4 100µF 16V + 4.6µA bias current –0.7V R2 4.7kΩ R3 680Ω TR2 BC556 R5 470kΩ –15V* Fig.13. CFP-modulated buffer (see Part 1, last month). Here, TR3 is a JFET, resulting in 0.003% THD. 62 Fig.15. A Veroboard prototype of Fig.14. I should have used a bigger bit of board. Yes, I know there are a lot of fat electrolytics! Practical Electronics | March | 2024 higher for a U1819 JFET current sink at 0.0035% because it is more difficult to get the full modulation required to obtain the distortion minima. A sense resistor of 680Ω had to be used to get sufficient drive. The distortion remained at the same level when used at ±25V, whereas it decreased with the bipolar circuits. An effective solution to the JFET voltage problem is to use a Darlington transistor for the constant-current load. Which leads us to the final Darlington circuit shown in Fig.14. The coupling capacitor (C6) needs to be 2.2µF. The Darlington has a 1.2V Vbe drop, double that of a normal transistor. This reduces the voltage across the constant-current set resistor (R10) by 0.6V. The residual voltage across R10 is 0.58V and this is used to set the current at 10.3mA. The Veroboard prototype is shown in Fig.15. To keep the input impedance high, TR1 should be a high Hfe device such as a BC550C. I did try a Darlington here, but there was an unexplained distortion rise at 10kHz to 0.005%, so I’m sticking with a regular NPN bipolar transistor. Components: ±25V/±12V version The component list for the ±25V version shown in Fig.14 is detailed below. Values for a ±12V version are provided in brackets. Resistors R1 100kΩ (150kΩ) 1% metal-film 0.25W R2 620Ω 1% metal-film 0.25W R3 omit, but for single-rail fit 33kΩ 1% metal-film 0.25W R4 omit, but for single-rail fit 47kΩ 5% carbon-film 0.25W R5 10kΩ 5% carbon-film 0.25W R6 47kΩ 1% metal-film 0.25W R7 4.7kΩ 1% metal-film 0.25W R8 390Ω (120Ω) 1% metal-film 0.25W R9 15Ω 5% carbon-film 0.25W R10 56Ω (47Ω) R11 47Ω 5% carbon-film 0.25W R12 47Ω 5% carbon-film 0.25W R13 4 7 Ω l i n k f o r s i n g l e r a i l 5 % carbon-film 0.25W R14 100kΩ 5% carbon-film 0.25W Capacitors C1 470nF polyester film 5mm 10% C2 100pF ceramic 5mm 5% C3 omit, but for single-rail fit 2.2µF 25V tantalum C4 470pF, (1nF) ceramic 5mm 5% C5 10µF 10V tantalum C6 2.2µF 50V, (35V) tantalum C7 220nF ceramic 10% 5mm (this is a rail-to-rail decoupling capacitor, not shown on diagrams) C8 100µF 35V (bi-polar electrolytic is preferable) C9 47µF 35V electrolytic or polymer C10 47µF 35V electrolytic or polymer Practical Electronics | March | 2024 1+1:3.6 step-up ratio audio transformer Repanco T/T3 or equivalent Input from 50Ω output of signal generator Output 50Vpk-pk at 1kHz Output to buffer under test 0V Fig.16. The nasty clip spike sometimes exhibited by over-driven modulated current loads. Semiconductors TR1 BC550CB (BC549C) NPN small signal TR2 BC556B (BC327) PNP small signal TR3 M PSA29 (MPSA13) NPN smallsignal Darlington LED1 5mm hi-eff red Rapid 55-0155 ZD1 3.9V 400mW BZY88C3V9 Testing to destruction Most designers don’t like to say what’s wrong with their circuits, or they never take the circuit out of its comfort zone. In my field I have a whole gang of musicians who ‘specialise’ in finding imaginative and unexpected ways of destroying electronics. The most recent example was a Tonebender pedal designed to work on a 9V battery, which emitted ‘confetti with fireworks’ at a wedding gig. It turned out a Hewlett Packard 33V printer power supply had been adapted and connected the wrong way to the pedal (reverse polarity supply voltage). So, I thought it prudent to stress-test the buffer, as explained below. Crazy clipping Fig.17. A 1:3.6 ratio audio transformer (type T/T3) can be used to step-up a signal generator output to overdrive buffers with ±25V rails. The transformer will distort at low frequencies and ring on square waves. description of a stage suddenly conducting or latching in an unforeseen fashion. I think the cause here is the output stage clipping, feeding an excessive distorted modulation signal into the current sink, breaking the modulation control loop. This effect occurs with input excursion magnitudes exceeding –18.5V (eg, –19V). Adding a Zener diode clamp (D1) across the sense resistor (R8) reduced the onset of this effect until –21V. So, if the input voltage stays well below this figure there will be no problem. This situation could be where a buffer precedes an amplifier with gain, such as a Baxandall active volume control. If almost full-rail voltage output is required, an op amp or diamond buffer (to be discussed next month) will have to be used. Clip testing Since these discrete buffers are designed to work with high ±25V rails and have unity gain, they have to be fed with a signal voltage as high as this to induce clipping. Very few signal generators can give this level. I suspect this is why the clipping behaviour hasn’t been mentioned much in the literature. One way of getting high levels is to use a step-up audio transformer, as I’ve come across many amplifier configuration that give lovely low distortion, but then clip horribly. This seems to be the case with the modulated current source loads used in these buffers. A brutal test I do with all audio circuits is to over-drive them and force a hard output clip at 10kHz into full load and then short the output. Sure enough, a problem came up that was invisible at the standard 1kHz test frequency. In this case it was a spike, as shown in Fig.16. These nasty clipping Fig.18. The high-voltage discrete op amp with a gain of 4x e f f e c t s w e r e c a l l e d makes an excellent signal generator booster, giving a much ‘mode changes’ by John cleaner waveform than a transformer. However, it does Linsley-Hood, an apt need a high voltage dual-rail power supply. 63 headroom from the total available power rail voltage of 50V. Also, because the circuit is asymmetrical, clipping occurs on the negative side first. The symmetrical voltage swing available into 600Ω is 42Vpk-to-pk. This is a headroom loss of 8V. The more complex discrete op amp (Audio Out, October 2023) gives 47Vpk-pk. Note that the 47Ω decoupling resistors (R12 and R13) drop an additional 0.8V full load if used. Specifications for the design in Fig.14 • Maximum current consumption running on ±25V, just before clipping at 10kHz into 600Ω is 18mA on both rails. Under no-signal conditions the current draw is 15mA. • The buffer’s power consumption is 0.9W max. • Loss of either rail switches the circuit off, and the output goes to the rail • There is an output offset of –1V, so a DC blocking capacitor is needed on the output. • The design’s frequency response is –1dB at 10Hz and 400kHz (slew limiting sets in at 350kHz at 14Vpk-pk). • Input impedance 90kΩ. • Output impedance is 34Ω (not including output resistor R11). • Distortion 0.0006% 0dBm (2.2Vpk-pk into 600Ω) at 1kHz. The distortion curve is shown in Fig.19. Fig.19. Distortion curve for the final buffer shown in Fig.14 using an MPSA29 for the current sink load, a BC550C for TR1 and ±25V rails gave 0.00055% THD. Distortion cancellation Fig.20. Distortion curve of a dual balanced buffer comprising two of the circuits shown in Fig.14 achieved an excellent 0.00045% THD. Maximum output The sense resistor R8 and the current sink consume quite a bit of voltage *Add these parts for 24V single-rail operation Unforeseen consequences 64 C3* + 2.2µF 25V 0V VIN R2 620Ω R7 4.7kΩ C1 470nF 47µF 16V 0V 2.2mA **Flip polarity for single-rail operation or use a bi-polar device C8** 100µF R11 16V 47Ω –1V TR3 MPSA13/14 0V R5 10kΩ R9 15Ω + C6 2.2µF 35V TR1 BC549C C2 100pF ZD1 3.9V R8 120Ω TR2 BC327 R3* 47kΩ R1 100kΩ +12V + C9 C4 1nF + In electronics, solving one problem usually leads to another. For the discrete buffer in Fig.14, the clamp diode (ZD1) passed a destructive current on the positive rail when the output was shorted. This was fixed by adding a 15Ω resistor (R9). Now when the output is shorted at full clip at 10kHz, there’s no smoke, just a current of 80mA flowing in the positive rail and 30mA in the negative. This is survivable shortterm with a BC556 (I c max, pulsed 200mA, continuous 100mA). Note that this current-limiting configuration assumes you have included R11 (47Ω) in series with the buffer output. For a buffer with higher current rating, use a higher-current transistor, such as a 2SA1275Y or ZTX751. R4* 33kΩ R12 47Ω 13mA R10 47Ω C5 + 10µF 10V VO R14 100kΩ R6 47kΩ LED1 Red high-efficiency 1.7V 0V + illustrated in Fig.17, although many transformers have distortion problems at these levels. The solution here is to use the discrete op amp discussed in Audio Out, October 2023, running at ±27V, as shown in Fig.18 to boost the amplitude of a typical signal generator. Just out of interest, I arranged two Fig.14 buffer circuits in balanced mode to see if the distortion could be reduced further by push-pull cancellation. This was easy to check since my AP distortion analyser has balanced inputs and outputs. Cancellation is much more R13 47Ω –12V C10 100µF 16V Fig.21. Fig.14 circuit modified to ±12V rails. Practical Electronics | March | 2024 3.14mA 6.2V C1 100nF TR1 BF244A 6.13mA Sense resistor R2 1.2kΩ R8 470Ω +15V R7 10kΩ TR2 BC556 2.8mA + one given here, but it still made a difference. When used in balanced mode the distortion went down to 0.00045%, as shown in Fig.20. ±12V version C6 22µF 35V A low-voltage version of the circuit is shown in R1 R3 Fig.21; running off ±12V. C8 290mV 1MΩ 82Ω R11 100µF It has slightly higher distor33Ω 16V 0V –0.14V tion at 0.0025% (compared VO to Fig.14, ±25V); its max58mV imum output is 20Vpk-pk. TR3 The voltage rating of the R6 BC546 2.2kΩ 9.6mA Fig.21 transistors and caLED1 1.6V pacitors can be reduced by Red R4 0.96V high-efficiency 100Ω half compared to Fig.14. A BC549C is used for TR1 –15V and a BC327 for TR2. TR3 can be the more common Fig.22. CFP buffer with JFET input. Note this circuit MPSA13/4 Darlington, diagram has been simplified by leaving off the which is rated at 30V. To decoupling and input filtering components. ensure lowest distortion, R8 is reduced to 120Ω and R10 to 47Ω. The BC327 will drive a loweffective when the basic distortion is er-impedance load than the BC556. TR2’s quite high and mainly low even-order quiescent current is 11mA. For the whole harmonics, such as second. This is the circuit it’s 13mA. If a BF244 JFET is case with JFET followers and valve pushused for TR1, as shown in Fig.22, distorpull output stages. It is less effective tion increases to 0.0035%, but the input with low distortion circuits, such as the VIN impedance is much higher and R1 can be increased to well over 2.2MΩ if needed. Single-rail operation + To run the buffer off a single rail, insert bias resistors R3/R4. These, along with R1, provide a mid-point bias. In practice, this has to be a bit higher than half rail to ensure symmetrical clipping. C3 decouples the bias to ensure no hum and noise can enter the input via the bias network. The negative rail and ground rail have to be linked. This is done by shorting both C10 and R13 with links. In single-rail operation, the ±12V version is run on +24V and the ±25V version can be run at 48V or 50V. Remember that C8 has to be reversed. The bias network components R3, R4 and C3 can also be used for bias current compensation to null the output offset on the dual-rail units. A total R3+R4 resistance of around 820kΩ should do it. Buffer stop Is that the end of the design? Well not quite, we’ll be doing more buffers, such as the diamond, next month. I also thought it would be a good idea to combine the balanced discrete op amp with a couple of buffers to make an ultra-low noise and low-distortion balanced input amplifier, which will also be discussed next month. Teach-In 8 CD-ROM Exploring the Arduino EE FR -ROM CD ELECTRONICS TEACH-IN 8 FREE CD-ROM SOFTWARE FOR THE TEACH-IN 8 SERIES FROM THE PUBLISHERS OF This CD-ROM version of the exciting and popular Teach-In 8 series has been designed for electronics enthusiasts who want to get to grips with the inexpensive, immensely popular Arduino microcontroller, as well as coding enthusiasts who want to explore hardware and interfacing. Teach-In 8 provides a one-stop source of ideas and practical information. The Arduino offers a remarkably effective platform for developing a huge variety of projects; from operating a set of Christmas tree lights to remotely controlling a robotic vehicle wirelessly or via the Internet. Teach-In 8 is based around a series of practical projects with plenty of information for customisation. The projects can be combined together in many different ways in order to build more complex systems that can be used to solve a wide variety of home automation and environmental monitoring problems. The series includes topics such as RF technology, wireless networking and remote web access. PLUS: PICs and the PICkit 3 – A beginners guide The CD-ROM also includes a bonus – an extra 12-part series based around the popular PIC microcontroller, explaining how to build PIC-based systems. £8.99 INTRODUCING THE ARDUINO • Hardware – learn about components and circuits • Programming – powerful integrated development system • Microcontrollers – understand control operations • Communications – connect to PCs and other Arduinos PLUS... PIC n’MIX PICs and the PICkit 3 - A beginners guide. The why and how to build PIC-based projects Teach In 8 Cover.indd 1 04/04/2017 12:24 PRICE £8.99 Includes P&P to UK if ordered direct from us SOFTWARE The CD-ROM contains the software for both the Teach-In 8 and PICkit 3 series. ORDER YOUR COPY TODAY! JUST CALL 01202 880299 OR VISIT www.epemag.com Practical Electronics | March | 2024 65