VFD PSU revisited. By Lee Davison.

I still have quite a few vacuum fluorescent displays that have been rescued from old video recorders, radios, clocks and the like. They aren't difficult to use but do need an AC filament supply and an anode supply of anything from twelve to about thirty volts.

This circuit is a copy of the circuit from an old calculator and provides the AC filament supply and high voltage anode supply for a vacuum fluorescent display from a single DC supply.

The advantage of this circuit is that the anode voltage is at the logic supply voltage so switching the anodes and grids does not require either high voltage open collector outputs or level shifters. A single small PNP transistor and two resistors is enough to interface each drive.

Clicking on each part of the picture will take you to a close up of that part.

Q2 and T1 form a blocking oscillator with the feedback current via C1 being amplified by the emitter follower stage Q1. R1 provides more feedback as well as the positive bias to the base of the emitter follower.

The output from the 160 turn secondary is rectified by D1 and smothed by C2 producing the negative bias for the VFD cathode (filament). This output is then fed back via the LM431 adjustable shunt regulator to the emitter follower's base with the result that the more negative this output goes the less time per cycle that Q2 conducts for.

The centre tapped 24 turn winding produces two to three volts AC to drive the VFD filaments and the centre tap allows symetrical biasing to the negative cathode voltage.

The anode voltage is the system supply voltage and so can be easily switched by a single PNP transistor.

All the components are what I had to hand at the time and none of them are particularly critical.

Transistors

The 2N2222 can be replaced by any high gain small signal NPN transistor such as BC108, BC547, 2N3904 etc and the BD437 can probably be replaced by any near equivalent. In this circuit the BD437 ran cool so could probably be replaced by a smaller transistor.

Transformer

This was wound, using 0.125mm enameled wire, on a small ferrite cored bobbin that was the feedback transformer in a PC power supply. After each winding was complete a layer of tape was laid over it. The start of each winding was marked on the bobbin with each winding being wound in the same direction. The dots in the circuit diagram denote the start of each winding.

The turns count and ratio could be varied if, for example, you need more anode volts without increasing the filament voltage then more turns can be added to the 160 turn anode voltage winding.

Diode

A random shottky diode was used to rectify the cathode voltage and any diode that can carry a few 10s of mA and withstand a reverse voltage of 50V or more, such as a 1N4148, should work here.

Regulator

The LM431 is used as a adjustable zener, the voltage being determined by Rx, and is used to set the voltage between the cathode and the anodes up to a maximum of about 36V. This device can be replaced by any equivalent such as TL431.

For cathode to anode volatges above 36V the LM431 and it's two resistors should be replaced with a zener diode of the required voltage.

Rx

Rx is used to set the cathode voltage. With the voltage control disconnected the unloaded output is greater than eighty volts. With the voltage control connected the approximate cathode to anode voltage is given by ..

		V = 3 + 2.5 * ((Rx / R2) + 1)

.. which for Rx = 220K gives a cathode to anode voltage of about 30 volts.

Capacitors

As this circuit oscillates at a far lower frequency than the prototype the rectfier resevoir capacitor is made up of a single 1uF 63V electrolytic. This value is not critical.

The 1000pF oscillator feedback capacitor was chosen to give the best running. The circuit will still oscillate even if this capacitor is omitted, though regulation suffers. The circuit seems to work fine over a few decades of range on this capacitor.

Resistors

The base bias resistor, R1, should be a fifth, or less, of the value of the LM431 reference voltage resistor R2. If it is much more than this regulation suffers as the current through Rx and R2 alone is enough to affect the bias point.

As the VFD anode supply is at the same potential as the logic supply the drives are just a single PNP transistor, in this case a BC557 but any equivalent should work, and two resistors. A third resistor is used to pull down the pin when the drive is off to prevent ghosting.

The anode/grid pulldown resistors, Rc, can be in SIL resistor packs and the value is not critical. 100K was only used because packs of this value were available, higher values could also be used.

Note that Rc pulls the anode or grid down to the VFD negative supply and not to zero volts.

The base pullup resistors, Rb, can also be in SIL resistor packs and the value is not critical. 100K was only used because resistors of this value were available, other values could also be used. This resistor ensures that the transistor is turned fully off when driven by TTL or open collector outputs. Rs, the series drive resistor, should be selected so that with a low output enough current is passed to turn on the transistor.

Last page update: 24th July, 2009.

e-mail me