Philips PE4801/00 regulated power supply (1960)

Somewhere in 2006 I bought this power supply through an Internet second hand site. I expected a versatile lab power supply, but it was a little more limited than I had hoped. It is one of the first transistorised regulated power supplies by Philips. The PE4801 goes from 0 to 30V and can deliver 300 mA output current. It has no electronic current limiting but a fuse to safeguard against short-circuiting the output. On the front there is an output meter. It comes in a grey hammer finish metal cabinet like many Philips instruments of the day. The series transistor is mounted on a heat sink that stands on three isolated mounts. Most of the other electronic components are on a PCB on the side of the chassis. The neon voltage stabiliser tube is in a socket on the metal back plate.

On the NVHR site I found documentation, including the circuit diagram. The circuit is quite simple: a series transistor in a darlington configuration is driven by a one-transistor error amplifier, that tries to copy the input voltage from the voltage control. This input voltage is derived from a floating 85V reference voltage. The line-up of semiconductors and valves is:

• ASZ18 power transistor (the manual says OC36)
• Three OC77 transistors as amplifier and driver
• Two OA210 rectifiers as power rectifier
• a B250C75 selenium bridge rectifier for the reference voltage
• 85A2 neon stabiliser tube as voltage reference

The circuit looks a bit awkward but it is effective. It seems the designers were still getting used to transistors.

According to my notes, the PE4801 did work when I first tried it. But I had more robust power supplies for my hobby so it landed on a shelf where it stayed for years.

Early 2026 I decided to try this power supply again and get to know it better. I checked the condition of the electrolytic capacitors. These were ok. Then I switched it on, regulating the mains voltage. The 85A2 lit. But the output voltage was low and the voltage control did nothing. I measured the voltages in the circuit and was a bit puzzled. The error amplifier transistor didn't seem to work. I didn't have a spare OC77 (where do you find a Germanium transistor with a V_ceo of 60V?) but an ASY77 should do. It didn't make much difference.

Checking again, I found that a resistor in the feedback circuit seemed interrupted. One of the solder joints was interrupted, it was badly soldered. I looked and found several other suspicious looking solder joints. After correcting these “dry joints”, the PE4801 was working.

There was one more strange thing. While measuring, I noticed that the chassis was somehow connected to the unregulated −45V power supply. That didn't make sense. The collector of the series transistor and heatsink are also connected to the − power. The heatsink has isolating mounts with a reason. I saw that one long M3-screw that holds the PCB to the heatsink was a few mm too long and was touching the chassis, causing a short. A 66-year old production fault. This hadn't caused damage to the power supply itself but it must have caused problems in use. I cut 3 mm from the screw and the circuit was floating again relative to the chassis.

Then I soldered the original OC77 back in place and had a PE4801 that was working well. The meter showed that it responded well to the voltage control and the output voltage remained constant when loaded.