Introduction

A Sovtek EL84 power pentode tube I was working on my Vox AC30 clone that I built in high school, around 2009-2010, trying to figure out why I always had an oscillation issue when I turned the master volume up too far. Investigating the power amp side of the circuit revealed that I had incorrectly connected some of the grid-stopper resistors. I removed the tubes from the sockets so I could solder on the socket terminals without risking any damage to the tubes while they were in place, only to realize that the tubes were so stuck in their sockets that 2 of the 4 tubes cracked in their base. This caused a fatal problem for a vacuum tube: air ingress. So realizing that the tubes were ruined, I ordered a new matched quad, sockets, and got to work rewiring the power amp. While I was doing this, I thought that it would be interesting to take apart one of the tubes and document the process.

How a tube works

Schematic and pinout of an EL84 pentode It’s always helpful to have at least a little bit of background on what you’re looking at so here I will describe basically how a vacuum tube works. The schematic above shows the schematic symbol and pinout of an EL84 power pentode tube, this is the kind that is used in the output stage of a Vox AC30 and is relatively common in other smaller guitar amps and hi-fi amplifiers as well. It is, therefore, the tube I will be looking at in this article.

Basic Structure

The circle in the symbol represents the glass envelope which forms the exterior of the tube. Pins stick out of the glass envelope to provide an electrical connection with the tube’s internal components. The EL84, and many other tubes, has 9 pins. In the EL84 the pins are connected as shown in the picture above. The internal components and their purpose are as follows.

Filament (or Heater)

The filament, also known as the heater, is arguably the most important part of a vacuum tube. Without it, the tube could be installed into a circuit and nothing much would happen at all. In this symbol, the filament is located at the bottom and connected to pins 4 and 5. If there were no other elements inside the tube, the heater would be quite similar to an ordinary lightbulb, a thin wire inside of a glass tube which has had all of its air pumped out. It becomes hot and glows when current is passed through it. When the filament breaks and no longer gets hot, the tube is ruined. In a vacuum tube, the heat is what we are after, and the heat is applied to the cathode.

Cathode

The cathode is represented in this symbol as the “]” shaped element at just above the filament and is connected to pin 3. The cathode is an electrode that acts as a source of electrons in the tube. The term “cathode” literally means a negatively charged electrode. This makes sense considering that the cathode is usually connected to a low voltage, sometimes even directly to ground. When heated by the filament, electrons are liberated from the cathode in a process known as “thermionic emission.” These electrons will be manipulated by electric fields that are formed by the other elements within the tube and be the source of electrical current through the tube.

Plate (or Anode)

The plate, also known as the anode, is represented by the thick flat bar at the top of the symbol connected to pin 7. Although the term “plate” is more commonly used, it is helpful to remember that the plate is also known as the anode, meaning “positively charged electrode.” The plate of the output tubes in an amplifier are typically connected to the highest voltage present inside the amplifier, B+. The great difference in voltage between the cathode and the plate creates an electric field between the two electrodes. We know that electrons are negatively charged, and we also know that opposite charges attract. When electrons that have left the cathode are exposed to this electric field, they will experience a force causing them to move towards the positively charged plate. Even though the electron source is the cathode and they flow up to the plate, we usually refer to the current in the vacuum tube using traditional current flow, which is opposite in direction to the actual electron flow, so the current we are most interested in inside of a vacuum tube is referred to as “plate current” rather than “cathode current.”

Control Grid

The control grid is where things get interesting. In this symbol, it is represented by the dashed line located above the cathode and connected to pin 2. This is where the musical signal is usually applied to the vacuum tube. Early vacuum tubes only had one grid, so the control grid is usually referred to simply as the “grid” rather than the “control grid.” To understand its purpose we have to consider two conditions: when the grid is at a low voltage, and when the grid is at a high voltage. We know that electrons are negatively charged and are attracted to positive voltages. So consider what will happen when the grid is at a low (or negative) voltage. The electrons are leaving the cathode on their way to the plate, but they are interrupted by a negatively charged grid. Naturally, the electrons will be repelled by this since they have the same polarity charge. The effect will be that the plate current will decrease. In the opposite case, the grid is positively charged. Now the electrons on their way to the plate are attracted to the grid, accelerating them even faster towards the plate and increasing the plate current. Small changes in the grid voltage can cause very large swings in plate current. So now using the control grid, we have a way to take a small signal applied to the grid and turn it into a large signal at the plate, providing amplification.

Screen Grid

If we were describing a triode (such as a 12AX7 commonly found in the preamp of guitar amplifiers) we would be done. The three electrodes: the cathode, grid, and plate, give the term “triode,” literally meaning three electrodes. The EL84 we are looking at however, is a pentode, and it therefore has more internal electrodes. The first extra electrode is called the screen grid, and is located on the symbol just above the control grid and connected to pin 9. In preamp tubes, a triode is usually sufficient, however, in power tubes more amplification is desired. The addition of a screen grid helps to increase the gain of a tube by reducing the influence that the plate has on the cathode. It is usually connected to a constant positive voltage and can be thought of as a shield (or indeed, a screen) for the control grid to minimize the effect of the large changing voltage that is occurring at the plate that would otherwise cause a sort of negative feedback.

Suppressor Grid

Finally, we have the last element in the tube, the suppressor grid. It is located above the screen grid and in some pentodes, it would be connected to an external pin, however, in the EL84 it is connected directly to the cathode. The primary purpose of the suppressor grid is to minimize an effect known as “secondary emission.” Because of the high current in power tubes, the speed of electrons flying towards the plate is also very high. When these electrons slam into the plate with such a high energy it is possible to knock electrons out of the plate and send them backwards into the tube. Without a suppressor grid, these backwards-moving electrons would be attracted to the positively charged screen grid. This would have the effect of reducing the plate current, which would obviously be counterproductive to the point of a power amplification tube where getting high plate currents is the goal. The suppressor grid solves this problem by being attached to a low voltage, in our case directly to the cathode, and therefore repels the secondary electrons back into the plate.

The Symptom

Getter indicating a leak on the left tube After I pulled out the tubes from their sockets, I noticed that they looked like the picture above and immediately knew that at least one of the tubes was bad. How did I know this just by looking at them? Notice how the tube on the right has a nice evenly shiny surface at the top of the tube, while the tube on the left has a dull whitish look to it. This was the giveaway to me that air had gotten into the tube on the left.

State of the getter a few days later A few days later, I got around to the idea that I wanted to take a tube apart and write this article. The picture above is the state of the same tubes after a few day’s time. The white color had fully spread across the entire top of the tube. The color comes from a coating on the inside glass surface of the tube that is actually deposited after the tube is sealed. It is called the “getter” and its purpose is not to provide an easy indication that your vacuum tube no longer contains a vacuum, but rather to trap any impurities that are still present inside of the tube when the vacuum is pulled. The physics taking place within the vacuum tube depend on there being no particles to interfere with the electrons, so the getter is added to ensure the purity of the vacuum. How the coating is applied from inside the sealed tube will be easier to understand once we are inside the tube.

Teardown

The tube after breaking the glass envelope To begin the teardown of the tube, I began by wrapping it in a napkin and simply whacked it a few times with the handle of a screwdriver, it is only glass after all. Before smashing, the diameter of the tube is 22.0 mm. After smashing I could measure the glass thickness to see that it was 1.04 mm. Cleaning up the broken glass lets us finally move on to looking at the features inside of the tube in more detail.

Inside the tube The inside of the tube is in three sections as seen in the picture above. On the bottom is the base, containing the pins and interconnecting wires. The middle is the electrode structure where all of the amplification takes place. And finally, at the top is the getter.

Interconnections between the pins and the active elements inside the tube Looking closer at the base, we can see the interconnections between the outside pins and the bottom of the electrode assembly. The exposed portion of the pins are about 1 mm in diameter and 7 mm long. They penetrate through the glass in the base and are connected to smaller connecting wires that go up to the various electrodes. The electrodes themselves can be seen as the rods, tube, and sheet metal pieces sticking through the flat greyish-white object towards the top of the image. The smaller wires are 0.47 mm in diameter. Connections are made between wires and electrodes by spot welds. From this angle, it is rather difficult to see exactly where all the connections are going, so we will look from a different angle after the base is separated from the electrode assembly.

The Filament

Pulling away the base after cutting the interconnections After most of the connections are snipped, I pulled the base away from the center section which slowly revealed the filament. When assembled, the filament is actually inserted inside of the cathode. I left the filament connected to the base, so I could use the base to aid in pulling it out of the cathode without damaging the filament in some way.

The filament. Looks just like an old-fashioned lightbulb Pulling the rest of the way reveals the entire filament. One could easily imagine how it got its name from looking at it in this state. It is very similar to a traditional lightbulb filament, consisting of a thin wire, only about 0.11 mm in diameter, that meanders back and forth for extra length. The total length of the wire is about 200 mm. Most of the wire besides where the contact is made with the connecting wires to the base is coated in a kind of ceramic material that insulates the filament electrically from the cathode, while still allowing good thermal conductivity. With the coating, the thickness is 0.35 mm. The wire itself is likely made out of a high resistance metal such as nichrome, which is also used in toasters. When powered up, it gets red hot.

Interconnections

Looking down at the base to see the interconnections In this image, we are looking down from the point of view of the electrode assembly towards the pins. There are 9 pins that are numbered starting from the one in the top right corner of the image and moving around in a counterclockwise direction. These pin numbers correspond to the symbol shown at the beginning of this article. Notice how some of the pins are connected to the electrodes not with just wires, but rather with thin metal ribbons. These are about 0.07 mm thick and 0.6 mm wide. Pin 1 has no internal connection, but it does go up to a standoff that is riveted to the insulating mica plate to add some structure to the tube. From pin 2, the control grid, we can see two ribbons spot welded to the pin traveling to the middle area. These originally connected to not just one, but both rods of the control grid. The other grids are not treated as fairly, getting only one connection to their pin. Pin 3 may look like it has two connecting ribbons, but one (on the right) goes to the cathode, and the other (on the left) goes to the suppressor grid. Pin 4 and 5 have their wires bent down to the middle of the tube, where they are spot welded to the two sides of the filament wire. Pin 6 and 8 do not connect to anything, but their wires are kept long so they can act as stoppers and extra supports for the whole electrode assembly. Pin 7 connects directly to the plate. Finally pin 9 gets a ribbon to the screen grid. You can also see in this image the crack that caused the air to enter the tube. It starts on the left side by pin 5, travels to the middle, and up to pin 1.

Mica

Looking up at the lower mica This grayish-white plate is made of an insulating material called mica. Mica is a naturally occurring mineral that is very easy to form thin flakes. In vacuum tubes, mica is used for its strength and insulating properties to hold all of the electrodes together and position them in their proper place. The micas in this EL84 are in the shape of an octagon that is 18.8 mm across the flats and 0.4 mm thick. The points of the octagon are sized so they almost perfectly fit into the glass tube which provides excellent stability. In the image above, we are looking up from the point of view of the bottom of the tube. This is where all of the connections are made between the pins and the electrodes. At the top, we see the tab where the connection to the plate is made. On the left, right, and bottom sides there are similar tabs that are folded over the mica that serve to firmly hold the plate to the mica. At the very center, we see the cathode. From this angle it’s clear to see how the filament can fit into the center of the cathode. We can also see a small piece of the the ribbon that’s used to connect the cathode to the pin. Immediately to the left and right of the cathode are the rods for the control grid. Further left and right still are the rods for the screen grid. And finally, the rods farthest out from the cathode are the suppressor grid. The remaining object in the lower right area is that standoff that is connected to pin 1 as described in the previous section for support of the mica.

The tab on the plate folded up on the right side To get further into the interior of the tube, I have to bend the tabs holding the mica to the plate. In the photo above I have bent the tab on the right up. This can be contrasted with the tab on the left and in the front which are still folded over. This angle also shows the ends of all the electrodes sticking out through the mica. Once each of the tabs has been bent up, the lower mica can be removed.

Inside

Looking into the tube with the lower mica removed Now we can finally see into the heart of the tube. The grids are constructed with two rods with a wire wrapped in an oval shape around them. The arrangement closely resembles the symbol for the tube, only wrapped around in a full circle. I can also now get some measurements of the plate. It is made from two pieces of stamped sheet metal that are spot welded together to form a sort of oval-shaped tube. The thickness of the sheet metal is about 0.23 mm. From the left to the right flat sides is 15.7 mm, and the height of the oval is 11 mm. From the bottom mica to the top mica, or the total height of the plate, is 28.9 mm.

The Getter

The getter from the top This image shows the getter viewed from the top. It is a sort of round tray with a diameter of 9.7 mm. The tray is filled with a reactive material like barium. When the tube is assembled and the air is vacuumed out, an induction heating coil is used to heat up this tray which causes the getter to become so hot that it evaporates and travels up to form the shiny metallic coating on the top of the glass tube. Any gas molecules that are still present in the tube will react with the getter material and become trapped. This process continues throughout the life of the tube, until of course, a crack in the glass forms and air is let into the tube and the getter cannot absorb all of those gas molecules. As the getter reacts with various gasses it turns white, like we saw at the beginning of this article.

How the getter is attached This image shows how the getter is attached to the rest of the electrode assembly, it is spot welded to the suppressor grid. To get the electrodes out, the getter has to be removed. Also in this picture we can see that the control grid also has a “U”-shaped piece of sheet metal connected between its two rods. I am not sure what purpose this serves, if you know, please let me know. The getter serves no electrical purpose and is simply attached to the rod on the suppressor grid for convenience and the fact that the screen grid is usually at a low impedance or even directly grounded.

Removing the Electrodes

Suppressor Grid

Suppressor grid removed After clipping the getter off the suppressor grid rod I can pull the suppressor grid out of the assembly. Here are some specs for the suppressor grid:

• Rod diameter: 1.0 mm
• Wire diameter: 0.13 mm
• Rod distance 12.1 mm
• Distance between turns: 3.0 mm

One point of interest is that the winding is relatively coarse except near the ends where the turns become much closer together. My guess for this would be that that they are trying to do some special beam forming at the extreme ends of the tube to keep the electrons directed inwards towards the plate rather than out towards the micas.

Screen Grid

Screen grid removed The screen grid comes out just like the suppressor grid, with nothing extra needing to be clipped off. The turns on the screen grid are much closer together than the suppressor grid and are uniformly wound the entire length of the grid.

• Rod diameter: 0.8 mm
• Wire diameter: 0.08 mm
• Rod distance 7.4 mm
• Distance between turns: 1.0 mm

Control Grid

Control grid removed The control grid needs the top of its rods clipped off in order to be removed. This is the electrode where the music signal is applied and has the finest wire and closest together turns. It is very close to the cathode and naturally has the most effect on controlling the electron flow in the tube.

• Rod diameter: 0.75 mm
• Wire diameter: 0.05 mm
• Rod distance 4.75 mm
• Distance between turns: 0.35 mm

Cathode

Cathode removed Finally, we have the cathode itself. It is made of a tube of sheet metal and has a metallic oxide coating applied to its surface that helps the emission of electrons. When new and healthy this coating is usually a more even and consistent layer. Over time as the tube ages this coating becomes damaged and results in less electron emission and therefore less amplification.

• Sheet metal thickness: 0.07 mm
• Width: 3.3 mm
• Height 1.6 mm
• Length: 33.2 mm

Grid Comparison

Comparison of the grids and cathode Comparison of the grids and cathode at an angle The two pictures above show each of the grids laid next to each other against a metric ruler. It is obvious as we move further from the cathode the spacing between the turns on the grids becomes less dense. From the angled shot, the oval shape of each grid is also clear.

The cathode and grids reinstalled into the mica without the plate For this close-up shot, I reinstalled all of the electrodes except for the plate into the lower mica to show just how compact this assembly is. You can also see from this angle the notches that are cut into the rods to allow the wires to be properly positioned along the rods. The wires are welded to each of these notches as well. All of the electrons contributing to the plate current must fly from the cathode, through each of these grids without hitting any of the wires.

All the Pieces

All of the parts of the tube laid out and compared with an EL84 Finally, here are all of the parts of the tube laid out together compared to another still intact EL84.

Conclusion

I hope you learned something about how vacuum tubes work and are constructed. I always find that it’s one thing to look at parts on a schematic and learn how to use them, but is is another to actually see how the parts are made and what is actually happening inside of them. Vacuum tubes are the origin of electronics as they are the first devices that allow us to control the flow of electrons using other electrons and provide amplification. I think they deserve a little bit of attention and respect for what they helped us achieve since their original invention in 1906 by Lee de Forest. I will leave you with a picture of the new matched quad of JJ EL84s glowing happily in their new place inside my AC30.

The new matched set of JJ EL84s glowing in the amp