• You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.

Solid-State Amplifiers: What is distortion and how do I reduce it?

eagle1911

Active Member
Jan 21, 2011
327
20
28
Based on a couple posts I made in several different threads, it occurred to me (after a suggestion from Moleculo) that this site could use a thread dedicated to solid-state amplifiers and how to keep their output clean. Maintaining clean output from a station is a foremost obligation of any radio operator and service, regardless of what class of license the op holds or if there is a license required at all.

I don't want to get into too much technical detail here, primarily because it's lengthy and dry and because I'm not an expert on the level where I would feel comfortable providing too much detail. Also, simple explanations are easier for anyone to understand. I'm a hobbyist first and foremost, so everything I share is based on my personal experience, which has been helped along by various texts.

First off let me describe two concepts which define amplifiers: Linearity and Non-Linearity.

Almost everyone in radio has heard RF amplifiers be called (most of the time incorrectly) simply "linear amplifiers". This refers to the ideal condition in which an amplifier provides a completely unmodified, but LARGER in amplitude (more powerful) copy of the input wave.

In reality, no amplifier is perfectly linear. When a wave is input, the resulting output is always altered in more ways than simply being more powerful. This extra alteration is referred to as "non-linearity" or more commonly "distortion". The factor that determines how close to being truly linear an amplifier comes is it's class.

Amplifier Classes:

There are 4 main classes of amplifier (more actually, but 4 for our purposes here): Class A, Class AB, Class B and Class C.

Class C amplifiers are the simplest amplifiers in terms of design complexity. To help describe things a little better I'll refer to the waveform in degrees, with a complete waveform consisting of 360 deg. and partial waveforms as less than 360 deg. If only half of the waveform appears at the output, we would say that the amplifier is conducting 180 deg. of the signal etc.

Every transistor has what is referred to as a "threshold". This is the voltage at which the transistor begins to conduct. By it's nature, a radio signal consists of a train of waves. A wave consists of varying voltages and/or currents. A sine wave, as it pertains to radio, typically has a positive portion (half of the wave consists of positive voltages) and a negative portion (the other half consists of negative voltages). The simplest and most common transistors (BJTs or bi-polar junction transistors) have a threshold of .7V. This means that when a wave is input into a single BJT transistor in the simplest possible amplifier, the transistor only produces output when the input wave is at .7V or higher. Obviously this means that that the rest of the wave (the entire negative portion and some of the positive portion) is missing from the output signal. The amplifier is conducting less than 180 deg. which means most of the original signal is missing. This is a classic example of nonlinearity. We put in a sine wave and we get only part of it back. This chopped up output wave is said to be "distorted". This describes "class C" amplifiers. Even though the output is distorted, class C amplifiers are still very useful, particularly with FM (by it's nature FM is unaffected by a distorted carrier) and collector-modulated AM (more complex and not very common with small brick amplifiers). Class C amplifiers are not suitable for pure AM and SSB usage however, as these modes are very much affected by distortion, causing bad sounding modulation and interference to other equipment and radio operators.

Since the threshold of a BJT transistor is ~.7V, we can help eliminate a lot of this distortion by ensuring that more of the input signal is always above threshold. When more of the input signal is above threshold, we get more of the signal at the output. We accomplish this by adding a DC voltage equal to the threshold directly to the base (input terminal) of the transistor. By adding a .7V DC voltage-only (negligible current) source we are adding what's called "bias". This bias insures that the signal cannot fall below .7V, which means that the entire positive portion of the input waveform will be above .7V, while the negative portion will flat-line at .7V. This gives us exactly half of the original waveform at the output. This refers to a class B amplifier, which conducts exactly 180 deg. of the input signal. Since 180 deg. at the output is more than what a class C amplifier produces, a class B amplifier is more "linear" than a class C amplifier but still leaves much to be desired in terms of linearity.

If we add bias consisting of both a .7V DC source AND a bit of current to go with it, we end up with a class AB amplifier. Since BJT transistors are current-controlled devices, the extra current helps ensure that even more of the input waveform is produced at the output, usually a bit more than 180 deg., which means that the majority (but not all) of the input waveform appears at the output. Obviously this means that a class AB amplifier is more linear than a class B amplifier, so now we are getting to the point where we can use the amplifier with AM and SSB, which require low distortion. Typically a class AB amplifier is the lowest class (in terms of distortion) that can be used with these modes.

Finally, if we add bias consisting of a higher voltage than the threshold (for example 1-2V) along with some current we can reach the pinnacle of low-distortion amplification, class A. The extra voltage and current push all 360 deg. of the input waveform above threshold and therefore the transistor conducts all 360 deg. of the input signal, giving 360 deg. at the output. This produces the lowest possible distortion level, and the cleanest output. Class A amplifiers can be used with any mode or type of signal.

This graphic shows each class along with the portion of the input waveform that is successfully duplicated for each. The shaded areas are what ends up being lost:

14179_81_1.jpg


After learning the details of the various classes it's easy for a person to think that class C is inferior and class A superior, which leads to the question: Why use anything but class A? The answer is Efficiency: A class C amplifier produces the most distortion, but also has the best efficiency. A class A amplifier has clean output but the least efficiency. As an example, this means that a single-transistor class C amplifier using an MRF455 will produce more power for a given input current than a class A amplifier with the same transistor at the same current level. Classes B and AB fall in the middle of the efficiency curve.

The above pertains to BJT transistors. MOSFET transistors are increasingly popular, and the main difference in terms of class of amplifier using MOSFETs is the fact that a MOSFET's threshold is usually on the order of 5-6V instead of .7V, so even class C MOSFET amps have some bias, and to bias a MOSFET requires higher voltage than a BJT. Also, since MOSFETs are voltage controlled (as opposed to current-controlled BJTs) the bias circuitry is actually simpler to implement.

Now that I've touched on distortion and how it relates to the various amplifier classes, I'll discuss how to clean things up.

There are 2 types of distortion found in RF amplifiers: Intermodulation Distortion (IMD) and harmonic distortion.

IMD is a function of the transistor itself, and most transistor datasheets specify a figure for output IMD at a given input level. IMD is therefore controlled and reduced by proper circuit design in the amplifier itself. A single-transistor amplifier (single-ended) will exhibit the worst IMD performance, and push-pull amplifiers, due to certain variables, help keep it comparatively minimized. Normally the very best way to minimize IMD is to "over-enigneer" the amplifer. Over-engineering means driving components well below their maximum ratings. If you've got a 70W transistor (as in 70W stated in the transistor datasheet) then good engineering would dictate that you only drive it to 50W for low IMD figures and enhanced ruggedness and therefore lifetime. In short, to minimize IMD NEVER overdrive a transistor and always use push-pull designs. A push-pull design uses 2 transistors in tandem, with each transistor helping to fill in the missing portions of the signal that are lost by the other transistor. Between the two transistors we get a fairly good copy of the input waveform.

The second form of distortion is Harmonic Distortion. Harmonics are sub-waveforms that are formed my an amplifier at 2X, 4X, 8X the original frequency etc.. Normally these harmonics can contain a good bit of power, and they occur well outside of the band that the amplifier is intended for. This causes interference in may ways, none of which are good. Fortunately, it's not hard to deal with them. All that needs to be done is to filter them out.

That brings me to filters, which I'll spend a bit of time describing since they seem to be shrouded in a bit of mystique, especially with CB enthusiasts.

Filtering Out Harmonics:

Harmonic filters are usually built from combinations of inductors and capacitors. These components are combined in to fairly simple networks with values that give a specific, desired response.

Here are some circuits and plots that should help demonstrate what harmonic filters are and how they work (made these for a different thread that some folks may already have read):

The red curve is the transmission dB of attenuation, which is the amount of incoming power that makes it through the filter. Any power that does not make it through the filter are losses. The blue curve is the VSWR curve, the high points of which are proportional to the transmission losses as can be seen.

here's a fairly basic 5-pole Chebyshev filter for 30MHz:

5-poleschematic.gif


and it's transmission/VSWR curves:

vswrandcutoff5-pole10mLPF.gif


You can see some large variation in the input VSWR indicating variations in input impedance which causes losses, heating, and potentially damage within the filter at those operating points where the SWR is high. You can see by the transmission curve (red) that there are slight losses occurring where the SWR is higher. At high power levels, even slight losses add up to a good amount of heating and damaging effects.

Here's the same filter with 2 more poles added, making a 7-pole Chebyshev:

7-poleschematic.gif


Plots for the 7-pole Chebshev:

vswrandcutoff7-pole10mLPF.gif


You can see that the input SWR across 10-12m is much more acceptable, resulting in less losses and heating at those frequencies.

Here's a whole different animal, a 7-pole Elliptic filter for 30MHz:

7-poleellipticRFProjectsschematic.gif


and plots for the 7-pole elliptic:

7-poleellipticRFProjects-1.gif


Here you can see that the input SWR is very low across the whole HF band, which makes this a particularly good design, but more complex to build and the power-handling capability is somewhat more limited.

Typically, performance is pretty much always better with more complex filters, within reason of course. That's not to say that simpler filters aren't useful. It's possible with the right math and the right components to have simpler filters (5-pole) with better SWR performance than seen in my example above. Also, more complex filters provide sharper cutoff and in the case of elliptic filters they can even be tuned to suppress specific harmonic frequencies. The transmission curve for the elliptic filter above shows these distinct and sharp suppression spikes after cutoff.

In conclusion, by using good engineering sense when designing or modifying an amplifier, along with well-designed harmonic filters it becomes possible to have an amplifier with good efficiency, very low distortion and therefore very little potential for interference. The bottom line is that this is about maintaining your station's output out of respect for others. Running a station that produces dirty output will almost invariably cause RFI problems across the radio spectrum. Whether you as the station operator are aware of this RFI or not, it is still there, and it is still causing problems for other people, whether it's messing with their phones, TVs, and toasters or bleeding on a public-service frequency used by a local PD or FD. Either way it causes problems, and some of those problems could threaten public service and safety in the worst case.

To put it bluntly, radio operators who only care how many watts they are "putting out" with no regard for whether that power is clean are almost always guaranteed to be interfering with someone else. I've heard and seen a lot of stuff online and on the air about how certain operators feel like people are unfairly calling them out for using poorly built, overdriven class C garbage on AM and SSB, but in reality it is NOT AT ALL unfair that they are being called out. Who likes it when people are so selfish that they don't care about anyone but themselves? NO ONE. That's just a fact of life, not just the radio hobby. That's why responsible, conscientious operators care about these things. The more people that care about proper engineering and good station operating techniques, the less noise and trash there will be on the bands, and the less interference there will be to other systems and neighbors. Obviously that's good for everyone.

Personally I would rather have very low power that is clean than lots of power that is dirty. Even though people might hear me better with gobs of extra juice, it's not worth it to me if it means I'll be causing other people and services grief. Basic "golden rule" stuff here.. not rocket science.

I'll follow up with more info as soon as I have more time regarding the construction of a set of filters, for the purpose of bringing the concepts discussed here regarding harmonic filtering into the real world.
 
Last edited:
  • Like
Reactions: 2 people

Nice, well thought out article. I especially like all the diagrams that help articulate the subject matter. I look forward to the rest!
 
This brings a question to mind. Most solid state RF amps I've seen (both for ham radio use and for CB) typically use transistor pairs, which I'm assuming are operating in a push-pull configuration, as described in the article. That is, for a given sine wave cycle, one transistor amplifies one half of the sine wave (when the voltage goes positive) and the other transistor amplifies the other half (when the voltage goes negative).

I would think that for push-pull, you would want both transistors biased for class B operation (each conducting for exactly 180 degrees of the input signal). But I've also heard people say: "Oh, that guy's amp is operating in class C." It's usually the case though that "that guy's amp" has a transistor pair, which to me says it's push-pull.

So I guess my question is: is it typical to design push-pull amps with something other than class B biasing? Using class AB seems like it would work though you'd lose some efficiency (there would be overlap), but using class C would increase distortion, which is what you're trying to avoid by using push-pull in the first place.

For all I know, it could have been that calling the amp "class C" in that context was just a misguided attempt at trash talk, but I honestly don't know enough about CB amp design to tell if there's any truth to it or not. Do they actually bias push-pull amps for class C (on purpose)? Are they using some other strategy besides push-pull? Inquiring minds want to know. :)

-Bill
 
You've already got a pretty good understanding it seems, as you're pretty much right on with most of your speculation. With class B and AB amplifiers it is pretty much standard operating procedure to use push-pull pairs, but with class C it's actually rare outside of CB amp setups. Most class C arrangements that I've seen in commercial and amateur designs are single-ended and many of them use several single-ended stages in parallel, but usually no more than three. By my logic though (and this is speculation) a push-pull class C amp will have slightly better linearity than a single-ended class C amp simply because the 2 transistors together still produce a more complete waveform than just one transistor in class C. The CB amp folks use push-pull exclusively, but many if not most transistor CB amps do not have any bias on the transistors, making them class C amps. The IMD performance would still be as described in the device datasheet (if specified, many have no published IMD spec) as long as the transistors are not being overdriven, and any harmonics could still be taken care of through filtering in commercial and amateur designs.

Also, push-pull designs have an inherent benefit: The design naturally results in a suppressed 2nd harmonic. You've still got to deal with the higher harmonics, but the 2nd harmonic in a single-ended design can be quite powerful and difficult to filter so it's nice to have it naturally suppressed. Also this inherent suppression is what allows for broadband HF amplifiers to utilize only 6 band-switched filters instead of needing a separate filter for each band. This applies to ANY push-pull design regardless of class, and this is why class B and AB designs are almost exclusively push-pull. I think that the choice between single-ended and push-pull when dealing with class C is a matter of cost and efficiency. Because of the efficiency provided by class C operation, a single device can go a long way and not generate too much heat. There is substantial cost savings if only a single transistor can do the job, and the added cost of filters with sharper cutoff to deal with the strong 2nd harmonic isn't so much that the cost savings are reduced much.
 
Last edited:
Agreed, and thanks.. I hadn't actually seen that the classes had already been discussed at length here, but really it's sort of common sense that it would have been lol. It is "The Worldwide Radio Forum" after all..

Describing the classes was sort of linked to the topic, which is dealing with distortion, so I might have still felt compelled to write it all down even knowing that there were already whole threads on the topic..

I wish that the OP was still editable.. there are a few typos, errors and misspellings I'd like to correct. I incorrectly described the transmission curve in the plots as representing "percentages", but they actually represent dB of attenuation.
 
Last edited:
" It's usually the case though that "that guy's amp" has a transistor pair, which to me says it's push-pull."


To me that says the guy is running a 2 pill amp.
 
" It's usually the case though that "that guy's amp" has a transistor pair, which to me says it's push-pull."

So what configuration would an amp with a single output transistor be considered as?
 
After reading your answer and thinking about it some more, I think I realize why I was initially confused about the use of class C biasing with push-pull. I think the drawing of the class C area of no conduction in the graphic included with the article is a little misleading. I understand what it's trying to convey, but I don't think it accurately reflects what you'd see if you were to observe the output of a single-ended class C amp on a scope, which is what I was thinking of when was looking at it.

The class C drawing is basically the inverse of the class AB drawing, and I can understand why: with class AB, the transistor is turned on for more than half the time, while with class C the transistor is turned on for less than half the time. However the drawing does not show the 'area under the curve,' as it were: it makes it seem as though there's just that bump sort of hanging there with nothing underneath it. Conceptually that make sense, but I think on a scope it would actually look like this (apologies for the blurriness):

classc.jpg


(In a push-pull configuration, instead of the second half of the waveform being completely blacked out, you'd have another negative bump shaped just like the positive one.)

Basically, the transistor switches on as soon as the input reaches a certain level, at which point the output voltage shoots up suddenly in a straight line, rather than rising an falling gradually in sync with the input like it would do with class B biasing. And if I understand correctly, the resulting square waveform shape is what creates additional distortion products.

Mind you, what we're talking about here is the unfiltered output. I'm assuming that filtering the distortion products from the output also has the effect of smoothing out the waveform to restore its sinusoidal shape, so the filtered output signal would look much more like the input signal.

Of course, CB amplifiers almost never have any output filtering. :(

-Bill
 
I would agree with your thoughts concerning class C sine waveforms. The filters do filter out distortion, resulting in a cleaner waveform. There will still be more distortion than other classes, but simply filtering the output helps a bunch in that department.
 
Depends on the configuration. You might have push-pull, or you might have 2 single-ended stages in parallel. There are differences in configuration between the two and in their respective spectral outputs, and a good description of parallel stages can be found in section 3.4.2 on page 31 in this paper: http://www.ieee.li/pdf/essay/rf_power_amplifier_fundamentals.pdf. I'll quote here (minus the figures):

3.4.2 Parallel connection
3.4.2.1 VHF AND UHF RANGES

Sometimes, in base stations for mobile radio, output
powers higher than those that can be delivered by a single
transistor are required. The simplest solution is then to
connect two transistors in parallel. However, because the
transistor impedances can be very low, it is not
recommended that this be done directly (i.e. by connecting
gate-to-gate and drain-to-drain). A better method is shown
in Fig.3-48.
Adjacent the transistors, separate matching sections are
used; elsewhere, the sections are common to both
transistors.
Resistors R1 and R2 are included to prevent push-pull
oscillations, and perform the function of a hybrid coupler.
Although this circuit does not fully isolate the transistors, it
does prevent oscillations. For the best performance,
R1 should be twice the equivalent parallel input resistance
and R2 twice the load resistance of one transistor.

In Fig.3-48, only the RF components are shown.
DC components such as RF chokes, coupling and
decoupling capacitors still have to be added.

3.4.2.2 HF RANGE
Another form of parallel connection is often used in
high-power amplifiers for the HF range. An extensive
description of such a system is given in application report
AN98032”. The combining transformer (transmission line
type) used is also described in report number “ECO6907”.

As shown in section 3.4, a main difference is in the phase differences, and there is also hybrid coupling as an option:

Most RF power amplifiers are ‘single-ended’, i.e. they
have one power transistor. Sometimes, however, there are
good reasons for using amplifiers having two or more
transistors, as for example described in Section 3.2.2.5.2.
In such cases, there are several ways to interconnect the
transistors, the most popular employing:
– Hybrid couplers (90° phase difference)
– Parallel connection (0° phase difference), or
– Push-pull or balanced connection (180° phase
difference).
 
Last edited:
Just realized I had a couple amps here that show what we're discussing so I took a couple pics and labeled some key parts:

This is a 136-174 MHz class C FM-only commercial/public safety amplifier that is frequently found on the ham bands. It uses 3 single-ended stages:
TPLPA3-1FE-4WEB.jpg


Here's a class C UHF amplifier for the same applications. It's generally useable from ~400-480MHz

HenryC40D02WEB-1.jpg


This amplifier has a 9-pole Chebyshev filter, but at UHF it only consists of 5 capacitors and 4 bent pieces of wire for inductors. It's mounted vertically on the tall standoffs on the left-hand side of the amp.
 
Last edited:

dxChat
Help Users
  • No one is chatting at the moment.
  • dxBot:
    Tucker442 has left the room.
  • @ BJ radionut:
    LIVE 10:00 AM EST :cool:
  • @ Charles Edwards:
    I'm looking for factory settings 1 through 59 for a AT 5555 n2 or AT500 M2 I only wrote down half the values feel like a idiot I need help will be appreciated