24 June 2012

Guitar Amp: How it works - Part 3

The Power Section

In previous articles, we saw that everything related to first amplification push, tone shaping and filtering occurs in the preamp section but, the power levels that the preamp section handle are not big enough to move an speaker, still.
And that's just the purpose of the Power Section, to make the last push to the signal to get enough power to move a set of speakers.


Driver or Phase Inverter tube

Even that the Phase Inverter (PI) or Driver is the last tube in the preamp section and, it is usually of the same family than the rest of preamp tubes (12AX7), it's really a link between the preamp and the power amp sections.
In fact, it wears before rest of preamp tubes, with same frequency as power tubes do.
It is highly recommended to swap the PI tube when you are swapping power tubes, while the rest of preamp tubes can usually last twice the life of power and PI tubes.

Its task is just to feed the input of the power tubes with the signal that was already sculpted in the several tubes, caps and resistors of the preamp section.
Usually, when the amp is operating in Class A mode, this tube is being named Driver and, when operating in Class A/B, it's being named Phase Inverter, because alternatively inverts the signal's phase to feed one or other tube (or rail of tubes) of the power amp, but, this will be discussed below.

Most of tube amps are using a 12AX7 or 12AT7 tube for this position.
The Driver or PI is not really adding anything to the already sculpted signal from the pre-amp but, it's highly responsible for the feel and responsiveness of our amp.
Tubes with high transconductance values are very valuable here, since they are really fast handling the input signal and converting it in the output signal that the power amp needs.

A 12AX7 tube gives a later break-up point and, power amp distortion begins when the Driver or PI starts to distort. Also, depending on the strength of the signal that the PI is sending to power tubes, these start to break up also.
So, an early break-up point in the Driver, creates a pre-power distortion that has different characteristics than the power tube's type of distortion. Therefore, one of the goals is to avoid the PI to distort but, providing power enough to the power amp to let power tubes to reach their sweet spot, when they deliver those outstanding harmonics that fill the sound of incredible nuances.

A 12AT7 is slightly different of a 12AX7 tube. In first place, it produces way higher current (density of electrons) and, in second place has a lower gain factor. One trick that many people uses to let the power section to break-up at lower volume levels is just to swap the Driver or PI with a 12AT7 or 5751 tube.

My experience says that not all tubes work the same in this position. While some tubes handle speedy riffs with ease, others seem to deliver a delayed response time. It's not so dramatic as it seems but, makes you to slow a bit your fingering technique to be in sync with the response of the amp. Other effect that I've noticed is related to legato between notes. Some tubes make the link between notes more clear, some others have a mushed behaviour. It's up to you to experiment with the PI but, it worths the time if you are experiencing problems like those I've explained.

Finally, not all the tubes last a long time in that high demanding position. In my experience, the Sovtek 12AX7-LPS is an outstanding tube for this position but, it lasts very few there.
One of the amps that I had, started to deliver a flubbed sound, as if the speaker's cone was broken. After swapping all the tubes, the issue was just an LPS going bad on PI position.
I felt similar issues, after short time of loading an LPS in PI position, in several amps.
It is a pity, since the LPS (while it stands) works awesomely in this  position (nice transconductance, power and transparency) but, it doesn't last too much.
Usually, tubes with spiral filaments last less in PI positions and, on cathode-follower configured triodes (the only exception seems to be Electro-Harmonix 12AX7, that works fine anywhere).
Sometimes, you have to sacrify tone for reliability. JJ tubes and Chinese tubes stand long in those positions. EH 12AX7 works fine in any position, also. Mullard 12AX7 reissue is an outstanding tube in any position, also (but its price makes me to reserve it for tone-shapping positions, instead).


Power Tubes

In most of cases, power tubes are Pentodes and, usually of family 6V6 (American tone) or EL84 (British tone) for low powered amps or 6L6 (American tone) or EL34 (British tone) for high powered amps.
There are other tubes that are used in a minor way as the 5881, KT66 or KT88.

Principle is always the same, the one or more power tubes of any of those classes should increase the power level of the pre-amp section up to the level required to move one or more speakers.
Election on number of tubes and tube layout will depend on which output power we need and, which kind of "color" we want for the sound.


Class A and Class A/B, what does it mean?

There is a lot of confusion related to Class A and Class A/B operation but, definition has nothing to see with First or Second class or the number or arrangement of power tubes. Operation classes define just the amount of time that a power tube (or set of tubes) is working, during a cycle.
Class A amps have their power tubes working the full cycle (100% of time).
Class A/B amps have their tubes working the 60% of the cycle. Since they cover more or less half of the cycle, they need to work in pairs to cover the full cycle. Imagine the pistons of a motbike's motor, one goes down while the other goes up.

Class A tubes are working highly stressed, since they are working even when idle (when no sound is coming from our guitar). Amps designed as Class A, deliver less power than same amp working as Class A/B. Reason is that since in class A/B, each tube is responsible for amplifying half of the signal (the positive or negative part), it has the same room to amplify a half than the class A tube for the whole signal. So, each tube of the pair is able to amplify more its half signal and, the sum is more output power at the end.
Remember the discussion about amp's efficiency?. The most efficient design is class A/B, in this case.
Class A amps tend to deliver warmer tones while Class A/B tend to deliver thicker tones. It seems that they distort in different ways and that, they generate different harmonic content.
 
Usually, people tends to call Class A to those amps that work with power tubes that doesn't need a BIAS but, this is definitively wrong. Let talk about this below.
 
 
BIAS, fixed bias and cathode-biased?
 
All tubes need to be biased, even transistors. What does it mean?.
 
Amp designers choose some resistor linked to the grid to avoid the electron's flow between cathode and anode, when no signal (from the guitar) is present in the circuit (quiescent state of tube).
Following the example of that typo N (mentioned before, during gain's discussion), imagine that we have our monitor with a bad setup, in a way that the center of the screen is some points higher or lower, righter or lefter. If we zoomed to the maximum that N (without distorting it), in a well setup monitor the typo would be clearly zoomed in, without loosing a pixel of its strokes but, same zooming in a monitor with a wrong centered setup would distort that N in some parts.
Talking about valves, a bad bias spot is quite similar to that uncentered monitor view (unbiased) so, the first goal of bias is to "center the image of our signal" to achieve the highest fidelity and lowest distortion.
 
A pair of tubes working in class A/B, are active 60% of the time so, one is "lighting" when the other is off then, the opposite happens. In the intermediate state when a tube is going off and the other one starting up, there is some overlapped work that generates what is called the crossover distortion. Other of the goals of the bias point is to minimize this effect.
 
Bias has also a lot importance to maintain the highest performance of the tube, avoiding the wearing of the same. During its operation, both, anode and cathode are working under high temperatures so, the surface of their metallic parts start to degrade. Part of the energy generated (or transformed) in the tube is being lost in form of heat (plate dissipation), what is an important value important to control, since a bad dissipation would destroy the plate (and, eventually create harder issues inside the amp, even fire).
 
Some people moves up a tad the bias spot specified by the maker of the amp, to make the sound feel thicker and "on steroids" but, this can lead to a quicker tube degradation. Even moving up the bias a bit has its security limits, there is a point were isn't safe to go further. Moving the bias down can cool down the sound of the amp, making it lifeless. So, please, consider this: the amp's maker is taking into account everything, the technical variables that should be controlled as well as the sound that the amp must deliver so, their recommended bias point usually makes all the sense.

OK. I said all tubes must be biased but, we are usually changing the bias of power tubes when swapping them. Usually, there is no need to re-bias any preamp tube. The value of the resistor is fixed during design phase and doesn't need any re biasing task.

In power tubes, we can have two different situations: cathode-biased and fixed bias designs.
Paradoxically respect of its name, fixed bias is the situation where we will need to re-bias the variable resistor that the maker leaves for that task. To get the best of the amp, we will always need to re bias after a tube swapping.
Cathode-biased designs automatically readjust their bias point while operating so, there is no need for us to re bias after a tube swapping.

I don't know why but, most of designs made around EL84 tubes are usually cathode-biased, while the rest of amp's designs seem to go for a fixed bias one.
This is one of things that drove to confusion about operating classes. Most of people tend to name an amp as operating in class A when they have cathode-biased tubes (that don't need re biasing) and, this is definitively wrong.
By example, the famous Vox AC30 is a class A/B amp with cathode-biased design. Vox Night Train 15W is a Class A, cathode-biased design. A Marshall 1923C (or DSL) is a class A/B, fixed-bias design, VHT Special 6 Ultra is a Class A, fixed-bias design.

The truth is that Class A and cathode-biased designs are more often seen on low wattage amps, while Class A/B and fixed-bias designs are more common on high powered amps (because of efficiency and stress).


Pull Push / Single ended

Finally, power tubes can be connected to the Output Transformer (OT) in two different configurations: pull/push and single ended. This determines also the type of OT to use.

The pull/push configuration needs at least two tubes. The power supply is connected to the center-tap of the OT and each tube is connected to one of the two connectors of the OT, the upper and lower end of the center-tapped primary.  In that way, each tube works with a phase of the signal.

To each of this sides one or more tubes (always in pairs) can be present, each group of tubes in the same row is known as a rail, and when there are more than two tubes (that is, 4, 6,or 8), this configuration is being called parallel pull/push. Usually, amps of 50W work with just two tubes in pull/push, amps of 100W have 4 tubes, 2 on each side or rail. Tubes in same rail work together during the same phase of the cycle and opposite to the tubes on the other rail of the pull/push.
Usually, when an amp has more than two power tubes, the tubes at right hand are in the same rail and the tubes at left hand in their own rail but, couples or pairs go from outsider to insider tubes so, the most lefty and the most righty are a pair, etc.

To downgrade the power of the amp, you can remove just as many pairs as you need. If the amp has 100W, then 4 tubes, you can remove the two outsiders or the two insiders (the outer or the inner pair) but, you cannot remove the two on the left or the two on the right since, in this case, you are leaving a rail void and, half of a cycle there will be no output of signal (despite of the damage that can take the amp in that case).

In the case of single ended configuration, the plate of the output tube (or tubes) is connected to one terminal of the OT while the other terminal of the OT is connected to the power supply.
As in the case of pull/push configuration, there can be more than a tube in the same (unique) rail, in that case, this configuration is known as parallel single-ended.

Class A/B amps are usually designed with a pull/push or parallel pull/push configuration while Class A amps are usually designed with singled-ended or parallel single-ended configuration. But, we aware that we can also have Class A in a pull/push or parallel pull/push configuration.


Swapping Power tube types?

If an amp was designed for 6L6s, it usually will not work properly if swapping those with EL34's and viceversa. While in the pre-amp section there is a bunch of freedom, possibilities in power section are very few.
Power and rest of technical specifications change from tube to tube and, they have a very different response related to peaks and valleys. So, at least that the amp's maker did a design that specifically says to you that will support tubes type X, Y and Z, a direct replacement is a bad idea.

There are some tricks in the market to achieve such a changes. By example, TAD (and, even VHT amps, if memory doesn't fail) has some kind of socket-addapters that can be plugged instead of the octal tubes (EL34, 6L6, 6V6) to mount EL84's (significatively lowering the power).
The VHT Special 6 Ultra, supposedly can support any tube BUT, going to high powered tubes (EL34 or 6L6) would require a substitution of the fast blown fuse with a slow blown fuse (because of the power peaks of the powerful tubes) and, you can run the amp in such a configuration for no more than half hour (because the Output Transformer will not support the stress longer).

For sure, there are designs which parts are perfectly tuned to support any kind of tube, as the THD Bivalve or other amps that are mounting some hybrid configuration of different types of power tubes.
Check the market, there is something new around every time.


Presence and Contour controls and negative feedback loop

You will see some Presence control in some amps. What this control does is to hijack the highest frequency range and let you to decide how much of this high end do you want to restore to the signal. This is usually done after the power tubes' amplification work.

Also, part of the interaction of the amp with the speaker is sent back and mixed in the input of the Power tube in some way that is called the negative feedback loop.  This makes the amp to sound thicker, more compressed and with less dynamics.

The Contour control usually is a control that links to a network of passive (or active) filters that equalize the sound with different shapes, increasing or decreasing some specific range of frequencies. This is usually also done in the power section.


Output Transformer

The Output Transformer (OT) is responsible for adjusting the power generated in the amp to the needs of the external speakers. The OT will have several output taps, corresponding to the several impedances that correspond to speaker's configurations that the amp will support.
The impedance is some kind of resistance (with a reactance, expressed as a Complex Number but, this makes the term even more cryptic and difficult to understand) and, we know that if we increase the resistance the power is lowered so, an amp that delivers 100W at 4 Ohms, will deliver 50W at 8 Ohms and 25W at 16 Ohms, by example. Volume will not be affected as dramatically, the difference between double of watts is just 3dB of volume difference so, going from 100W to 25W we are loosing 6 dB in terms of volume.
Usually, a gain pedal with a Boost switch will provide an extra of +3dB to the signal, just to make the instrument to sound a bit louder in the mix, during solos (just to give you an idea of what 3dB of difference mean).
So, we can put an speaker with higher impedance (16 Ohm) in an exit of the amp prepared for a lower impedance (4 Ohm) and, it will work but, the power will drop to 1/4 and, volume will decrease in 9dB. This can be an useful trick to tame your beast of amp and make it not so loud.
The opposite is not recommendable in any way, to put a lower impedance speaker in a higher impedance exit of the amp can severely damage the OT, among the Power tubes and, even burn the amp. So, don't do that.
In terms of efficacy and reliability, you should match the speakers impedance the the same impedance in the output of the amp but, if there is no way (no available output for such an impedance), just put the higher impedance speaker in a lower impedance output of the amp (16 in 8 or 8 in 4 or 4 in 2, by example).

It seems that the OT, differently of the PT, has some clear influence in the sound. It seems that bigger OTs, with certain iron cores, reinforce the bass frequencies, while tinnier OT are more loose in that range.
The OT seems to add also other kind of nuances to the sound so, it's a very well watched part when designing an amp.

No comments:

Post a Comment

Please, feel free to add your comments.