The Amplifier Explained

[toobfreak]

Here is a simple explanation of how amplification is achieved. I will use the vacuum triode as an example, but the same theory applies basically to the diode, other tube types (pentode, tetrode, etc.), and to the transistor.

The main difference with the transistor, especially as it is employed in computer chips is as a switch. It is either switched full on or full off, giving a high and a low (1 or 0), whereas in amplifier circuits, it is continually biased along the linear trans-conductive part of the conduction curve so that equal bias changes on the input results in an equal conduction level on the output.

The basic concept of an amplifier then is to effect a LARGE change/swing in output for a small change/swing in input. The basic vacuum tube amplifier:

You have an input side where a small, line-level voltage signal is fed in, which controls
a much larger output voltage and current on the output side.

This is accomplished with three (tri) basic parts:

1. The cathode which is usually doped with some material which when heated, emits electrons, so is negatively charged (often at ground potential).
2. The anode side (often called B+) which holds a highly positive voltage (may be 300-400 VDC above the cathode).
3. And the grid, which carries a small negative charge, connected to the input signal (decoupled from DC from the earlier input stage with a decoupling capacitor).

In the heated cathode design, the cathode is heated up to emit its electrons. It is this heating up process which is why tubes must "warm up" before operating.
As the cathode heats and emits electrons, they form a "space charge" or cloud of electrons, which are the "fuel" of the circuit, held between the very negative cathode and the slightly less negative but still negative grid. Since like charges repel, the space charge is held here and all electrons are repelled from flowing to the anode. The volume is "down" and the amplifier emits no sound.

Then when the input AC signal is applied to the control grid, the bias of the grid is biased more or less negative, increasing and decreasing the repulsive force of the grid. When the grid signal is biased less negatively enough, eventually, its repulsive force begins to be overcome by the strong attraction of the highly positive anode and electrons escape the space charge, flow past the grid and make it to the anode.

In this way, small variations in the low level input signal close to the cathode effect large swings in electron flow through to the anode (sometimes called The Plate). These cause large voltage swings dropped across the output resistor and current flows back to ground, thus, a tiny input signal in millivolts is /amplified/ to create an analogous but much larger output signal.

A diode is the same except it just has the two poles and the bias determines whether it is closed (shorted) to allow conduction or biased open to block all conduction.

Same with a transistor. The principle is pretty much the same except a semiconductor junction replaces the grid and much lower voltages are used. This is the basic principle behind all amplification.

 

[okfine]

The very first tube invented was a diode tube which is a rectifier. The millivolts you ascribe isn't in the signal chain. That drop is called sag. A tube rectifier cannot keep up with a solid state rectifier in converting AC to DC.

 

[toobfreak]

The very first tube invented was a diode tube which is a rectifier.

The first thermionic tube was the simplest, a tube diode created by John Ambrose Fleming, the Fleming valve. It relied on simple biasing of the inputs to switch its conduction on or off, thus could be USED as a rectifier, by converting the AC bias into a DC current. I have both tube and solid state rectifiers, their use depends upon the application.

 

[trevorjohnson83]

When they dope the silicon with a resistor its like putting boulders in a river, which has real physics to the effect on current.

 

[toobfreak]

When they dope the silicon with a resistor its like putting boulders in a river, which has real physics to the effect on current.

What they actually do is add impurities to the base silicon or germanium. Things like antimony, which gives the outer valance shells of the material a moderate number of electrons. This means that pressed enough, the silicon becomes a conductor but not a good one, but it also becomes a resistor, but not a good one--- it becomes a semi-conductor.

So, given the electric pressure, the substrate can be biased "on" to conduct, or be biased off to not conduct. Better still, once the flow of electrons is interrupted, they tend to stay off, while once biased forward, the current-carrying capability tends to "give in" and current tries to keep flowing.

That is one of the big differences between a solid state transistor and a vacuum tube--- while transistors prefer to be either full on or full off like a switch, tubes operate much more happily like a valve holding an infinite number of positions between full on and full off, which is why they are much more musical than a transistor.

 

[scruffy]

For.most useful.devices, the input-output relationship is an S shaped curve. It's "sigmoidal".

Vacuum tubes tend to be more linear, but transistors also have a linear region, in the middle of the S.

This is an example of an S shaped input-output relationship:

You can see the "approximately linear" region in the middle of the curve. If you want linear amplification (like for audio, for instance), you'll try to keep your device operating within the linear bounds. (Any deviation from linearity results in "distortion").

In an amplification device or circuit, the vertical scale will be bigger than the horizontal scale. For instance a microphone signal (or guitar pickup) is in the millivolts, so that'll be your horizontal scale, whereas the output of the amplifier will be in volts, which will be your vertical scale.

Interestingly, this same S-shaped curve underlies all of modern AI. It's called the "activation function" in neural networks. The machine learning people use variants of it, like tanh, ReLU, or the logistic function.

In machine learning, they use the nonlinear areas too. The nonlinearity is introduced deliberately, and it's essential for the proper behavior of the circuits. In digital logic you're basically ignoring the linear region, and you're only interested in whether your device is "on" or "off". But in brains, you can see how the entire curve can become useful. If you think of the curve as representing a behavior, you get "none of" the behavior on the left side of the curve, whereas if you're on the right side the behavior dominates (perhaps to the exclusion of other behaviors). If you're in the middle of the curve in the linear region, you can control "how much" of the behavior you get.

In brains as in tube amps and transistor amps, the "bias" is important. It's a set point that determines the quiescent behavior (when there's no input signal). In audio amplifiers you want your set point approximately halfway, in the middle of the S, that way you get an approximately equal swing in either direction, and your circuit is designed to keep the swing within the linear region (in other words you can't have "too much gain", otherwise you'll saturate on one side or the other and distortion will result).

In brains the set point is more flexible, depending on the application. In retinas the set point is near the middle so you can swing between conditions of low light and bright light, and there is automatic gain control that tends to keep the contrast within useful bounds. On the other hand, most motor systems are biased way to the left, you want "none of" the behavior until it's specifically turned on - and things like reaching get "how much" whereas fight-or-flight tends to saturate on the right.

 

[toobfreak]

For.most useful.devices, the input-output relationship is an S shaped curve. It's "sigmoidal".
Vacuum tubes tend to be more linear, but transistors also have a linear region, in the middle of the S.

The plot you show is more characteristic of the smooth transconductance of a tube in Class A.

It has a generous region at center where one gets a straight (linear) transition in conductance to an equal swing in input voltage. And yes, special transistors can be designed and forced to behave more linearly and decades have been spent (since the 1970s) trying to force transistors into a linear operation, the truth is that transistors generally want to respond non-linearly. Here is a common plot to a typical field effect transistor; note the far steeper slope to the linear (ohmic) region, which quickly transitions into full on saturation leading to breakdown and permanent damage to the device.

 

[trevorjohnson83]

What they actually do is add impurities to the base silicon or germanium. Things like antimony, which gives the outer valance shells of the material a moderate number of electrons. This means that pressed enough, the silicon becomes a conductor but not a good one, but it also becomes a resistor, but not a good one--- it becomes a semi-conductor.

So, given the electric pressure, the substrate can be biased "on" to conduct, or be biased off to not conduct. Better still, once the flow of electrons is interrupted, they tend to stay off, while once biased forward, the current-carrying capability tends to "give in" and current tries to keep flowing.

That is one of the big differences between a solid state transistor and a vacuum tube--- while transistors prefer to be either full on or full off like a switch, tubes operate much more happily like a valve holding an infinite number of positions between full on and full off, which is why they are much more musical than a transistor.

yes boulders in a DC current of the river flowing. we'll just skip past what you said, it's like what we do to each other around here cause sometimes we just want to hear what we have to say and the reason is just because it has a lot to do with not caring so much about what you think and just taking it easy at the computer like you were asleep at the beach and you were laying there relaxing and you could catch up on some zz's this way without it being a problem.

 

[trevorjohnson83]

I learned boulders from a park ranger website and it goes like this the boulders slow down the current on the side facing the flow and on the other side the current speeds up. Current also speeds up when the streem meets with rock bed that makes it shallow the water speeds up where shallow. Also unrelated to boulders and rock, the center of the stream is most rapid and the outside and bottom are apt to resistance that slows it.

 

[52ndStreet]

Can you now explain how today integrated transistor Amps work, without that old time tube amplification ??. They don't need to warm up in order to operate.??

 

[toobfreak]

Can you now explain how today integrated transistor Amps work, without that old time tube amplification ??. They don't need to warm up in order to operate.??

Most consumer electronics is now driven by Op amps (operational amps) instead of discreet circuitry, but the principle is still basically the same. BTW, an integrated amp is generally considered a preamp and power amp together in the same chassis.

On paper at least, op amps appear ideal, however, nothing still sounds as good as tubes.

Operational amplifier - Wikipedia

en.wikipedia.org

 

[52ndStreet]

Transistor integrated circuit amps can also shut themselves off, to protect your equipment from power surges, or if they get to hot internally. Old tube amps can blowout from a power surge, or if they get to hot from prolonged use. That is why I prefer Transistor integrated circuit amps. They have internal protection circuitry inside. At least the ones that I have purchased.!!? Do the research.

 

[toobfreak]

Transistor integrated circuit amps can also shut themselves off, to protect your equipment from power surges, or if they get to hot internally. Old tube amps can blowout from a power surge, or if they get to hot from prolonged use. That is why I prefer Transistor integrated circuit amps. They have internal protection circuitry inside. At least the ones that I have purchased.!!? Do the research.

Not true. PS: I design the shit. Any amp can be made to protect itself from any fault condition. Transistor amps have more of that stuff because they need it. Rather hard to change out a power transistor compared to just swapping out a bad tube!

As to power surges, etc., you have it backwards, transistor amps are far more susceptible to damage from a line fault or some internal damage. Most tube gear can withstand an EMP without damage.

Tubes have 300 volt rails. Transistors have 5 volt rails. Guess which one is more easily damaged?

Main reason to get a transistor amp is:

1. More compact
2. Lighter
3. Cheaper to operate
4. Less servicing

Tube gear definitely relies on greater user skill than transistorized stuff.

 

[HaShev]

I thought this was gonna be a topic on
optical amplifiers which sweeps frequencies to cool the radiation on
high intensive gas lasers.

 

[scruffy]

Can you now explain how today integrated transistor Amps work, without that old time tube amplification ??. They don't need to warm up in order to operate.??

Not only that, they work straight off the AC line. They'll suck up as much current as you're willing to give them. That's why you see 2000 watt amps that are only an inch and a half tall. No transformers.

 

[52ndStreet]

Not true. PS: I design the shit. Any amp can be made to protect itself from any fault condition. Transistor amps have more of that stuff because they need it. Rather hard to change out a power transistor compared to just swapping out a bad tube!

As to power surges, etc., you have it backwards, transistor amps are far more susceptible to damage from a line fault or some internal damage. Most tube gear can withstand an EMP without damage.

Tubes have 300 volt rails. Transistors have 5 volt rails. Guess which one is more easily damaged?

Main reason to get a transistor amp is:

1. More compact
2. Lighter
3. Cheaper to operate
4. Less servicing

Tube gear definitely relies on greater user skill than transistorized stuff.

If you design Amps, you should know that not all old Tube amps, have protection automatic shut off circuit breakers.?, which is incorporated into many newer integrated , and transistor circuit amps. The old tube amps tend to be bigger, and more bulky, than the integrated circuit amps. I still prefer the integrated circuit transistor amps. They are smaller, lighter, and can be rack mounted more efficiently. The sound quality, is just as good, with some brands , as tube amps. Just my opinion.? Many integrated circuit transistor amps can out put a lot of power also..

 

[okfine]

Not only that, they work straight off the AC line. They'll suck up as much current as you're willing to give them. That's why you see 2000 watt amps that are only an inch and a half tall. No transformers.

Class D

 

[toobfreak]

Not only that, they work straight off the AC line. They'll suck up as much current as you're willing to give them. That's why you see 2000 watt amps that are only an inch and a half tall. No transformers.

No transformers? You must be talking about Class D and Class H amps.

 

[okfine]

If you design Amps, you should know that not all old Tube amps, have protection automatic shut off circuit breakers.?, which is incorporated into many newer integrated , and transistor circuit amps. The old tube amps tend to be bigger, and more bulky, than the integrated circuit amps. I still prefer the integrated circuit transistor amps. They are smaller, lighter, and can be rack mounted more efficiently. The sound quality, is just as good, with some brands , as tube amps. Just my opinion.? Many integrated circuit transistor amps can out put a lot of power also..

Older tube amps generally used a slow blow fuse on the mains. Then engineers used at least 3 MOV's in series with a regular fuse.

 

[scruffy]

lol - I had lunch with a 21 year old today, I go "you're probably too young to remember vacuum tubes" -

She goes, "you mean like hoses on a vacuum?"

I go "no, like an old radio, you know, tubes". Blank stare, glassy eyes.

I'll have to take her to the Guitar Center and give her an education.

I like the 4x or 6x 6550 amps. Ampeg V9 or SVT, Marshall Major, old 300 watt Sunn, like that. Roadies are essential, but the sound is incomparable. Guaranteed to deafen even the hard of hearing.

In the Marshall "Pig" (the earliest version of the Major) they were running 700 volts on the plates, and the cabinets were so good that if you cranked the amp it would blow the output cable right out of the jack. Which in turn made the amp blow up, it would literally light on fire because the tubes got cherry red and melted. But before they did that the plate wires would arc over inside the chassis.

Which is probably why Pigs are extremely rare these days.

Ah, the good old days. Now they use locking jacks to keep that from happening. If you back down to 600 volts it won't arc either, you'll be a lot safer.