class C linear amplifier

[kb3ouk]

First off, I know this does work for AM, and AM only. But what I can't figure out is if there is if the layout has any effect on how it works. Here are the 4 major amp layouts I could think of:

Single ended grid driven
Parallel grid driven
Push-pull grid driven
Grounded grid (single ended or parallel)

I already know that a GG amp biased to class C will work as a linear for AM, but will a class C grid drven amp work too? I also know that the reason a class C amp can be driven with a modulated carrier and amplify it is because the carrier acts to reduce the bias voltage, so the tube is only actually in class C when it isn't being driven. When the carrier is applied, it reduces the bias voltage, which causes the tube to shift into operating in the linear portion of it's curve.

[W4NEQ]

Class C amplifiers cannot be used to linearly amplify AM, but can be themselves modulated by injecting audio into cathode, grid, screen, or plate circuits.

Chris

[K1JJ]

I also know that the reason a class C amp can be driven with a modulated carrier and amplify it is because the carrier acts to reduce the bias voltage, so the tube is only actually in class C when it isn't being driven. When the carrier is applied, it reduces the bias voltage, which causes the tube to shift into operating in the linear portion of it's curve.

Yes, from my own experiments, I believe you have it correct.  Any amplifier, whether it be grid driven or grounded grid is not linear when biased into class C. Put a ssb signal through and part of the signal will be missing when below cutoff. However, it is a different thing when an AM carrier is used to drive the amplifer. The carrier acts to shift the operating point back above cutoff and it becomes linear again, but now operating in the AB class.  No free lunches. The carrier acts like a substitute "positive" bias.

I wouldn't have noticed this unless I did it myself by experiment some years ago. I had a 4X1 GG linear that was biased by a string of diodes in the cathode. I could switch in any bias level from class A to class C.  When in class C, ssb was distorted, of course. But an AM carrier brought the idling current up to near class A and the audio was clean when modulated.

I THOUGHT the AM efficiency looked better compared to normal bias in class B just by looking at the plate glow, but not sure about that.  Not sure theoretically why the overall linear efficiency wud change, if at all, just becuz a carrier is used for bias vs: standard DC bias. Seems like the same thing in the end. But it sure feels cool to bias the amp into class C and then drive it with linear AM...

There is a limit, whereas if the bias is really large so that the tube is biased close to class D, it just takes that much more AM carrier RF drive to bring the amplifier up to a normal carrier output to put it into the linear curve again.

T  

[kb3ouk]

http://amfone.net/Amforum/index.php?topic=16488.msg113118#msg113118
http://amfone.net/Amforum/index.php?topic=29516.msg229169#msg229169

And this article:
amfone.net/Amforum/index.php?action=dlattach;topic=30787.0;attach=31340
That article discusses class BC linear amplifiers, which are pretty much what I am asking about (biased beyond class B but not deep into class C). Take note of figures 1 and 2. If you look closely, you can see how it works. The bias is set for class C. If you look how the input and output carriers are plotted against the curve, you will notice that when the input carrier is at rest, the output carrier falls in the middle of the curve. The modulation then is kept inside the linear portion of the curve. As long as the 100% positive and negative peaks are within the linear portion, a class C amp can act as a linear for AM ONLY.

Edit: thanks for chiming in Tom.

[IN3IEX]

This is a modern "envelope replicator" class B - C linear amplifier
 
http://www.813am.qsl.br/artigos/py2ko/burrico/quaggi.pdf

Il looks very interesting. Too complex for most of us.

Giorgio

PS: by comparison this is a very compact grounded grid linear amplifier. Class B, less power.

http://www.ing.unitn.it/~fontana/liexnear.pdf

[kb3ouk]

That first amp is basically a tube version of an EER amp, which are usually class E SS amp made linear.
The second amp could be biased into class C and still work for AM. I had a thought last night that might explain the increases efficiency with a class C linear. The thing is, the peaks are probably more effiecient with class C than Class AB. Class AB is about 66% efficient on peaks, where class C is probably about 75% efficient. Of course the carrier is still gonna be low to allow for headroom, but even that should be slightly higher with class C than class AB.

[WQ9E]

In the early SSB handbooks there are discussions of using standard AM/CW transmitters as a linear amplifier for SSB exciters.   One details the use of the Viking 1 by feeding the SSB signal into the VFO input with the transmitter set for CW mode and reducing drive to the point that no grid current is observed..  The author notes that this will not work on bands where the VFO signal is multiplied.

[W3NE]

I used my homemade 20A knockoff to drive a Viking II straight through on 75 & 40 as described by Rodger. I seem to remember putting a 470-ohm resistor in the key jack to ground per reccomendation in QST. Worked very well as I recall.

Bob - NE

[W1RKW]

Doesn't the tank circuit play a key role in class C plate modulation.

[kb3ouk]

Doesn't the tank circuit play a key role in class C plate modulation.

Yes, in class C the tube really only amplifies the signal in pulses, the rest of the signal is formed in the tank circuit due to the flywheel effect.

[kb3ouk]

 I sure did open a can of worms when I tried to explain this on the air tonight, that it is possible to use a class C biased amplifier to linearly amplify AM. An interesting point came up: sure you might get a result that sounds like it is linear, but is there any garbage that being generated from using class C bias that wouldn't be there if the amp was biased to class AB?

[W2VW]

I sure did open a can of worms when I tried to explain this on the air tonight, that it is possible to use a class C biased amplifier to linearly amplify AM. An interesting point came up: sure you might get a result that sounds like it is linear, but is there any garbage that being generated from using class C bias that wouldn't be there if the amp was biased to class AB?

I heard some of that.

The amplifier is not class C while being driven.

Some of those guys stopped listening as soon as they heard class C.

If the thing were biased hard into class C it would not work the way you desire.

[kb3ouk]

Yea, I tried explaining that to them that as soon as the thing gets drive, the carrier throws it out of class C and in to class AB, and they just didn't seem to follow. And I did tell them that it can't be biased too deeply into class C, since the more bias you have means the more carrier you need to overcome the bias, until you get to where you can't add any more carrier to pull the bias up.

[W2VW]

It does work. I used to bias my Gnational NCL-2000 to give less idling current than I could get away with on ssb. Based on a couple years usage and plenty of on air reports from intelligent folk with panadaptors, I'd give it the green light.

Of course the concept can't be taken too far....

[W3RSW]

So why Hasn't this mode been the commercial standard all these years?
Relative efficiencies?  The power bill being the main factor?  Anyone actually superimpose this set of changing classes on tube curves, plotted differential biasing on real load lines, etc.?
Hmmm?

[kb3ouk]

Does something like this help?

[K1JJ]

Yes, it does indeed work and is clean.  But it does fly in the face of conventional ham wisdom and beliefs until we specify what really is a class C amplifer.

The question still is, are we gaining anything or just trading off use of a carrier for less bias?   As I mentioned earlier, I never ran any precise tests to determine if the masquerading "class C" amplifier was running any more efficient when the carrier was providing the shift in position on the transfer curve  -  or when it was being shifted by real applied bias.  Is it technically still in class C?  Probably not.

It wud be an easy experiment to do.   The initial question is,  "Can we use a standard class A, AB or B linear on AM when biased into class C?"  The answer is yes.   But we must be sure this class C amplifier does not have grid leak or sloppy screen regulation, a screen choke, etc., as with most plate modulated finals, or this will cause distortion all its own.

This technique only applies to a normal ssb linear amplifer that has been biased into class C. The jury is still out whether it is more efficient to run it this way.    A simple test is to add a string of diodes in the cathode to a standard grounded grid pair of 3-500Z's in linear to bias it into class C. Then run AM and note the differences.

T

[kb3ouk]

If you had a RF deck that was driven by an exciter and used some form of high level modulation, an interesting experiment to try would be to turn off the high level modulator and then modulate the exciter, without making any changes to the RF deck. Something else that kinda goes against conventional ham wisdom is having a modulated amplifier where the driver is modulated too. The same concept applies here, we have a class C final that is being driven with a modulated signal. This idea came up at one point last night too, where the input to a final was modulated, but the final was also getting some kind of high level modulation at the same time. I actually know of one example of a commercial transmitter that used this, the Harris VP-50. The purpose of modulating the driver of a high level modulated class C amplifier is to improve the linearity, which is the complete opposite of what was being told to me last night, which was that "class C" and "linear" don't belong in the same sentence.

[K1JJ]

If you had a RF deck that was driven by an exciter and used some form of high level modulation, an interesting experiment to try would be to turn off the high level modulator and then modulate the exciter, without making any changes to the RF deck.

In this example, severe distortion would be generated by the Heising choke, mod transformer, grid leak resistor (unregulated), screen choke and likely unregulated screen supply. If all of these parts were eliminated and the final simply had a higher class C bias (and regulated), then the AM carrier would cause the amplifier to work linearly due to the carrier generated bias point as we discussed.

which was that "class C" and "linear" don't belong in the same sentence.

A <180 degree plate modulated class C final's output actually becomes a linear signal when plate modulated and run with the proper class C grid and screen currents and plate operating point.  If it didn't, then there would be severe distortion.  Same way a mixer works in a receiver.  So, the resultant class C final's output actually should be linear...    (linear output when compared to the modulated audio input signal)

I think we are all talking the same thing - just different ways of thinking about it.

T

[kb3ouk]

If you had a RF deck that was driven by an exciter and used some form of high level modulation, an interesting experiment to try would be to turn off the high level modulator and then modulate the exciter, without making any changes to the RF deck.

In this example, severe distortion would be generated by the Heising choke, mod transformer, grid leak resistor (unregulated), screen choke and likely unregulated screen supply. If all of these parts were eliminated and the final simply had a higher class C bias (and regulated), then the AM carrier would cause the amplifier to work linearly due to the carrier generated bias point as we discussed.

Ok, that makes sense. It would make more sense to try it with a triode final , then you could also run fixed bias (not that you couldn't do that with a tetrode or pentode) and have so way of switching the mod iron out of the circuit, which might be worth trying.

[ke7trp]

If I am understanding this correctly, we are talking about driving a grid driven Class C amp with an exciter.

I tried this a few moments ago with my ranger/ Thunderbolt combo.  I turned HV off on tbolt, Put Tbolt into Class C mode.  Still used Swamped input to Grids.  Loaded up ranger for 5 MA of Drive on Tbolt meter.  When I modulate ranger, Tbolt Grid meter goes to 15 MA.

I turned on High voltage and keyed up.  Loading of Tbolt changed slightly. I had to rotate the loading knob a bit. 

The result, Super clear audio, Less Plate current at carrier, MORE pep output. Seemed to have gained some efficiency?  Less heat blasting out of Tbolt.

Got great reports on air.

Hmm.. Class C linear.   Sounds like most cheap 11 meter amplifiers.  Transistors with no bias supply at all for AM only use.

C

[K1JJ]

If I am understanding this correctly, we are talking about driving a grid driven Class C amp with an exciter.

I tried this a few moments ago with my ranger/ Thunderbolt combo.  I turned HV off on tbolt, Put Tbolt into Class C mode.  Still used Swamped input to Grids.  Loaded up ranger for 5 MA of Drive on Tbolt meter.  When I modulate ranger, Tbolt Grid meter goes to 15 MA.

I turned on High voltage and keyed up.  Loading of Tbolt changed slightly. I had to rotate the loading knob a bit.  

The result, Super clear audio, Less Plate current at carrier, MORE pep output. Seemed to have gained some efficiency?  Less heat blasting out of Tbolt.

Got great reports on air.

Hmm.. Class C linear.   Sounds like most cheap 11 meter amplifiers.  Transistors with no bias supply at all for AM only use.

C

Yep, you have it exactly right, Clark.

 It could be done with a triode, GG linear too.

I thought I saw less heat too, when I tried it with my 4X1 linear amp.  I have been told by others the same thing.

Try this if you can....  Load the amplifier up in normal linear fashion (normal bias) and set the output power at 200 watts AM carrier.  Put a thermometer over the final air flow and see what it reads after five minutes of constant carrier.

Then let the amp cool down and repeat the same test again with the class C bias setting, same power output, same duration.  

The second test could be to see how the efficiency changes under modulation.  Put thru a 1kc audio tone at 70% modulation and run the same test for 3 minutes each.

I'd be curious what you find.


Just to repeat for readers,  on ssb, this class C linear technique would create crossover distortion and bad splatter. It works on AM only. (with a constant carrier)

T

[ke7trp]

I will do that tonight Tom.  I am home sick with the flu and need something to do anyways  

I can vary the "class" with the ranger ANT Loading knob.  I played around with it. I can get 250 watts carrier out with less current and an upward swing on plate in Class C and more pep.

The guys said the rig was pounding into CA and way up North to Kingman AZ on 40 meters.  They said the radio ocupied the same Bandwidth and there was no splatter.

I will report back on heat and plate current tests.

C

[ke7trp]

Ok.  Here is the data Tom.

Class C mode:
250 watts carrier on Bird43
300MA plate current


104F output at rest.  This was stable.
After 5 minutes 140F. 
It hit 140 and stayed there at 3 min point

Class AB (linear mode)
250 watts carrier on Bird43
400MA plate current
After 5 minutes air temp was 170F.  Rising steady... Was going up when I let off.

In Class C,  The 4-400s have very little color.  Slight red spot the size of a dime or less.

In Class Ab,  The 4-400s are Red from top to bottom and white/yellow in middle

A HUGE decrease in Temp and current.  In Class C, the plate current swings UPward.  My Friend called to say that the Class C mode is Crystal clear, loud and is Clean on the air.

[KM1H]

Hmmm, so I should add another bias zener in series with the existing one in the LK-500ZC and add a switch to bypass it for SSB?

Probably means a bit more CW efficiency also.  Gotta give it a try and if that works do it to the Alpha 76PA which has 1200W of PD available.

There is also a NCL-2000 that sometimes gets used on AM with the HT-32B and that could sure use less heat.

It also means that guys with the SB-200 and 220's can get some more oomph without melting tubes and transformers