PS5V
Nostalgia for analog ammeters and precision potentiometers.
0.1 ~ 5.0 V, 1 A, Precision Power Supply

Designed by Takayuki HOSODA
Rev.1.95 (Oct. 19, 2022)

This small laboratory power supply was basically designed and built in 1990 out of nostalgia for the use of analog ammeters and precision potentiometers, which were disappearing from laboratory power supplies at that time.
It was also designed to provide a stable supply of low-voltages of 100 mV to 5 V with high resolution due to the trend toward lower voltages in electronic circuits.

This design shown below is basically a maintenance time improvement of the PS5V Rev.1.4, and is NOT intended to be used for new production.

Japanese edition is here.

Schematics

The main modifications include:
The voltage reference U1 was replaced with the state-of-the-art LT6657A (1.5 ppm/°C), U2 with the Precision Low-Noise CMOS Operational amplifier TLC2202IP (0.5 μV/°C, 18 nV/√Hz@10Hz, 110dB PSRR).
The resistor for the voltage divider was replaced with a pair of ultra-high precision resistors PTF5610K000AZEB (0.05 %, 5 ppm/°C).
The output capacitors were changed to high capacitance MLCCs GRM32EB31C476KE15 (27 μF @ V_DC = 5 V, 2.5 mΩ ESR)
If available, it might have been better to use matched pair resistors instead of ultra-high precision resistors.
Also, it might have been better to add an LDO local regulator of 7.5 V for U2.

Fig.1 : PS5V (Rev.1.95) Schematic diagram

Download

Schematics : PS5V1R95-sch.pdf

Voltage drift and long term stability

Fig.2a : Initial voltage drift just after replacing U1 to LT6657A, at bench top (V_out = 5 V, 22 °C < T_a < 26 °C)

40 μV/div (8 ppm/div)

Although there is some initial voltage variation after soldering, the LT6657A is significantly more stable than the TL431A.
However, it is very sensitive to deflection stress, so considerable care must be taken in mounting and wiring.

Fig.2b : Voltage drift after replacing U2 to TLC2202, at bench top (V_out = 5 V, 22 °C < T_a < 26 °C)

40 μV/div (8 ppm/div)

The TLC2202IP is considerably quieter and more stable compared to the TLC272CP.

Fig.2c : Voltage drift after 15 days of aging at bench top (V_out = 5 V, 22 °C < T_a < 26 °C)

20 μV/div (4 ppm/div)

The small ripples are due to the effect of the temperature control of the air conditioner.

Fig.2d : Voltage drift characteristics for 244 hours on benchtop after 54 days (1296 hours) of aging (V_out = 5 V, 21 °C < T_a < 26 °C)

50 μV/div (10 ppm/div)

To tell the truth, with the replacement of key performance components, the main cause of temperature drift now appears to be the thermal EMF (electromotive force) of the precision potentiometer contacts.
Oh, precision potentiometers, forever!

• Low-noise Pre-regulator for LT6657A (LTspice simulation)