Laboratory Power Supply
A complete description of my adjustable laboratory power supply, including the circuit, component selection, digital metering and mechanical construction.
Introduction
After a long break, I decided to build a new device. Because my laboratory lacked a power supply with adjustable current limiting, I chose to build one myself. After searching online, I found a design that matched my requirements very well: adjustable current from approximately 2 mA to 4 A and adjustable output voltage from 0 to 30 V.
The supply works very well up to nearly 30 V; the actual maximum is slightly lower because of voltage drops across the rectifier diodes. At 4 A, the output voltage drop is at most about 1 V. I had heard many negative comments about this circuit, including reports of failed components and oscillation. However, when the circuit is assembled carefully and component values remain within tolerance, it should operate reliably. I recommend increasing the values of the suppression capacitors that prevent the operational amplifiers from oscillating.
Power Stage and Components
The original circuit uses a 2N3055 as the final power transistor. I recommend using a KD503 or KD501 instead, because some inexpensive 2N3055 devices sold as branded parts may be poor-quality copies and can handle only a fraction of the specified current. The 2N2219 driver transistor can also be replaced with a KF506.
All components are commonly available, and the component cost should remain below approximately CZK 250, excluding the potentiometers and power transistor. Suitable KD-series power transistors and large filter capacitors can often be recovered from older televisions. The original supply was designed for 3 A, but it can provide 4 A after reducing the value of R7. I also used two output transistors connected in parallel.
Circuit, PCB and Wiring
Bill of Materials
| Resistors | R1 2.2 kΩ/1 W; R2 82 Ω; R3 220 Ω; R4 4.7 kΩ; R5, R6, R13, R20, R21 10 kΩ; R7 0.47 Ω/5 W; R8, R11 27 kΩ; R9, R19 2.2 kΩ; R10 270 kΩ; R12, R18 56 kΩ; R14 1.5 kΩ; R15, R16 1 kΩ; R17 33 Ω; R22 3.9 kΩ. |
|---|---|
| Controls | RV1 100 kΩ trimmer; P1, P2 10 kΩ linear potentiometers. |
| Capacitors | C1 3300 µF/50 V electrolytic (a larger value is preferable); C2, C3 47 µF/50 V; C4 100 nF polyester; C5 200 nF polyester; C6 100 pF ceramic; C7 10 µF/50 V; C8 330 pF ceramic; C9 100 pF ceramic. |
| Diodes | D1–D4 1N5402; D5, D6, D9, D10 1N4148; D7, D8 5.6 V/2 W Zener; D11 1N4001; D12 red diffused LED. |
| Semiconductors | Q1 BC547/BC548; Q2 2N2219 or KF506; Q3 BC557/BC327; Q4 KD506/KD501 or 2N3055; U1–U3 TL081. |
My Construction
As usual, I refined and miniaturised the design as much as possible. I replaced ceramic capacitors with higher-quality, more stable film capacitors, used precision IC sockets, good-quality wiring, precision potentiometers and a toroidal transformer.
During development, I found a roughly twenty-year-old aluminium enclosure at an online auction. It was better made and more robust than many modern low-cost enclosures. For the voltage and current indicators, I used a proven digital meter design. I programmed the Atmel microcontroller with a USBasp programmer and AVR eXtreme Burner.
The meter works very well, but one important detail is not widely documented: a 10 nF capacitor should be soldered directly between pins 21 and 22 of the Atmel microcontroller. Without it, the reference voltage may oscillate and the reading becomes unstable. To enter the settings menu, hold the button while switching on the power.
The complete project cost approximately CZK 2,000. The construction photographs are shown below.
