Class C Power oscillator design

Dear All
Hi
Thank you for your time and attention.
I want design a high power class c oscillator , but i don't know any thing about it . so what should i do ?
Is it possible that you guide me please?
Thanks in advance
Goldsmith

Are you building a Hi power transmitter using the Class 'C' ? If so ,what frequency ,power level,modulations used ,and conduction /radiation antenna needs be specified/designed initially.

Hello,

As vimalkhanna says, what frequency, power, supply voltage, etc?

Building a real class C oscillator can be very difficult to almost impossible. Real class C (with conduction angle < 120 degrees) is sensitive to envelope instability.

Because of the harmonic-rich content of the current waveform, getting the correct termination for the harmonics may be difficult in low voltage, high power applications. Vacuum tube implementation is easier to operate in real class C operation at HF.

Is "class C" a real requirement, or are "class E" circuits (for example) permitted also?

Dear vimalkhanna and WimRFP
Hi
Thank you for your replies.
I want to learn designing ways of high power self exciting oscillators ( I Read some articles about class B and C And D and E and F oscillators , but i just can design class D oscillators as well as . and now , i want learn about another classes oscillators such as C and B and F and E )
So what should i do for learning the design steps of them ?
for example , consider , please , that i want design a 80w 3MHZ self exciting oscillator . what should i do ? Thank you
Best Regards
Goldsmith

Hello Goldsmith,

80W at 3 MHz is certainly possible with cheap (switch mode) semiconductors. The capacitors will probably be the most expensive components.

In case of low supply voltage, push-pull circuits are attractive (tuned Royer or Baxandall oscillator all with center tapped transformer or resonant circuit). The tuning capacitor avoids the need for a transformer with low leakage inductance. Even a center tapped LC circuit without magnetic core does the job. Drain voltage is about 3.3*supply voltage. Both semiconductors are at same potential (ground).

You need some Q factor in the resonant circuit. At 3 MHz you need good capacitors. The problem in this type of oscillator is that you need good 50% duty cycle, otherwise the efficiency drops and/or you get smoke. So you need a form of bias setting feedback that maintains 50% duty cycle. With BJT the BC junction can be used for that, but with mosfet/IGBT this doesn't work. Advantage of the resonant push-pull oscillators is that these oscillators are not load critical (they even run without any load).

When the tuning capacitor is split in two capacitors (each capacitor goes from drain to ground), you will get a balanced class E oscillator. Main disadvantage is that even under no-load situation, all oscillating current goes through the mosfet and the body diode, of course this is no problem when the oscillator runs with nominal load only. Note that not all mosfets have fast enough body diodes for operation at 3 MHz. Main advantage is that the duty cycle isn't very critical.

When running from high supply voltage (where silicon mosfet properties are not that good anymore), the half or full bridge oscillators are best, as the drain voltage equals the supply voltage. Also here you can have zero voltage/current switching reducing the switching loss to a negligible value (with respect to conduction loss). There exists a so-called current fed half bridge oscillator circuit, but I have no experience with it.

Older CFL or TL use (for example) a voltage fed half or full bridge oscillator directly fed from rectified/filtered AC mains. Most of them are (quasi) resonant to reduce switching loss and setting the frequency. Most modern types use additional control now (dedicated chip or microcontroller).

For all oscillators: you should evaluate the turn-on behavior and behavior under non-optimal load. Several oscillators require a certain minimum load, or can't accept too heavy load. Depending on the bias circuit, you may need a starting pulse (in older electronic ballasts frequenly generated with a diac).

Many oscillators will not work well when followed by a filter (for example to suppress harmonics). It may oscillate on some other (parasitic) circuit resonance.

If you need high efficiency in the low MHz range from low supply voltage (using mosfets), I would not go for a true class C approach.

Dear WimRFP
Thank you very much for your reply . can you suggest me a circuit for this aim , or a way for design , please?
Appreciate
Goldsmith

Hello Goldsmith,

I think the best you can do is try to understand how the various DC to AC topologies function. Also check how they behave under your load (inclusive load variations if present).

Application notes may be helpful in mastering DC to AC conversion. "Solid state radio engineering" also has a chapter on efficient RF power amplifiers.

Next thing may be figuring out what components are available at your site (inclusive power supply and passive components). Based on this you can select the best topologies given your components and required power.

After this you need to dive into the feedback network to turn the DC to AC converter stage into a self-oscillating (and self-starting) converter. When the output stage has sufficient frequency selectivity, the output stage may determine the frequency (as is frequently done in resonant converters). In case of non-resonant topology, the feedback network should determine the frequency.

A non-linear simulation tool (for example spice based) will save you lots of time, frustration and smoking components. You may very likely run into some challenges, and then the forum can be helpful.

Dear WimRFP
Hi
Thank you very much for your answer and help.

Application notes may be helpful in mastering DC to AC conversion. "Solid state radio engineering" also has a chapter on efficient RF power amplifiers.

Click to expand...

I have no problem with RF amplifiers . those are my best friends!

Next thing may be figuring out what components are available at your site (inclusive power supply and passive components). Based on this you can select the best topologies given your components and required power.

Click to expand...

I have not problem with inductors because , i wound all of my inductors until now .
And about capacitors . i can find them in local markets however hardly . and about transistors and diodes , they are available in local markets of my country at high frequencies.

A non-linear simulation tool (for example spice based) will save you lots of time, frustration and smoking components. You may very likely run into some challenges, and then the forum can be helpful.

Click to expand...

I have simulated all of my circuits with pspice and LT spice ( i trying to learn it currently ) until now ( circuits such as class A oscillators at frequencies up to 100MHZ and some simple transmitters ).


So , consider , please , that you want design an oscillator with specifications that i have mentioned at top . what you'll do ? i'll do it too .
Appreciate and Regards
Goldsmith

Hi Dear All again
I have researched for this oscillator , but unfortunately until now , i couldn't find my answer . so is it possible any body , that can help here , help me , please ?
Regards
Goldsmith

The class C oscillators are using only the 15% of the incoming drive pulse(sine)as a trigger sync. and the parallel tuned circuit oscillates to full sinewave at heavy power .Thus the overall efficiency is very very high nearing to >85%-90% ..
However , the modulation (hi level) needs to be done seperately as the FM /PM /AM purity does not exist from drive end .
These transmitter are /were manufactured by marconi /Nippon etc.for radio transmission.and creation of 10KW to 100KW power levels .
Such units are noisy in generation and have difficulty in Transciever designs for Military applications ...

Dear vimalkhanna
Hi
Thank you for your reply.
How can i design one of these oscillators ( at class C ) ? and how will be the arrangement of this circuit ?
Thanks In advance
Goldsmith

This article discusses the oscillator in a popular transceiver unit. It is specifically referred to as being class C.

**broken link removed**

Hello Goldsmith,

Though in Dutch language, It may be helpful as the text in de images is in English **broken link removed**

It also shows the topology for a class C oscillator with mosfet and automatic bias (as you don't have grid current or BC diode in a mosfet).

When you overdrive a class C amplifier, have a reverse diode present, and change the tuning, you may get a class E stage.

http://www.tetech.nl/divers/classEoscHV.pdf shows an example of a class E oscillator and some theory of operation (Dutch, but the graphs may be helpful).