High Voltage Power Supply for Vacuum Tube Projects

Before 1960 or so, opto-electronic instruments and devices employed vacuum tubes. Here we describe a regulated variable high voltage power supply that finds general use in the restoration of vintage laboratory equipment as well as at the experimenter’s workbench. The supply is based on earlier work by Leigh Bassett1 and Cam Hartford2. A goal was to make use of the previous work and to create a printed circuit board (PCB) that could be used by many builders for unique applications. To demonstrate performance and to encourage readers to build something useful, a battery of tests were performed on the PCB and I breadboarded a two-channel preamp using two 6SN7 vacuum tubes. As we’ll see at the end of this note, the bare PCB is available from the Antique Wireless Association AWA) (

Consider the schematic shown in Fig. 1. The circuit utilizes the LR8N high voltage regulator and the TIP50 pass transistor in the configuration of Bassett and Hartford. AC input to E1/E2 is supplied by a suitable isolation transformer. Alternatively, for use in existing DC circuits, DC voltage can be applied at E8. Input voltage should not exceed 265 VAC or 375 VDC.

Figure 1. Schematic of the regulated variable high voltage power supply PCB.

Variable resistor, VR1, is used to select the output DC voltage at E6 and can be mounted on the board or remotely on a chassis using E4 and E5. If a fixed output voltage is desired, R4 and VR1 can be omitted and R2 can be used in conjunction with R1.

The TIP50 resides on a suitable heat sink, which can be mounted on the PCB or remotely. When mounted properly on the PCB, the heat sink is at B- potential. There is also a provision to “run barefoot” for very low current applications using connection point E9, which is the low-current output of the regulator. B- is at E7, and is isolated from chassis ground to allow for a floating power supply. The circuit also allows the use of a center-tapped power transformer using E1/E2 and E3 for the center tap. In this configuration, D1 and D2 are omitted.

Figure 2. Regulated variable high voltage power supply PCB, with components.

While the circuit in Fig. 1 is quite general, a representative case with the following requirements and components is considered:

Voltage, Output: 90-330V DC, variable

Voltage, Input: 120V AC

C1: 100μF 450V electrolytic capacitor

C2: 1μF 450V electrolytic capacitor

C3: 22uF 450V electrolytic capacitor

D1-D5: 1N4007 diode

HS1: Depends on user requirements, Ohmite FA-T220-38E fits on PCB

PCB: AWA Rev 6, 8/2020

Q1: TIP50 NPN power transistor

Q1 Mounting Kit: AAVID 4880MG

R1: 2k, 1/4-watt resistor, 1%

R2: Used for fixed output voltage only; not used with VR1

R3: 150k, 2-watt, 5%

R4: 130k, 1/4-watt resistor, 5%

T1: Power transformer, Triad VPS230-110

U1: LR8N3-G 3-terminal adjustable regulator

VR1: 500k 1/2-watt, Bourns 3296Y-1-504LF

It is important to note that the TIP50 transistor will get hot at relatively low output voltage and high output current, and heat sinking is required. The power dissipation in the TIP50 is given by Vce*Ic. In our example, the unregulated DC voltage will be about 365V DC depending on the input AC voltage. For illustrative purposes, at an output voltage of 265V DC and ambient temperature of 25°C, the maximum power dissipation, Pd-max, for the TIP50 is given by


where Rjc is 3.125 °C/W and HSrth is about 10.4 °C/W for the Ohmite FA-T220-38E (measured to be much higher than the specification of 3.8 °C/W). Using these values,

Pd-max=9.24 watts.

If Vce=365V-265V=100V and Pd-max=9.24 watts, then the maximum load current, Ic-max=92.4ma. This is just an illustrative calculation and the actual allowable values depend on the specific heat sink being used, operational duty cycle, safety margin, and unobstructed air flow. Some thought must be given to power dissipation in cases where the PCB, with on-board TIP50, is mounted in a location where cooling is compromised.

The minimum input-output voltage differential across the LR8N should be greater than about 20v. The following table gives an approximation of output voltage as a function of R2, given R1=2k, using 1% resistors (in this explanation, R4 and VR1 are not used.)

Output VoltageR2

Figure 3 shows the LR8N PCB energizing a two-channel vacuum tube line stage. Barbour and Kittleson3 gave a schematic and a detailed description, which allowed the build to be completed in an afternoon using a repurposed chassis. It operates at B+=320V DC, gain is 15 and essentially flat from <10Hz to >100KHz and it’s plenty to drive most any tube or solid state amplifier. The PCB really streamlined the process and gave regulated B+ as a nice benefit.

Figure 3. LR8N PCB, powering a stereo 6SN7 preamplifier.

The LR8N PCB was created to be used in a wide variety of user-specific applications. I’ve used the circuit widely in buffers, pulse generators, signal generators, oscillators, amplifiers, HV triggers/shutters, differential amplifiers and active filters. It comes in handy when rebuilding vintage electro-optic instruments and devices.

Speaking of the older equipment that shaped the future, please do not scrap it. Too much of our optics heritage has been landfilled already. Please send me an email about it (dci at perluma dot com) and consider donating it for study and preservation.

As mentioned earlier, a professional quality, silk-screened circuit board is available (High Voltage Power Supply PCB) and it ships with a schematic and representative parts list. Their website will soon include test results using several different power transformers with this PCB.

There is a long-standing commitment to supporting builders of vacuum tube audio amplifiers, HiFi equipment as well as restorers of vintage test equipment and amateur radio gear. I’m trying to highlight the need for a similar commitment to help preserve our optics heritage.


1. Leigh Bassett, “A solid-state filter choke or field coil replacement,” edited by Ken Owens, AWA Old Timers Bulletin, 45-2-14, 2004.

2. Cam Hartford, “A Power Supply for April’s Two-Tube Transmitter,” CQ Magazine Digital Edition, June 2011.

3. Eric Barbour and Charles Kittleson, “VTV Octal Line Stage Project,” Vacuum Tube Valley, Issue 11, p. 10, Spring 1999.


Their Heads in the Clouds

“The base of the lowest cloud, or ceiling, can be considered an important meteorological observation, although emphasis on its measurement comes primarily from aircraft navigational needs. In the middle 1920’s aviation had increased to an extent which required the Weather Bureau to measure ceiling with small balloons and ceiling light projectors. In the absence of these devices, the cloud height was estimated by eye. The importance of ceiling for aircraft operations, in particular the importance of having it recorded in cases of litigation in connection with aircraft accidents, led to the development of the practical instrument now used—the automatic ceiling projector. Light modulated (to be detectable in daytime) is projected vertically and scanned automatically from some distance away. The angle of the scanner when it is pointing at the spot on the base of the cloud is recorded continuously. Thus ceiling is measured automatically and could be telemetered.”

Automatic ceiling projector.
Ceilometer computer. [1]

In this configuration,[2] the detector is scanned up and down the vertical light beam and the scan angle of maximum detected illumination is recorded. Given the displacement between the light transmitter and receiver and the measured angle, ceiling height can be calculated. Danish scientist, inventor, and meteorologist, Poul la Cour, is the likely inventor of the modern ceilometer in 1871.

By the mid 1960’s, the use of pulsed laser radar (LIDAR / LADAR) for atmospheric measurements was established firmly.[3]

Cloud height measurement technology, 1966.

Today, it’s done a bit differently, but most techniques still use laser radar in one form or another. Over time, we’ve developed better materials, tools, and understanding. We learned from the experience of the engineers that came before us.


1. Paul Meissner, “Design and Operation of the Ceilometer Computer,” National Bureau of Standards, 64 (1960).

2. C.A. Douglas and R.L. Booker, “Visual Range: Concepts, Instrumental Determination, and Aviation Applications,” National Bureau of Standards Monograph 159 (1977).

3. J.E. Masterson, J.L. Karney, W.E. Hoehne, “The laser as an operational meteorological tool,” Bulletin American Meteorological Society, Vol. 47, No. 9, SEP 1966.


The Barons Rayleigh and Corning UV Glass

The last post discussed early UV-VIS spectrophotometers. Those systems required phototubes with envelopes that pass UV. Corning’s ( Corex ultraviolet transmitting glass fit the bill and found application.

Apparently, there is another feature of Corex, although the author does not state exactly which material he studied:

Double Refracting Structure of Corex Glass

Nature, Volume 126, p. 845 (29 November 1930)

Some years ago I found that silica glass showed a doubly refracting structure (Proc. Roy. Soc., A, vol. 98, p. 284; 1920. Also Proc. Optical Convention, Part I, p. 41; 1926). This structure is quite distinct from any due to bad annealing, and seems kindred to the ‘liquid crystals’ of Lehmann. Nothing of the kind could be found in the ordinary glasses consisting of silica with metallic oxides.

I now find that the ultra-violet transmitting Corex glass of the Corning Co. shows a similar structure. This glass is said to consist in the main of calcium phosphate, though I have not seen an analysis. The subject evidently requires detailed examination, which I hope to make as opportunity allows.

(Robert John Strutt, 4th Baron) Rayleigh
Terling Place, Chelmsford,
Nov. 14.

This Lord Rayleigh is the eldest son of TheLord Rayleigh (John William Strutt, 3rd Baron Rayleigh). The unit, Rayleigh, is named in honor of #4, commonly known as Robin.

#3 established a laboratory at Terling Place, the family’s residence. #4 resigned his chair in physics at Imperial College after #3 died in 1919 and continued his research in the laboratory at Terling Place. Gentleman (or independent or citizen) scientists understand the benefits of WFH (working from home).

Unlike electronics, vintage optical datasheets are harder to find online. Here are some comments uncovered from a confidential US document that looked at materials and methods from 1920-1950 (G.L. Harvey, “A Survey of Ultraviolet Communication Systems,” NRL Report 6037, March 13, 1964):

“Corex D is useful for bulbs on sunlamps as it transmits the erythema or sun tanning wavelengths which are most effective at about 2967A while cutting out the less desirable germicidal mercury line at 2537A. The most effective wavelength for germicidal effectiveness is 2600A, dropping to very little bactericidal action at 3200A.”

Corex D appears to be Corning 970. Corex A (red-purple glass) appears to be Corning 986 or 986A (think black-light lamp). Clear Corex A appears to be Corning 980. But, this could be wrong. When material properties or some datasheets are found, they will be posted. It would be nice to see confirmation of #4 Lord Rayleigh’s observations of double refraction in Corex.

Corning Print Ad, 1938

For more information on the optical work of #4, see his son Charles R. Strutt’s article, “The Optics Papers of Robert John Strutt, Fourth Baron Rayleigh,” Applied Optics Vol. 3, No. 10, p. 1116-1119 (October 1964). Charles is the father of #6, John Gerald Strutt, The Lord Rayleigh.