Solid State Amps, Classes, Cleanliness..

[eagle1911]

OK, so I've been doing a bit of research into RF amplifier design, and I've got a few questions for the experts here. First off, let me apologize in advance for any n00bish questions I might ask, as I am fairly new to the theory behind amplification.


I have several HF/10/12m amps, mostly of the two transistor variety. I've been hearing for years about how class C amps just won't cut it on SSB.. Only one of my amps was class AB biased from the factory. I've added a rudimentary bias to another that appeared to be class C from the factory (KL-300p). I also have an unmodified KL-300p exactly like the one I just described that I added bias to. I run low pass filters on the output to minimize harmonics.

My question is this: Why is it that I get good signal reports consistently on ALL of my amps, regardless of mode, biasing, class or any other feature that I can tell? I do run my amps based on the specs for the transistors, rather than amp manufacturer specs which are almost always WAY overrated, so if I have 2 sd1446s I set things up so that the amp swings to 140-160W, if I have mrf455s I set it to run 120-140W etc.. Is this why the class/biasing doesn't seem to affect the output? It just seems odd to me all the down-talking I hear about class C amps on AM and SSB, but I've never heard so much as a peep about dirty output/splatter from the receiving end of any QSO.. Even the two KL-300p's (which ground the bases of the transistors via the input transformer) get nearly identical *excellent* reports, even though one is crudely biased AB..

Your thoughts are appreciated!

 

[dxhound]

Class C is described as any device that amplifes less than 180 degerees of an original 360wave form. In this case, the waveform is your voice.

Class AB is some variation of amplification between 180 degrees and 360 degrees of the original waveform. AB1 is closer to 180 deg. vs AB2 being closer to 360 deg(otherwise known as class A)

The reason class C is not the greatest on SSB is due to 2 things primarily.

1. With less than 180 deg of the input being amplified, that means in all reality less than half of your voice is being amplified. Depending how deep into class C the amp operates in, you can some times hear their words sounding "fuzzy" or broken up due to this fact.

2. Class C by nature creates more distortion than ab1/2 because less of the original signal is being amplified and the remaining >180 is noise, or whats known as null.

Electronic amplifier - Wikipedia, the free encyclopedia

This wiki article explains it pretty well, go toward the bottom and there is where you will see the diff. class' of amplifiers.

And for the most part, The reason you get good reports using these amplifiers, is because the avg cb operator doesnt know what a good signal should sound like. He just knows that he hears you fine and that your loud.

All that said, you can use (i dont recomend it) a class C amplifier on SSB. But you will cause interference, regardless of how small that amount may b and weather its in or out of band. If both amps (ab or c) are being driven corectly, AB will always produce less distortion and harmonics just given their design.

P.S. The Low pass filter is a great Idea. however it only attenuates the harmonics above 30mhz(typically they are tuned to there). That filter does nothing for in band harmonics.

 

[eagle1911]

Thank you for that response!

I'm not a CB operator, only ham band usage for me. My call is WB0LSR.. I use my amps with an HTX-10 or a Yaesu FT-817.

I am aware of the basics behind the operation of the various classes of amps, but your statement about how deep into the class the amp operates intrigues me.. Is it possible that an amp designed to be class C will operate in other regions depending on some input factor?

Regarding the term "in-band harmonics", I'm not sure what those are. By my limited knowledge, a harmonic is a multiple of a fundamental frequency. It would seem to me that once you've multiplied a fundamental that is within a given band, that harmonic would no longer be in that same band.

I'm currently building a set of band-pass filters for the output, which will hopefully go a long ways toward eliminating much potential harmonic interference, since I do know that low-pass filters can still allow for some unwanted frequencies to pass.

 

[dudmuck]

The problem of class-C amp is that of pills shutting down with low RF drive on input.
Because with class-C its RF only which turns on the transistors. With enough RF drive, the transformers (and capacitor) on input and output make up a tank circuit which has inertia that stores energy for the zero-crossing part of the cycle. When RF drive is low, that inertia becomes too small to maintain over the whole cycle.

But this has alot to do with the input drive required of the type of pill in use. The 2sc2879 requires more RF input watts than an SD1446.

In other words: A pair of sd1446 takes low-drive, and it could be that your SSB transmitter puts out enough power even with no audio into microphone to keep the SD1446 pair turned on thru the whole RF cycle.

 

[eagle1911]

That falls in line what what I know about class C amps. You may be right about my radios being able to maintain output even while I'm not talking, but you can hear the relay clicking off during pauses in speech so the transistors must be shutting down during that time if it's a class C amp.. Maybe it's just that I was feeling like I wasn't really paying a penalty in performance and output quality using the amp that I'm sure is class C, and hearing so much about how class C is bad for certain modes I felt I must be overlooking something. Thanks for the replies so far..

 

[dxhound]

Thank you for that response!

I'm not a CB operator, only ham band usage for me. My call is WB0LSR.. I use my amps with an HTX-10 or a Yaesu FT-817.

I am aware of the basics behind the operation of the various classes of amps, but your statement about how deep into the class the amp operates intrigues me.. Is it possible that an amp designed to be class C will operate in other regions depending on some input factor?

Regarding the term "in-band harmonics", I'm not sure what those are. By my limited knowledge, a harmonic is a multiple of a fundamental frequency. It would seem to me that once you've multiplied a fundamental that is within a given band, that harmonic would no longer be in that same band.

I'm currently building a set of band-pass filters for the output, which will hopefully go a long ways toward eliminating much potential harmonic interference, since I do know that low-pass filters can still allow for some unwanted frequencies to pass.

Class C is a pretty broad region, with conduction ranging from 90deg to 180deg. Depending on design its possible the amps bias may vary enough to skirt into other classes. Texas star amplifiers are known for being class AB1. But when they get hot, and because their bias is not thermally tracked(correctly) They drop into class C.

You could potentialy play with the bias a little bit and get it close or even into class AB1. thats the fun of this hobby, experimentation

I really used the wrong term above about "in band harmonics" Its actually refered to as In band IMD (inter-modulation distortion) . You are correct about the harmonic being a multiple of the fundamental freq. But you need to remember that your audio has a bandwidth and freq of its own. That can cause intereferce within a few khz of your operating freq. All transmitters exibit some of this but usually so low that it does not affect the avg rxer, However when the radio is improperly tuned, and then put into a "dirty" amp, it can become VERY pronounced!

What style filters are you building?

 

[bob85]

stock kl300's sound class C on ssb, adding bias removes the switching distortion making them sound ok,
for some reason unknown to me some folk don't seem to be able to hear it while to others class C is unacceptable for ssb.

 

[Beetle]

Rather than using signal reports from other stations (where the operators, as someone else correctly said, may not really know what a good, clean signal sounds like), try looking at your signal with a spectrum analyzer. That's the best way to tell IF you're splattering, how badly, and on what frequencies.

Don't try using a plain oscilloscope; you have to look in the frequency domain.

 

[eagle1911]

You could potentialy play with the bias a little bit and get it close or even into class AB1. thats the fun of this hobby, experimentation

I really used the wrong term above about "in band harmonics" Its actually refered to as In band IMD (inter-modulation distortion) . You are correct about the harmonic being a multiple of the fundamental freq. But you need to remember that your audio has a bandwidth and freq of its own. That can cause intereferce within a few khz of your operating freq. All transmitters exibit some of this but usually so low that it does not affect the avg rxer, However when the radio is improperly tuned, and then put into a "dirty" amp, it can become VERY pronounced!

What style filters are you building?

I've purchased a set from Filters, 80-10m. They're kits so I'm having to build them, but they're very simple so it's no trouble really. They are actually low-pass filters, but they're each tuned to a different band.

Regarding the experimentation, that's what drew me to Ham Radio years ago, and I know that CBers do their fair share as well.. I've only recently decided to really figure RF PAs out.. eventually I'd like to build one from scratch, solid state, but I figured a good way to start out learning would be to buy a few cheap amps, since the core push-pull configuration isn't overly much different between any of them. There's a lot of room for hacking in improvements and just messing around trying to figure them out. This way if I blow one up, it neither cost me time building it or my own effort.. and minimal money. I've already disconnected the input transformer from ground on one of the kl-300s, and I've added a simple diode clamp to provide rudimentary bias. The diode is stuck to the top of one transistor with Arctic Silver.

@Beetle: Thanks for that input, that makes sense. Unfortunately I don't have a spectrum analyzer at my disposal.. I've got a scope and a few other types of test equipment that I've been making do with.