LPF after a final amplifier is it always needed?

[neazoi]

Hello,
Assuming an oscillator that has it's harmonics filtered and that is followed by a preamplifier/power amplifier stage.
Should an LPF always present after the amplifier?
If one does not send harmonic frequencies to the amplifier, these should never get out of it, at least not in considerable power levels if talking about 5-10w final amplifier. Intermodulation signals from a linear could impose powerful unwanted signals after it so that an LPF is needed indeed?

 

[tony_lth]

If you output 5W, then LPF after final PA is needed.

 

[drkirkby]

Hello,
Assuming an oscillator that has it's harmonics filtered and that is followed by a preamplifier/power amplifier stage.
Should an LPF always present after the amplifier?
If one does not send harmonic frequencies to the amplifier, these should never get out of it, at least not in considerable power levels if talking about 5-10w final amplifier. Intermodulation signals from a linear could impose powerful unwanted signals after it so that an LPF is needed indeed?

A purely linear amplifier would generate no harmonics and produce no IM products. In practice, unless is is highly linear, the harmonics might well be an issue. But it depends on the application. Personally I'd always fit some form of low-pass filter.

For most systems, the IM products will fall close to the carrier, and so can't be fitered with a low-pass filter. That's a very differernt issue. You must consider IM distorsion and harmonics as different issues, and not expect to solve them the same way.

Dave

 

[biff44]

Since there is no such thing as a truly linear amplifier....the question you should instead ask is "what level of harmonics is allowed". Then you drive your amplifier with the oscillator, and measure the harmonic levels, and add the LPF IF it is needed to meet your specification.

 

[FvM]

If one does not send harmonic frequencies to the amplifier, these should never get out of it, at least not in considerable power levels if talking about 5-10w final amplifier.

If you are targetting to harmonic levels required by RF regulations, this sounds like a grandiose illusion. Even PAs with mW output power often need harmonic filtering.

 

[chuckey]

The distortions that affect you are mainly the third order because the distortion products fall within your transmitting channel. However second order products all fall around 2 X your transmitting frequency, so they will not effect you but will effect other users of the air waves. So to mop them up, a low pass or band pass filter is often required.
Frank

 

[DeboraHarry]

If one does not send harmonic frequencies to the amplifier, these should never get out of it, at least not in considerable power levels if talking about 5-10w final amplifier. Intermodulation signals from a linear could impose powerful unwanted signals after it so that an LPF is needed indeed?

If the amplifier gets no harmonics (an impossibility in practice), then it will still generate harmonics.

I suggest you add a filter, but work out what attenuation you require. Then add some more attenuation, as the source and load impedances will probably not be 50 Ohms at the harmonic frequencies, so the filter may perform worst than you think.

I recall once transmitting on the 7 MHz amateur band, and picking my frequency precisely, I was able to lift the squelch on a 432 MHz amateur band receiver. That would have been around the 90^th harmonic!

 

[neazoi]

Thank you all!
What if a class-A amplifier is used? (rarely used on hf) Does this change the situation at all or the same thing applies?

 

[FvM]

Not quite right. There's a lot of class A PA applications requiring high linearity, e.g. GSM. They will still need harmonic filtering.

 

[DeboraHarry]

In theory, if there were no harmonics fed to a class A amp, there would be no haromics on the output. But in practice nobody has made the perfect amplifier yet. So in practice, unless you are using this for something like RF heating, you probably need a filter.

- - - Updated - - -

Not quite right. There's a lot of class A PA applications requiring high linearity, e.g. GSM. They will still need harmonic filtering.

With GSM you even need to worry about the intermodulation products created by the antenna, as you don't want multiple carriers creating 3rd or 4th order product that are received.