Silicon ChipDual Tracking ± 50V Power Supply - April 1990 SILICON CHIP
  1. Outer Front Cover
  2. Contents
  3. Publisher's Letter: Myths and microwave ovens
  4. Feature: Servicing Your Microwave Oven by Leo Simpson
  5. Vintage Radio: Finding receivers from the 1920s by John Hill
  6. Project: Relative Field Strength Meter by Ralph Holland
  7. Project: Build A VOX With Delayed Audio by Darren Yates
  8. Feature: Computer Bits by Jennifer Bonnitcha
  9. Project: Dual Tracking ± 50V Power Supply by John Clarke & Greg Swain
  10. Serviceman's Log: It's an ill wind... as they say by The TV Serviceman
  11. Back Issues
  12. Feature: Taking the BASF CD Challenge by Leo Simpson
  13. Project: 16-Channel Mixing Desk, Pt.3 by John Clarke & Leo Simpson
  14. Feature: Remote Control by Bob Young
  15. Feature: Amateur Radio by Garry Cratt, VK2YBX
  16. Subscriptions
  17. Market Centre
  18. Advertising Index
  19. Outer Back Cover

This is only a preview of the April 1990 issue of Silicon Chip.

You can view 48 of the 104 pages in the full issue, including the advertisments.

For full access, purchase the issue for $10.00 or subscribe for access to the latest issues.

Articles in this series:
  • Computer Bits (July 1989)
  • Computer Bits (July 1989)
  • Computer Bits (August 1989)
  • Computer Bits (August 1989)
  • Computer Bits (September 1989)
  • Computer Bits (September 1989)
  • Computer Bits (October 1989)
  • Computer Bits (October 1989)
  • Computer Bits (November 1989)
  • Computer Bits (November 1989)
  • Computer Bits (January 1990)
  • Computer Bits (January 1990)
  • Computer Bits (April 1990)
  • Computer Bits (April 1990)
  • Computer Bits (October 1990)
  • Computer Bits (October 1990)
  • Computer Bits (November 1990)
  • Computer Bits (November 1990)
  • Computer Bits (December 1990)
  • Computer Bits (December 1990)
  • Computer Bits (January 1991)
  • Computer Bits (January 1991)
  • Computer Bits (February 1991)
  • Computer Bits (February 1991)
  • Computer Bits (March 1991)
  • Computer Bits (March 1991)
  • Computer Bits (April 1991)
  • Computer Bits (April 1991)
  • Computer Bits (May 1991)
  • Computer Bits (May 1991)
  • Computer Bits (June 1991)
  • Computer Bits (June 1991)
  • Computer Bits (July 1991)
  • Computer Bits (July 1991)
  • Computer Bits (August 1991)
  • Computer Bits (August 1991)
  • Computer Bits (September 1991)
  • Computer Bits (September 1991)
  • Computer Bits (October 1991)
  • Computer Bits (October 1991)
  • Computer Bits (November 1991)
  • Computer Bits (November 1991)
  • Computer Bits (December 1991)
  • Computer Bits (December 1991)
  • Computer Bits (January 1992)
  • Computer Bits (January 1992)
  • Computer Bits (February 1992)
  • Computer Bits (February 1992)
  • Computer Bits (March 1992)
  • Computer Bits (March 1992)
  • Computer Bits (May 1992)
  • Computer Bits (May 1992)
  • Computer Bits (June 1992)
  • Computer Bits (June 1992)
  • Computer Bits (July 1992)
  • Computer Bits (July 1992)
  • Computer Bits (September 1992)
  • Computer Bits (September 1992)
  • Computer Bits (October 1992)
  • Computer Bits (October 1992)
  • Computer Bits (November 1992)
  • Computer Bits (November 1992)
  • Computer Bits (December 1992)
  • Computer Bits (December 1992)
  • Computer Bits (February 1993)
  • Computer Bits (February 1993)
  • Computer Bits (April 1993)
  • Computer Bits (April 1993)
  • Computer Bits (May 1993)
  • Computer Bits (May 1993)
  • Computer Bits (June 1993)
  • Computer Bits (June 1993)
  • Computer Bits (October 1993)
  • Computer Bits (October 1993)
  • Computer Bits (March 1994)
  • Computer Bits (March 1994)
  • Computer Bits (May 1994)
  • Computer Bits (May 1994)
  • Computer Bits (June 1994)
  • Computer Bits (June 1994)
  • Computer Bits (July 1994)
  • Computer Bits (July 1994)
  • Computer Bits (October 1994)
  • Computer Bits (October 1994)
  • Computer Bits (November 1994)
  • Computer Bits (November 1994)
  • Computer Bits (December 1994)
  • Computer Bits (December 1994)
  • Computer Bits (January 1995)
  • Computer Bits (January 1995)
  • Computer Bits (February 1995)
  • Computer Bits (February 1995)
  • Computer Bits (March 1995)
  • Computer Bits (March 1995)
  • Computer Bits (April 1995)
  • Computer Bits (April 1995)
  • CMOS Memory Settings - What To Do When The Battery Goes Flat (May 1995)
  • CMOS Memory Settings - What To Do When The Battery Goes Flat (May 1995)
  • Computer Bits (July 1995)
  • Computer Bits (July 1995)
  • Computer Bits (September 1995)
  • Computer Bits (September 1995)
  • Computer Bits: Connecting To The Internet With WIndows 95 (October 1995)
  • Computer Bits: Connecting To The Internet With WIndows 95 (October 1995)
  • Computer Bits (December 1995)
  • Computer Bits (December 1995)
  • Computer Bits (January 1996)
  • Computer Bits (January 1996)
  • Computer Bits (February 1996)
  • Computer Bits (February 1996)
  • Computer Bits (March 1996)
  • Computer Bits (March 1996)
  • Computer Bits (May 1996)
  • Computer Bits (May 1996)
  • Computer Bits (June 1996)
  • Computer Bits (June 1996)
  • Computer Bits (July 1996)
  • Computer Bits (July 1996)
  • Computer Bits (August 1996)
  • Computer Bits (August 1996)
  • Computer Bits (January 1997)
  • Computer Bits (January 1997)
  • Computer Bits (April 1997)
  • Computer Bits (April 1997)
  • Windows 95: The Hardware That's Required (May 1997)
  • Windows 95: The Hardware That's Required (May 1997)
  • Turning Up Your Hard Disc Drive (June 1997)
  • Turning Up Your Hard Disc Drive (June 1997)
  • Computer Bits (July 1997)
  • Computer Bits (July 1997)
  • Computer Bits: The Ins & Outs Of Sound Cards (August 1997)
  • Computer Bits: The Ins & Outs Of Sound Cards (August 1997)
  • Computer Bits (September 1997)
  • Computer Bits (September 1997)
  • Computer Bits (October 1997)
  • Computer Bits (October 1997)
  • Computer Bits (November 1997)
  • Computer Bits (November 1997)
  • Computer Bits (April 1998)
  • Computer Bits (April 1998)
  • Computer Bits (June 1998)
  • Computer Bits (June 1998)
  • Computer Bits (July 1998)
  • Computer Bits (July 1998)
  • Computer Bits (November 1998)
  • Computer Bits (November 1998)
  • Computer Bits (December 1998)
  • Computer Bits (December 1998)
  • Control Your World Using Linux (July 2011)
  • Control Your World Using Linux (July 2011)
Articles in this series:
  • 16-Channel Mixing Desk (February 1990)
  • 16-Channel Mixing Desk (February 1990)
  • 16-Channel Mixing Desk (March 1990)
  • 16-Channel Mixing Desk (March 1990)
  • 16-Channel Mixing Desk, Pt.3 (April 1990)
  • 16-Channel Mixing Desk, Pt.3 (April 1990)
  • 16-Channel Mixing Desk; Pt.4 (May 1990)
  • 16-Channel Mixing Desk; Pt.4 (May 1990)
  • Modifications To The 16-Channel Mixer (November 1990)
  • Modifications To The 16-Channel Mixer (November 1990)
Articles in this series:
  • Remote Control (October 1989)
  • Remote Control (October 1989)
  • Remote Control (November 1989)
  • Remote Control (November 1989)
  • Remote Control (December 1989)
  • Remote Control (December 1989)
  • Remote Control (January 1990)
  • Remote Control (January 1990)
  • Remote Control (February 1990)
  • Remote Control (February 1990)
  • Remote Control (March 1990)
  • Remote Control (March 1990)
  • Remote Control (April 1990)
  • Remote Control (April 1990)
  • Remote Control (May 1990)
  • Remote Control (May 1990)
  • Remote Control (June 1990)
  • Remote Control (June 1990)
  • Remote Control (August 1990)
  • Remote Control (August 1990)
  • Remote Control (September 1990)
  • Remote Control (September 1990)
  • Remote Control (October 1990)
  • Remote Control (October 1990)
  • Remote Control (November 1990)
  • Remote Control (November 1990)
  • Remote Control (December 1990)
  • Remote Control (December 1990)
  • Remote Control (April 1991)
  • Remote Control (April 1991)
  • Remote Control (July 1991)
  • Remote Control (July 1991)
  • Remote Control (August 1991)
  • Remote Control (August 1991)
  • Remote Control (October 1991)
  • Remote Control (October 1991)
  • Remote Control (April 1992)
  • Remote Control (April 1992)
  • Remote Control (April 1993)
  • Remote Control (April 1993)
  • Remote Control (November 1993)
  • Remote Control (November 1993)
  • Remote Control (December 1993)
  • Remote Control (December 1993)
  • Remote Control (January 1994)
  • Remote Control (January 1994)
  • Remote Control (June 1994)
  • Remote Control (June 1994)
  • Remote Control (January 1995)
  • Remote Control (January 1995)
  • Remote Control (April 1995)
  • Remote Control (April 1995)
  • Remote Control (May 1995)
  • Remote Control (May 1995)
  • Remote Control (July 1995)
  • Remote Control (July 1995)
  • Remote Control (November 1995)
  • Remote Control (November 1995)
  • Remote Control (December 1995)
  • Remote Control (December 1995)
Articles in this series:
  • Amateur Radio (November 1987)
  • Amateur Radio (November 1987)
  • Amateur Radio (December 1987)
  • Amateur Radio (December 1987)
  • Amateur Radio (February 1988)
  • Amateur Radio (February 1988)
  • Amateur Radio (March 1988)
  • Amateur Radio (March 1988)
  • Amateur Radio (April 1988)
  • Amateur Radio (April 1988)
  • Amateur Radio (May 1988)
  • Amateur Radio (May 1988)
  • Amateur Radio (June 1988)
  • Amateur Radio (June 1988)
  • Amateur Radio (July 1988)
  • Amateur Radio (July 1988)
  • Amateur Radio (August 1988)
  • Amateur Radio (August 1988)
  • Amateur Radio (September 1988)
  • Amateur Radio (September 1988)
  • Amateur Radio (October 1988)
  • Amateur Radio (October 1988)
  • Amateur Radio (November 1988)
  • Amateur Radio (November 1988)
  • Amateur Radio (December 1988)
  • Amateur Radio (December 1988)
  • Amateur Radio (January 1989)
  • Amateur Radio (January 1989)
  • Amateur Radio (April 1989)
  • Amateur Radio (April 1989)
  • Amateur Radio (May 1989)
  • Amateur Radio (May 1989)
  • Amateur Radio (June 1989)
  • Amateur Radio (June 1989)
  • Amateur Radio (July 1989)
  • Amateur Radio (July 1989)
  • Amateur Radio (August 1989)
  • Amateur Radio (August 1989)
  • Amateur Radio (September 1989)
  • Amateur Radio (September 1989)
  • Amateur Radio (October 1989)
  • Amateur Radio (October 1989)
  • Amateur Radio (November 1989)
  • Amateur Radio (November 1989)
  • Amateur Radio (December 1989)
  • Amateur Radio (December 1989)
  • Amateur Radio (February 1990)
  • Amateur Radio (February 1990)
  • Amateur Radio (March 1990)
  • Amateur Radio (March 1990)
  • Amateur Radio (April 1990)
  • Amateur Radio (April 1990)
  • Amateur Radio (May 1990)
  • Amateur Radio (May 1990)
  • Amateur Radio (June 1990)
  • Amateur Radio (June 1990)
  • Amateur Radio (July 1990)
  • Amateur Radio (July 1990)
  • The "Tube" vs. The Microchip (August 1990)
  • The "Tube" vs. The Microchip (August 1990)
  • Amateur Radio (September 1990)
  • Amateur Radio (September 1990)
  • Amateur Radio (October 1990)
  • Amateur Radio (October 1990)
  • Amateur Radio (November 1990)
  • Amateur Radio (November 1990)
  • Amateur Radio (December 1990)
  • Amateur Radio (December 1990)
  • Amateur Radio (January 1991)
  • Amateur Radio (January 1991)
  • Amateur Radio (February 1991)
  • Amateur Radio (February 1991)
  • Amateur Radio (March 1991)
  • Amateur Radio (March 1991)
  • Amateur Radio (April 1991)
  • Amateur Radio (April 1991)
  • Amateur Radio (May 1991)
  • Amateur Radio (May 1991)
  • Amateur Radio (June 1991)
  • Amateur Radio (June 1991)
  • Amateur Radio (July 1991)
  • Amateur Radio (July 1991)
  • Amateur Radio (August 1991)
  • Amateur Radio (August 1991)
  • Amateur Radio (September 1991)
  • Amateur Radio (September 1991)
  • Amateur Radio (October 1991)
  • Amateur Radio (October 1991)
  • Amateur Radio (November 1991)
  • Amateur Radio (November 1991)
  • Amateur Radio (January 1992)
  • Amateur Radio (January 1992)
  • Amateur Radio (February 1992)
  • Amateur Radio (February 1992)
  • Amateur Radio (March 1992)
  • Amateur Radio (March 1992)
  • Amateur Radio (July 1992)
  • Amateur Radio (July 1992)
  • Amateur Radio (August 1992)
  • Amateur Radio (August 1992)
  • Amateur Radio (September 1992)
  • Amateur Radio (September 1992)
  • Amateur Radio (October 1992)
  • Amateur Radio (October 1992)
  • Amateur Radio (November 1992)
  • Amateur Radio (November 1992)
  • Amateur Radio (January 1993)
  • Amateur Radio (January 1993)
  • Amateur Radio (March 1993)
  • Amateur Radio (March 1993)
  • Amateur Radio (May 1993)
  • Amateur Radio (May 1993)
  • Amateur Radio (June 1993)
  • Amateur Radio (June 1993)
  • Amateur Radio (July 1993)
  • Amateur Radio (July 1993)
  • Amateur Radio (August 1993)
  • Amateur Radio (August 1993)
  • Amateur Radio (September 1993)
  • Amateur Radio (September 1993)
  • Amateur Radio (October 1993)
  • Amateur Radio (October 1993)
  • Amateur Radio (December 1993)
  • Amateur Radio (December 1993)
  • Amateur Radio (February 1994)
  • Amateur Radio (February 1994)
  • Amateur Radio (March 1994)
  • Amateur Radio (March 1994)
  • Amateur Radio (May 1994)
  • Amateur Radio (May 1994)
  • Amateur Radio (June 1994)
  • Amateur Radio (June 1994)
  • Amateur Radio (September 1994)
  • Amateur Radio (September 1994)
  • Amateur Radio (December 1994)
  • Amateur Radio (December 1994)
  • Amateur Radio (January 1995)
  • Amateur Radio (January 1995)
  • CB Radio Can Now Transmit Data (March 2001)
  • CB Radio Can Now Transmit Data (March 2001)
  • What's On Offer In "Walkie Talkies" (March 2001)
  • What's On Offer In "Walkie Talkies" (March 2001)
  • Stressless Wireless (October 2004)
  • Stressless Wireless (October 2004)
  • WiNRADiO: Marrying A Radio Receiver To A PC (January 2007)
  • WiNRADiO: Marrying A Radio Receiver To A PC (January 2007)
  • “Degen” Synthesised HF Communications Receiver (January 2007)
  • “Degen” Synthesised HF Communications Receiver (January 2007)
  • PICAXE-08M 433MHz Data Transceiver (October 2008)
  • PICAXE-08M 433MHz Data Transceiver (October 2008)
  • Half-Duplex With HopeRF’s HM-TR UHF Transceivers (April 2009)
  • Half-Duplex With HopeRF’s HM-TR UHF Transceivers (April 2009)
  • Dorji 433MHz Wireless Data Modules (January 2012)
  • Dorji 433MHz Wireless Data Modules (January 2012)
Dual tracking ± 50V power supply Looking for a dual tracking power supply that can really deliver the goods? This switchmode design can provide up to ± 50V DC and features a LED dropout indicator and short circuit protection. By JOHN CLARKE & GREG SWAIN There are many situations where you need a high voltage dual tracking power supply. For example, this design could be used to power prototype audio amplifiers or high voltage op amp circuits, or it could be used in any application requiring a high voltage rail of up to 100V (eg, for servicing some TV sets). Perhaps the most important feature of the design is its tracking positive and negative DC outputs. These are adjustable from 0V to ± 50V (ie, the supply can deliver up to 100V DC) and are fully protected against short circuits and voltages generated by external loads. By contrast, most other currently available dual tracking power supplies only go to about ± 20V or perhaps ± 30V. 46 SILICON CHIP Another important feature is the output current capability. Take a look at Fig, 1. This shows the maximum load current for output voltages between 0 and 100V (ie, 0 to ± 50V). As shown, the maximum current is 1. 7A between 0V and ± 43.5V. After that, the output current capability is reduced but the supply can still deliver 1A at ± 50V (ie, 100V). The line and load regulation figures are very good. On the prototype, the line regulation was within ± 5mV of a given output voltage setting for mains voltages between 220V AC to 260V AC. The load regulation is within 500mV at ± 50V for an output current 1A (ie, 1 % ). The ripple output, which consists of 100Hz and other noise superimposed on the DC rails, is less than 3mV p-p for all load currents up to 1.7 A. A LED indicator on the front panel lights when the output ripple exceeds 5mV p-p. This indicates that the supply rails are beginning to drop out of regulation. Front panel hardware There are just three basic controls on the front panel: a Power ON/OFF switch, a Load switch and a 10-turn Voltage Adjust pot. The 10-turn pot makes it easy to adjust the output voltage to precisely the value you want. A conventional pot would be much too coarse in this position, so don't be tempted to save a few bucks. Th€! Load switch is a handy facility. It allows you to switch power to the load without having to switch the supply itself on or off. In operation, the Load switch simply switches the output rails to the positive and negative output terminals. Also on the front panel are a power on/off LED indicator, the aforementioned Dropout LED indicator, and four binding post output terminals (plus, minus, 0V and Specifications 10011 Type Dual tracking with switchmode preregulators for high efficiency Output Voltage 0 to ±50V Output Current 1.7A from Oto 87V; 1.5A at 91V; 1A at 1 OOV Tracking Accuracy Within ±1 % Load Regulation Better than 500mV at ±50V and 1 A Line Regulation Better than ± 5mV for mains voltages from 220-260VAC Ripple output Less than 3mV p-p at full load Protection Fully protected against output short circuits and forward and reverse voltages connected to the output; fuse protection for the power transformer GND). The rear panel carries a large finned heatsink, the mains cord entry and a panel-mount fuse holder. In practice, you can use the power supply in several different ways. You can use it to provide balanced positive and negative rails; you can take the output from between the positive and negative output terminals to obtain up to 100V DC output; or you can use it to obtain a single supply rail. Because the circuit is fully floating [ie, not tied to mains earth), the output can be referenced to earth by tying either the positive, negative or 0V terminals to the GND terminal. Our last dual tracking power supply was described in the 1.7 Fig.2: basic adjustable positive regulator circuit. The LM317 maintains a constant 1.25V between its OUT and ADJ terminals, which means that 12.5mA flows through the 1000 resistor and VR1 at all times. \ 1.5 I\ 0.5 I I 10 20 I 30 I 40 I I 50 60 I 70 I I 80 90 OUTPUT -1.25V REFERENCE January 1988 issue of SILICON CHIP. That unit was capable of supplying 0 to ± 18.5V at about 1-amp and was based on adjustable positive and negative 3-terminal regulators [the LM317 and the LM337). Our new design also uses LM317 and LM337 3-terminal regulators but there is quite a bit more to it than that as we shall see. Let's first take a brief look at how these devices work. Fig.2 shows an adjustable positive regulator circuit based on an LM317. Capacitor Cl is used to filter the DC input to the regulator, while potentiometer VR4 adjusts the output voltage. In operation, the LM317 produces a nominal 1.25V between its adjust [ADJ) terminal and the output [OUT) terminal. The 1000 resistor between these terminals thus has Design considerations +1.25V C1 100 OUTPUT VOLTAGE Fig.1: maximum output current vs. output voltage. The supply can deliver 1.7A for outputs up to 87V and 1A at 100V (ie, ± 50V). 1.25V across it which means that 12.5mA flows through VRl at all times When VRl is set to on, the output voltage [ie, at the OUT terminal) sits 1.25V above point A. This point is set at - 1.25V by a - 1.25V reference circuit, and so the output sits at 0V with minimum VRl. As the resistance of VRl increases, the voltage on the ADJ terminal is " jacked up" and so the output voltage also increases. OK, so what do we have to do to obtain a high voltage, high current dual tracking supply? Unfortunately, you can't just use high voltage regulators and substitute a bigger transformer, bigger heatsink and larger filter capacitors. The problem is that as the power dissipation in the device increases, its temperature also increases and the device shuts itself down by current limiting. This means that the amount of current that you can obtain using the circuit of Fig.2 is severely limited at low output voltages due to the high voltage developed across the regulator. Pre-regulator circuit Fig.3 shows how we solved this problem by employing a switchmode pre-regulator circuit ahead of the 3-terminal regulator. The pre-regulator acts to minimise the input voltage to the regulator and thus reduces power dissipation for a given voltage and current setting. The result of this scheme is that the regulator can supply heaps more current. It also means that we can now use the cheaper lowvoltage 3-terminal regulators inA PRIL 1990 47 REGULATOR L1 01 100n Cl JUUL PULSE WIDTH MODULATED SIGNAL stead of the more expensive high voltage units. Now take a closer look at Fig.3. Qt, D5, Lt and C2 form a basic switchmode circuit. What happens is that Qt is switched on and off rapidly by a pulse waveform into its base. If the pulse waveform has a short duty cycle (ie, the transistor is off most of the time), very short current pulses will be fed to 11 and the resultant DC voltage across C2 will be low. Conversely, if the duty cycle is high, the transistor will be on most of the time and the DC voltage across C2 will be high. By switching Qt on and off to control the output voltage, its power dissipation is low and the overall circuit efficiency is high. D5 protects Qt against the inductive kickback from 11 when the transistor switches off. So the voltage at the input of the regulator is controlled simply by adjusting Qt 's duty cycle. The switching pulses to Qt 's base are provided by a high gain er- VT /\ /\ V[/I'<iVMK ror amplifier circuit consisting of IC3c, IC2b and Q3. This circuit monitors the voltage across the regulator and then generates a pulse width modulated (PWM) feedback signal at the output of comparator IC2b. This signal then switches Q3 which in turn controls Qt. As shown in Fig.3, the inverting input of op amp IC3c is fed from a voltage divider network consisting of ZD7, R2 and R3. This network is connected between the regulator input and OV. Similarly, the noninverting input is fed from voltage divider network R4 and R5 which is connected between the regulator output and OV. Because equivalent values are used for the resistors in each divider network (ie, R2 = R4 & R3 = R5), the inputs to IC3c are equal only when the regulator's input is 4.7V greater than its output. This 4.7V differential is necessary to compensate for the 4. 7V drop introduced by ZD7 in one of the divider networks. IC3c thus effectively monitors the L<T/\/\ "'I nf VP-1...____.~ . _ _ _ _ _ . ~ ~ r (a) HIGH VOLTAGE (b) LOW VOLT AGE Fig,4: how the error voltage VE and the triangle waveform VT interact. The higher the error voltages, the wider the voltage pulses (Vp) produced at the output of IC2b. 48 SILICON CHIP Fig,3: adding a preregulator circuit drastically reduces the power dissipation in the LM317 for a given output. IC3c monitors the voltage across the regulator & produces a DC error voltage which is fed to IC2b. IC2b compares this error voltage with a constant triangular waveform and produces a pulse width modulated (PWM) signal to drive transistors Q3 & Qt. voltage across the 3-terminal regulator and generates an appropriate error voltage (VE), If the voltage across the regulator begins to fall below 4.7V, the error voltage goes up. If the voltage across the regulator begins to rise, the error voltage goes down. This error voltage is applied to the non-inverting input of IC2b where it is compared with a triangular waveform applied to IC2b's inverting input. This triangular waveform (VT) is derived from an oscillator circuit (not shown in Fig.3). Since IC2b is wired as a comparator its output can only be high or low, so when VE is above VT the output will be high and when VE is below VT, the output will be low. The interaction of VE and VT via IC2b is shown in Fig.4. Fig.4(a) shows that when VE is high (ie, the voltage across the regulator is beginning to fall), the output from IC2b is a series of fairly wide pulses (Vp). Thus, Q3 and therefore Qt will be pulsed on for fairly long periods of time and this will increase the pre-regulator output voltage. Similarly, if VE is low as in Fig.4b, the output from IC2b will consist of a series of narrow pulses and Qt 's duty cycle will be low. What this all means is that Q 1 is pulsed on and off at exactly the correct rate to give the required input voltage to the regulator. If the voltage across the regulator begins to move in either direction away from 4.7V, the pulse signal at the output of IC2b automatically adjusts to switch Ql on for longer or shorter periods as required. Circuit details Now take a look at Fig.5 which shows all the circuit details. While this may appear daunting at first sight, all the circuit elements in Fig.3 can be directly related to Fig.5. The main difference between these two diagrams is that Fig.5 also includes all the circuitry necessary for the negative supply rail. Its pre-regulator circuit operates in a similar fashion to that used for the positive rail but uses a PNP driver transistor (Q4) and an NPN switching transistor [Q2). IC3b is the error amplifier for the negative rail, IC2a the comparator, Q4 the driver transistor, Q2 the The supply is easy to wire but you should take extra care with the mains wiring. Use a cord-grip grommet to secure the mains cord and sleeve the switch and fuse terminals with heatshrink tubing. The finned heatsink on the rear panel ensures adequate cooling for the power devices. main switching transistor, and L2 the inductor. The output of this pre-regulator circuit drives an LM337 negative 3-terminal regulator. Power for the circuit is derived from a 160VA toroidal transformer with an B0V centre-tapped secondary winding. This drives bridge rectifier Dl-D4 which, combined with the four l000µF filter capacitors, gives ± 60V DC rails. Short circuit protection for the transformer is provided by a 250mA fuse in the mains active lead and a 2A fuse in each leg of the secondary winding. The ± 60V rails are connected directly to the emitters of switching transistors Ql and Q2. These are BD650 and BD649 Darlington devices which have a collector to emitter voltage rating of 100V and require only about 12mA of base current to deliver 3A. In addition to the ± 60V rails, there are five other supply rails in the circuit: ± 30V, ± 15V and + 12V. These supply rails are derived using zener diodes ZD1-ZD5. The ± 15V rails feed dual comparator IC2 and quad op amp IC3, while the ± 30V rails feed IC4, a high voltage op amp. The + 12V APRIL 1990 49 F2 2A F1 25 0mA +60V +30V _ _ _,......+.... 15.,.v_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _..,..._ _ _ _ +15V pm A N 39k 1000 63VW E + 1000 + - 63VW _ TRIANGULAR WAVEFORM +l 2V GENERATOR 10k 24 0VAC ltl 7555 1W -15V 1M + ZD1 LM336 -2.5 J - A 100k COMPARATOR 100k 1.Bk E +15V A POWER LE01 10k 10k 04 BC640 330 63VW A X 1M OV OUTPUT SET VR1 10k 1000 63VW + 1000 + _ 63VW - -15V 2.7k -1.2V OFFSET AMPLIFIER 10k ERROR SIGNAL 06 BY229 10k -15V ,w -30V :t50V, 1.5A DUAL TRACKING POWER SUPPLY BC546,BC556 VIEWED FROM BELOW rail is used to power ICl. ICl is the triangle waveform generator. It consists of a CMOS 7555 timer which runs at 10kHz. This frequency is set by the 15k0 and 4 7k0 resistors and by the .0015µF capacitor on pins 6 and 2. 50 SILICON CHIP Because the circuit is wired in astable mode, the .0015µF capacitor repeatedly charges to 2/3Vcc and discharges to 1/3Vcc (Vee = the 7555 supply rail). Th_e resulting triangular waveform developed across the capacitor is then fed to pins 2 and 6 of IC2a and IC2b via 10k0 isolating resistors. These stages compare the triangle waveform with the error signals from IC3c and IC3b to produce the pulsed waveforms to drive Q3 and Q4 as described previously. + - 2.2M DROPOUT LED2 15k 4x 1N4148 LOAD LM317T IN OUT S2a +0-50V ~ ADJ 100n ZD7 4cio1.:w j 100 63VW 100k . - 22k 1% 15k 39k 15k ERROR SIGNAL 470 50VW SET 50V OUT VR2 10k 330 63VW 03 BC639 10 50VW BP • A 010 1N4002 METER CAL VR3 20k E 100k 2.7k 2.7k -15V DV +15V 2.7k 2.7k +3DV GNO 100k J, CASE + 10k 100k 0-SOV \ 470 sovw + 15k 011 1N4002 100 63VW + _ 22k l¾ 15k 012 1N4002 L4 LOAD 0-0 S2b -0-SOV LM337T LH4 : 33T. 0.5mm ENCU ON NEOSIO 17-732-22 TOROID 013 1N4002 A CMOS 7555 timer must be used for ICl, by the way. Don't substitute a standard 555 timer, as this will impose discharge spikes on the supply line and disrupt the operation of the switchmode circuitry. ICZa and ICZb are part of an Fig.5: the final circuit uses the parts depicted in Fig.3 to derive the positive rail, while the negative rail is derived using an LM337 negative regulator and a similar switching pre-regulator arrangement. IC4 ensures that the negative rail tracks the positive rail. LM393 dual op amp package. These have open collector outputs which means that pins 1 and 7 can switch to the - 15V supply. Because of this, diode D7 is included to prevent the base of Q3 from being pulled APRIL 1990 51 I. ± SOV 1.SA DUAL TRACKING PC • DROPOUT • POWER VOLTAGE ADJUST • • L This artwork can be used as a drilling template for the front panel. The meter is supplied with its own mounting template. negative when pin 7 of IC2b goes low. The 10k0 resistor between Q3's base and the OV rail ensures that the transistor completely switches off. No diode is necessary for IC2a, since its output will only go to OV. This is by virtue of the 10kn pullup resistor connected to the OV rail. Offset amplifier IC3a provides the - 1.25V offset voltage for the LM317 regulator. This op amp is connected as an inverter, with feedback via the 2.7k0 resistor connected between pins 1 and 2. Its non-inverting input is connected to OV via a 1.BkO resistor, while the inverting input samples the voltage from ZD1 via trimpot VRl. ZD1 is an LM336 precision 2.5V reference diode. IC3a simply inverts and attenuates this reference to provide the nominal - 1.25V offset voltage which is applied to the bottom of voltage adjust pot VR4; ie, IC3a operates with a gain of - 0.5 . In practice, VR1 is adjusted so that the output voltage is at OV when VR4 is at minimum setting. Kick start circuit The circuit shown in Fig.3 won't start up when power is initially applied. The problem is, Ql cannot switch on and charge C2 until a pulse signal appears at the output of IC2b. And this pulse signal can52 SILICON CHIP not be generated until IC3c generates an error voltage which in turn cannot be generated until voltage appears at the input to the regulator. So we have a classic catch 22 situation - Ql cannot be pulsed on to give an output because there is no output in the first place. The way around this dilemma is to "kick start" the circuit by fitting 100k0 resistors between the supply rails and the inverting inputs of error amplifiers IC3b and IC3c. Take a look at IC3c on Fig.5. It has a 100k0 resistor fitted between pin 9 and the - 15V rail. Now, when power is applied, pin 9 is initially pulled to - 15V and so a large error voltage appears at the output (pin 8). IC2b thus produces a pulse waveform with a high duty cycle and so Ql quickly charges the 330µF capacitor on the regulator input to the required voltage. As the voltage to the regulator input rises, the voltage on pin 9 of IC3c also rises and so the error voltage decreases and the circuit quickly stabilises. The negative rail pre-regulator is kick started in similar fashion by using a lOOkn resistor to pull pin 6 of IC3b to + 15V. As well as providing the kick start facility, the 100kn resistors also have the effect of increasing the input/output differential applied to the 3-terminal regulators. This is because the resistors provide an additional voltage offset at the inverting inputs of the error amplifiers. As a result, the circuit of Fig.5 stabilises when the differential voltage across the regulators is 8V instead of 4. 7V as is the case for Fig.3. Voltage adjustment VR4 provides the output voltage adjustment for the positive supply rail. This is wired in parallel with a series pair consisting of trimpot VR2 and a 15k0 resistor. VR2 sets the maximum output voltage and is adjusted to give exactly ± 50V out when VR4 is at maximum setting. OK, so the output of the positive regulator is adjusted using VR4 but what about the negative regulator'? It doesn't have a potentiometer on its ADJ terminal but uses a voltage tracking circuit consisting of IC4, D14, D15, Q5 and Q6 instead. IC4 is an LM344 high voltage op amp while Q5, Q6, D14 and D15 form an output buffer stage to provide the necessary lOmA drive to the LM337. Q5 and Q6 drive the ADJ terminal of the LM337 via two 1.2k0 resistors in series. Because a current of 10.4mA flows through the 1200 resistor between the OUT and ADJ terminals, it follows that the voltage across the two 1.2kn resistors is 25V. This allows IC4 ,weR SUPPLY 7 • LOAD • ON GND • • ov + _J and its buffer stage to drive the ADJ terminal of the 1M337 to - 48.75V in spite of the fact that the negative supply rail to the op amp is only -30V. In practice, the output of the op amp buffer stage (ie, the junction of the two 330 resistors) swings between about + 26.25V and - 23.75V to provide the full O to - 50V range of the negative supply output. So how does the circuit work? The idea behind the op amp circuit is to provide a mirror of the voltage on the positive supply rail. IC4, its output buffer stage Q5 and Q6, and the 1M337 regulator can all be regarded as a power op amp. This op amp can be regarded as an inverter with a gain of minus one, set by the 22k0 resistor to the positive output rail and a second 22k0 resistor connected to the negative output rail. Thus, if the positive output rail is at + 50V, the negative rail will automatically be at - 50V. 13 and its associated 470µF capacitor filter the output from the positive regulator while 14 and another 470µF capacitor filter the negative regulator output. These two filter networks remove any residual switchmode ripple due to radiation from the pre-regulators into the positive and negative output supply lines. The output voltage of the supply is monitored by a lmA FSD meter PARTS LIST 1 PCB, code SC04104901, 167 x 126mm 1 plastic instrument case, 260 x 1,90 x 80mm (Altronics Cat. H-0482 2 aluminium panels to suit case (Altronics Cat. H-0488) 1 front panel label, 255 x 73mm 1 80V 2A centre-tapped toroidal mains transformer (Altronics Cat. M-3075 or equivalent) 1 heatsink, 110 x 33 x 72mm (Altronics Cat. H-0560 or equivalent) 1 pushbutton mains switch with plastic body 4 Neosid 17 · 732-22 toroids (Altronics Cat. L-511 0) 1 DPDT toggle switch 1 red binding post 1 black binding post 1 white binding post 1 green binding post 1 panel-mounting 3AG fuse holder 4 PC-mounting 3AG fuse clips 1 250mA 3AG fuse 2 2A 3AG fuses 3 solder lugs 1 mains cord grip grommet 1 3-way mains terminal strip 1 mains lead & plug 1 MU52E 1 mA panel meter with 0·50V scale (Altronics Cat. 0-0538) 1 5k0 1 0-turn potentiometer 1 20k0 horizontal trimpot 2 1 OkO horizontal trim pots 4 T0-220 mica washers and insulating bushes 1 95 x 125mm metal plate 3 300mm lengths of heavy duty hookup wire (red, green, blue) 4 250mm lengths of medium duty hookup wire (red, yellow, brown, white) 1 500mm length of green/yellow mains earth wire 500mm length of brown mains active wire 4 1-metre lengths of 0.8mm enamelled copper wire Semiconductors 1 LM31 7T positive adjustable 3-terminal regulator LM337T negative adjustable 3-terminal regulator 1 7555 CMOS timer (IC1) (do not substitute a 555) 1 LM393 dual comparator (IC2) 1 LF34 7 , TL07 4 quad op amp (IC3) LM344H high voltage op amp (IC4) - (available from Geoff Wood Electronics) 1 BD650 PNP Darlington transistor (01) BD649 NPN Darlington transistor (02) 1 BC639 NPN transistor (03) 1 BC640 PNP transistor (04) 1 BC546 NPN transistor (05) 1 BC556 PNP transistor (06) 1 LM336 -2.5V reference diode (ZD1) 4 15V 1 W zener diodes (ZD2-ZD5) 1 1 2V 400mW zener diode (ZD6) 2 4 . 7V 400mW zener diodes (ZD7,ZD8) 4 1 N5404 3A diodes (D1 -D4) 2 BY229, MUR1550 fast recovery diodes (D5,D6) 7 1 N41 48 diodes (D7 ,D14,D15,D16-D19) 6 1 N4002 1 A diodes (D8-D13) 2 5mm red LEDs (LED1 ,LED2) Capacitors 4 1 OOOµF 63VW PC electrolytic 2 470µF 50VW PC electrolytic 2 330µF 63VW PC electrolytic 2 1 OOµF 63VW PC electrolytic 2 1 OµF 50VW non-polarised PC electrolytic 1 0 .15µF 1 OOV metallised polyester 1 0.1 µF 100V metallised polyester 1 .0015µF metallised polyester 1 22pF ceramic Resistors (0.25W, 1 2 .2MO 2 2 1MO 5 6 1 OOkO 1 1 47k0 1 2 39k0 2 2 22k0 1 % 1 8 1 5k~ 2 9 10k0 1 1 1 OkO 1 W 1 2 6.8k0 2 1 4.7k0 5%) 4.7k0 1W 2 .7k0 1.8k0 1.2k0 1.2k0 0.5W 1 kO 5600 5W 1200 1000 330 Miscellaneous Screws , nuts, star washers, heatshrink tubing, solder etc. APRIL 1990 53 The 3-terminal regulators and switching transistors are bolted to (but electrically isolated from) the rear panel (see Fig.7). Their mounting bolts are also used to secure the finned heatsink. (with a 50V scale) connected across the positive and negative rails via a 39kn resistor and trimpot VR3. Diodes DB, D9, D10, Dll, D12 and D13 protect the regulators from reverse voltages which may be generated by capacitive or inductive loads connected across the outputs. Drop-out indicator When the regulators are working as designed, the ripple voltage superimposed on the DC rails will be very low. However, if the current drain is higher than the regulator can supply while still maintaining about BV between its input and output terminals, the ripple voltage will suddenly become quite high and the output voltage will fall. IC3d detects the onset of this condition and flashes a warning LED indicator. This op amp is wired as an inverting amplifier and monitors both the positive and negative regulators via 15kf2 resistors and a 0.15µF capacitor. The amplified ripple at the output of IC3d is fed to a full wave re.ctifier consisting of diodes D16-D19. When the ripple on one of the regulator outputs exceeds about 30mV, the output of IC3d swings sufficiently 54 SILICON CHIP high and low to drive LED 2 via the rectifier circuit. Construction The ± 50V 1.5A Dual Tracking Power Supply is housed in a standard plastic instrument case measuring 260 x 170 x 80mm. Most of the components are mounted on a PCB coded SC 04104901, while the transformer is mounted on a metal baseplate measuring 95 x 125mm. Metal front and rear panels are also used to provide power supply earthing and heatsinking for the 3-terminal regulators and Darlington switching transistors. In addition, a large finned heatsink is mounted on the rear panel to ensure adequate cooling for the power devices. Fig.6 shows the wiring details. Start construction with assembly of the PCB. Install PC pins at all external wiring points first, then install the fuse clips, wire links, resistors and trimpots. The two 5W resistors should be installed 1-Zmm proud of the PCB so that air can circulate under them for cooling. Next, install the semiconductors on the PC board. Make sure that each device is correctly oriented before soldering it in place. Note in particular that diodes D5 and D6 (BY229) face in opposite directions. Zener diodes ZD2-ZD4 should be mounted with a small loop in one end to provide stress relief under changing temperature conditions. IC4 (LM334H) is in a round metal TO-5 package. You will have to align its 1-4 leads and its 5-8 leads as shown in Fig.6 so that the leads fit the PCB. The tab on the bodv of the IC is adjacent to pin 8. Note that IC3 faces in the opposite direction to ICl and IC2. Ql, Q2 and the two 3-terminal regulators should all be mounted at full lead length and with their metal tabs facing away from the PCB. Make sure that you don't mix these devices up - you will run into big trouble if you do. Construction of the PCB can now be completed by installing the capacitors and making up and fitting the inductors. All the inductors (Ll-L4) are the same and consist of 33 turns of 0.5mm enamelled copper wire on a Neosid 17-732-22 toroid. You will need about 1 metre of wire for each coil. The best procedure is to first slide the toroid half way along the wire, then wind 16 ½ turns using one end of the wire. The remaining 16½ turns can then be wound using the other end and the completed inductor dipped in epoxy resin (eg, Araldite). This will prevent the windings from coming loose and also reduce the winding buzz from Ll and L2. Clean and tin the leads before installing L1 and L2 on the PCB (L3 and L4 are left till later). Drilling the case The case can now be drilled to take the various items of hardware. Most kits will be supplied with a pre-drilled front panel but if you're working from scratch, you will have to use the published artwork (or a self-adhesive label) as a drilling template. If you have a selfadhesive label, it's best to affix this to the front panel before drilling. Drill small pilot holes initially, then carefully ream out each hole to the correct size. The meter cutout can be made by drilling a series of small holes just inside the circumference of the marked circle and then filing the resultant cutout CORD GRIP GROMMET METAL REAR PANEL Q E~niH Al _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __j - (GREEEN/ ) MOUNTING PLATE A EJ\RTH POWER TRANSFORMER <at> © ~½ S1 LUG 470uF A~ LED1 LE02 METAL FRONT PANEL Fig.6: take care to ensure that all polarised parts are installed with correct polarity on the PC board. You can use medium-duty cable to wire up the meter, Voltage Adjust pot and the LEDs but the remaining wiring must be run using heavy-duty (23 x 0.2mm) cable. APRIL 1990 55 ;#"' ii This close-up view is of one inductors on the PC board. Wind the turns on tightly and dip the finished inductor in epoxy adhesive (eg, Araldite) to stop winding buzz before mounting it in position. The bottom CRO trace shows the triangular waveform produced at pin 6 of 555 timer IC1 and applied to the inverting inputs of IC2a and IC2b. The top trace shows the waveform on the collector of switching transistor Ql. to a smooth circle. The meter is supplied with its own mounting template. On the rear panel, you will need to drill holes to accept the mains fuse (Fl ], the cord grip grommet and the earth lug. Fig.6 shows the locations of these holes. You will also have to drill holes to accept the mounting bolts for the 3-terminal regulators and switching transistors. These same mounting bolts are also u:sed to secure the finned heatsink. The locations of these holes can be determined by mounting the PCB on the integral standoffs inside the case and sliding the rear panel into position. Bend the leads of the power devices so that their tabs sit flat against the panel, then check that the heatsink can be positioned so that the mounting holes will go between the fins. Adjust the power devices as necessary to achieve this, then mark the locations of the holes. This done, drill one hole and bolt the heatsink in position. The remaining holes can then be drilled at the appropriate locations from the heatsink side. That way, there's no danger of a hole running into a fin. Carefully deburr all holes on the inside surface of the panel and smear heatsink compound on the mating surface of the heatsink. Once drilling is completed, the various items of hardware can be mounted on the front and rear 56 SILICON CHIP panels. Note that the black anodising on the panels can form a good insulator so scrape this away from around the mounting holes for the earth lugs to ensure a good contact. The earth lug on the front panel is secured by one of the meter mounting screws. The power devices must all be electrically isolated from the rear panel using mica washers and insulating bushes. Fig.7 shows the mounting details for each device. As before, smear the mounting surfaces of these devices with heatsink compound before sliding the panel into position and installing the heatsink and mounting screws. When the devices have all been MICA WASHER \ SCREW ! ~ J PCB l METAL REAR PANEL ' FINNED HEATSINK Fig.7: mounting details for the switching transistors and 3-terminal regulators. The metal tab of each device must be electrically isolated from the rear panel. secured, switch your multimeter to a low ohms range and check that the metal tabs of the power devices are all isolated from the metal panel. If you encounter a short, clear the problem before proceeding further. Transformer mounting The transformer mounting plate can now be drilled to accept the transformer, mains terminal block, earth lug and mounting screws. Fig.6 shows the location of all these parts. Note that one of the mounting screws (at back right] is not shown in Fig.6 but you can see where it goes from the photographs. You will also have to drill corresponding holes to accept the four mounting screws in the base of the case. Check that everything fits in the case before marking out these holes. If the mounting plate is too far forward, the transformer will foul the meter. The remaining hardware can now be installed in the case and the wiring completed. Medium-duty hook-up wire can be used to connect the meter, potentiometer and the two LEDs but the remaining wiring must be run using heavy-duty insulated cable (eg, 23 x 0.2mm]. This includes the wiring to the output terminals and to the load switch (S2). Inductors 13 and 14 and their two associated 470µ,F capacitors are mounted adjacent to the Load 0 NOG 0 The PC board is coded SC 04104901 and measures 167 x 126mm. switch to ensure maximum reduction of switching hash. Use epoxy resin to glue the capacitors to the front panel so that they are not just supported by their leads. 13 and 14 can then be supported by gluing them to the ends of the capacitors. The 3-core mains cord is clamped to the rear panel using a cord grip grommet but first strip back the outer insulation by about 60mm. Connect the active (brown) and neutral (blue) leads directly to the terminal block as shown in Fig.6, and solder the earth lead (green/ yellow) to the adjacent solder lug. The remaining wiring can then be run to the mains fuse, power switch, front and rear panel earth lugs, and to the GND terminal using 240V AC cable. Sleeve the mains switch with plastic heatshrink tubing to avoid the possibility of accidental shock. In fact, it's a good idea to sleeve the switch terminals first, and then sleeve the whole of the switch body up to the mounting nut. Note that we have specified a mains switch with a plastic body and actuator. Don't use a miniature toggle switch here. If you substitute a switch with a metal body, make sure that it is correctly earthed to the front panel. Switching on Before switching on, wind the Voltage Adjust pot (VR4) to minimum setting and set VRl, VR2 and VR3 to mid-range. With that done, you can apply power and check the supply rails. Check that the unregulated voltages to the emitters of Ql and Q2 are at ± 60V, and check that the ± 30V, ± 15V, + 12V and + 2.5V rails are all present. Now connect your digital multi- meter across the positive and negative output terminals and check that the positive and negative output voltages increase as the Voltage Adjust pot is wound up. Check also that the two rails track each other within ± 1 % (by measuring between each rail and OV). If you want the tracking closer than that, you will need to pad one of the 22k!l 1 % resistors. If everything is OK, wind the Voltage Adjust pot back to its minimum setting, connect your DMM between the positive and OV output terminals, and adjust VRl so that the DMM reads OV. This done, wind the Voltage Adjust pot up to maximum setting and adjust VR2· for a reading of 50V. Finally, adjust VR3 so that the front panel meter reads 50V. All that remains is to secure the lid of the case and your new power supply is ready for work. ~ APRIL 1990 57