Building a Valve (vacuum Tube) Prototype Board Out From Recycled Materials.

For a long time, I wanted to experiment and build amplifiers and other audio-related devices based on electron tubes.

To achieve this goal, I needed a suitable tool for testing and experimentation.

That's why I made this simple device, the materials and components I used are mostly recycled or inexpensive.

As I said in the introduction, most of the materials used in this project are recycled.

-I just bought a cheap kit box with M3 screws, washers and nuts and a NOS set of sockets for miniature, noval and octal tubes.

Obviously, the electrical tape, shrink tubing, rivets, plastic ties, double-sided tape, epoxy, paint, etc. are new. :-)))

-A side cover from a pc case salvaged from a dumpster was used to build a chassis.

-A bunch of screw terminal strips and colored wires with crimp terminals taken from a scrapped cnc machine made the circuitry.

-A power cord and an IEC socket from an old pc power supply.

-A chassis fuse holder with a 2A, 5mm fuse from my parts stock.

-A DPST toggle switch wit neon lamp also from my parts stock.

-Assorted sef-tapping screws from one from my infinite "Tachos de los Tornillos"

When working with any device connected to the power grid, always follow safety procedures to avoid any risk of electric shock and electrocution.

Electron tubes usually operate with high voltages, both AC and DC (which is even more dangerous), even when isolated from the mains by a transformer. So, additional safety procedures must be taken.

In addition, the filter capacitors in the power supply and other stages can hold elevated voltages after the device is turned of, which can cause unpleasant surprises. Therefore, it is always advisable to check that the capacitors are discharged before any intervention in the circuit.

Many types of capacitors, including electrolytic capacitors, have a property called "dielectric absorption" that causes their charge voltage to spontaneously rise after they have been discharged. This can lead to unpleasant surprises too.

Finally, valves operate at high temperature due to their nature. Glass envelopes often reach temperatures above 200°C, so be careful not to burn yourself with your fingertips.

I selected this case cover from my hoard of materials for its dimensions and shape. The sheet metal is very thin, just 0.7mm, which makes it very easy to work with. The perforated area facilitates centering and fixing of the components.

That perforated area has plenty of room to fit two miniature sockets, two novals and two octals, plus the terminal blocks to give access tho the socket pins, plus few extra blocks for components connections, line power and ground.

The making process was very straightforward: The cutting, drilling and bending areas were traced on painter's tape.

The corners and the rectangular holes were cut using scissors and a small grinder. The large holes for the tube sockets were drilled with hole saws.

The smooth areas were folded down to create a box-shaped chassis.

A couple of rivets at each corner give rigidity to the whole.

A layer of epoxy covered the corner seems and other imperfections.

Once the epoxy cured, it was sanded smooth and a quick spray of paint was applied to give it a nice finish.

The resulting chassis measures 320mm long x 290mm wide x 65mm high.

The available terminal blocks have various numbers of terminals. Most of them have four terminals, some have eight, one six and one ten.

I decided to assign the 8 terminals blocks to the 8 and 7 pins sockets.

There is only one ten terminals block, but two of the nine pin tube sockets, so I made a ten terminals block by joining a six and a four terminals blocks together to make a ten terminals block from them.

Once the chassis was built, the components were fitted into place.

On the front plate, the power switch was inserted. On the back plate, the power socket and the fuse.

At the top, the tube sockets were secured with M3 screws and nuts. The terminal strips were fixed mainly with self-tapping fasteners, except at those ends where the end terminals were connected to ground.

The colored cables were screwed to their corresponding arranged by the numerical color code (the same of the resistors).

-Brown = 1

-Red = 2

-Orange = 3

-Yellow = 4

-Green = 5

-Blue = 6

-Violet = 7

-Grey = 8

-White = 9

The wires passage through the chassis was protected using shrink tubing and cable ties, these wires were soldered to the corresponding socket leads.

Finally, the power circuit and ground cables were installed.

At this point, the project is completed, now it's time to do some tests.

Well, no building project is complete if it is not used for what is was intended for...

The inaugural test was the assembly of a very basic class A amplifier. I gathered some components from my parts bins. A power transformer from an old record changer, an output transformer from some table radio, a very used 6BQ6/EL84 valve and some resistors and capacitors.

The distribution of the blocks on the chassis permits a quick assembly and modification of the arrangement of the components.

The main goal of this project is to be an instrument for developing and testing devices with valves, but it can also be used as a learning tool.

So: What is a valve? How does it work?

In simple words, a valve/vacuum tube/electron tube is an electronic component able of controlling an electric current that flows through it from a voltage applied on a control electrode. I don't want to boring you with math, so I'll try to explain the subject in the simplest way I can.

It consists of an evacuated envelope (usually made of glass) which contains several metal electrodes.

The working principle is based on the "thermionic emission", which is the liberation of charged particles (electrons in this case) from a hot electrode. As a metal is heated (for example, by an electric current passing through it), its electrons acquire kinetic energy and tend to escape from it. This hot metal electrode takes the form of a filament, a thin metal wire made of tungsten.

This filament, enclosed in glass envelope with a high vacuum inside, makes a light bulb in its simplest form.

In this condition, nothing can be done about the cloud of electrons flying around the filament. If a second electrode is added to the device, a metal plate ("THE PLATE:"), it can attract these electrons if a positive potential is applied to it with respect to the filament, generating an electric current.

This device has two electrodes: The filament, also called cathode and the plate, also called anode. This device is called a DIODE, its rectifying function is analogous to the silicon diode. However, its characteristics are very different from the solid state counterpart. The forward voltage drop is increasingly larger as the current increase.

I made a simple demonstration assembly that shows a diode in action as seen in the photos. The magnitudes and waveform can be observed on the meters and oscilloscope.

The only way to control the current is by reversing the applied voltage or by modifying the load.

To control the electrons flow at will, the addition of a third electrode is needed.

This control electrode is usually in the form of a grid ("THE GRID") made of thin wire, located between the cathode and the plate. A valve built in this configuration is called a TRIODE.

By applying to the grid a negative relative to the cathode, the plate current can be modified.The more negative the grid voltage, the lower the plate current.

A varying voltage (input signal) applied to the grid has a varying plate current (output signal) in response. There is a ratio between the variation of the input and output signal, this ratio is called "Mutual conductance" or "Transconductance" (gm): [gm=Δiout/Δvin] expressed in microsiemens (μS -NOT micro seconds-). If you look at old datasheets, the units are expressed with different symbols: μS are μmho (Ohm backwards!), pF are μμF, KHZ are K.C. (weird huh?)...

This is the key parameter that permits the valve to amplify an electrical signal.

I assembled a simple triode set up to demonstrate it in action as an amplifier.

Since the first valves were available in the early 1900s, inventors, engineers and scientists improved the component's performance over decades of development: Filaments were coated with substances that enhance electron emission. Indirectly heating cathode was created by enclosing the filament within a metal tube. Additional grilles (screen, suppressor, etc) were added to cope with some limitations and improve performance.

All these years of evolution led to the creation of a wide variety of models and applications, giving rise to radio, television, electronic computers, radar, hifi audio, microwave ovens, etc, etc.

Valves have a finite life span by nature, the cathode emission decays over time to a limit where they are no longer operational. Also, because valves are assembled using multiple small parts with intensive handwork, they can often fail due to mechanical causes such as insulation leakage, short-circuits from loose particles or vacuum loos due to cracks in the envelope seal.

That's why it's necessary to test valves, for example to repair a malfunctioning device.

Today, you can still purchase valves on the market, both brand new (yes, there are still active electron tubes factories around the world), new-old-stock or used. You may need to test these valves to check they are of the quality claimed by the vendor or to create matching pairs where required (e.g., push-pull amplifiers).

There are several types of valve testers, both vintage, new or DIY.

Almost all of them can check gas intrusion, insulation leakage, noise, emission, transconductance, etc.

I don't have a valve tester (not yet), so this is a good opportunity to improvise some basic tests on the prototype board:

Emission test is one of the simplest ways to check a valve's condition. It consists in apply a preset dc voltage between the cathode and the plate with the grid, screen and other auxiliary electrodes connected to the plate, and measuring the current using a milliammeter.

Such instrument has its scale plate marked with a green and a red zone in a Pass-Fail fashion.

If the plate current is within the green zone, the valve is OK. Otherwise, it is weak, worn out or non-functional.

The first, second and third pictures show typical emissions testers and its milliammeter with its color scale.

I put together a simple set-up on the breadboard to perform an emission test based on a simplified diagram from my RCA data manual: An adjustable power supply using a variac, isolation transformer, rectifier and filter; my restored "San José" power supply provides the 6.3v for the filament; the plate voltage and current are measured with two DMMs.

Old commercial testers don't even bother to rectify the applied voltage since the valve act as a diode in this configuration, but I'm using the proposed RCA setup.

The valves of choice are three 6AQ5 in different conditions: One NOS, one very worn and a troubled one.

Applying 100Vdc, the NOS draws 21.44mA; the worn down, just 3.06mA and the defective unit runs almost 100mA under 20V!

Note that the NOS valve plate current should be around 30mA according to the curve shown in the photos. However, emission is within the acceptable range (barely). That's why it's necessary to check the valves when you bough them, even if they claim to be in mint condition.

Having said that, this a perfectly useful component.

Emissions test is a quick and easy testing method, but it does not accurately show the performance of a valve. After all, this is a static test that is not based on any key parameters. A better testing method is required.

The other widely used test method is the transconductance test, which is based on a well-defined parameter under more realistic dynamic conditions.

The first photo shows a classic transconductance tester (this unit is mil-spec, highly valued today among people involved with valves).

This time, I have built a transconductance test rig based on the figure taken from the RCA data manual.

The set up consists on wiring the valve under test applying preset voltages to each electrode according to the recommended working conditions in the datasheet. Then, by applying an AC signal to the grid, the corresponding AC current on the plate can be measured (this AC current is superimposed on the static DC plate current). The same variac-based power supply was used as in the previous test, the "San José" power supply provides the grid negative bias and another power supply feeds the filament with 6.3V DC. The DMMs display DC plate voltage and AC RMS plate current. The scope injects a 1V RMS signal into the grid and shows its amplitude and the grid bias.

If the grid signal is 1VRMS, the plate RMS AC current is equivalent to the transconductance according to the formula gm=Δiout/Δvin. Therefore, the AC mA reading multiplied by 1000 is the transconductance.

The test conditions are shown in the table taken from a typical datasheet. The valves chosen are two of the three 6AQ5 from the previous test (the failed unit was excluded).

The last two photos shows the test readings of the NOS 6AQ5 and the worn one. The new valve has a transconductance that nails on specs, 4111 micromhos/microsiemens, while the oldie reads just 1700 microsiemens (It will work on an amplifier, but the sound will be very thin).

This test is very useful, not only for checking a valve's condition. Valves, like any electronic component, have dispersion and tolerances in their parameters. Datasheets only show typical characteristics, but they may be slightly higher or lower than these figures from one unit to another. To build push-pull amplifiers, for example, its necessary to match pairs with similar parameters to each other to get the best performance.

A better testing method is to plot curves using a curve tracer. The transconductance test is good enough for most uses, but it focuses on one specific working point. A curve tracer can show the dynamic behavior of a component under various conditions at the same time.

Off course, I don't have a tracer (not yet!), but its possible to do some basic testing using my equipment.

Using the isolation transformer connected to the variac, a 1K resistor as a load, a 1 ohm resistor as shunt and the scope in XY mode it is very easy to plot the curve of a diode and compare it with the datasheet.

The valve under test is a high run hours 6AX4GTB (worn, but still functional). My "San José" power supply powers the amps-thirsty filament, the X axis show the forward voltage drop and the Y axis the current (50mV/div represents 50mA/div). Comparing the oscilloscope curve to the datasheet, it's noticeable how worn this valve is. The forward voltage drop is 40V @ 125mA, while it should run 350mA with the same drop, according to the datasheet. However, this valve may still be useful for some limited power uses...

Over the course of more than 40 years of work, I have collected many old books related to vacuum electronics in my library. Most of them are freely available on the internet, but these are my treasures.

The Information and tests I shared in this publication is not intended to be in-deep or highly accurate, but I found this experience very satisfying. I hope it was useful, inspiring, or at least entertaining to someone else.

If so, my mission is mostly accomplished.