[TowerTalk] Are All Low-Pass Filters Alike? -

Tom Rauch w8ji at contesting.com
Wed Aug 17 01:04:41 EDT 2005


> Those spurious frequencies are, correctly as stated, not
filtered by a
> LP between the Xcvr and amplifier.
> The original question I raised was concerning HARMONICS,
not spurious
> frequencies generated by the amplifier.

That's what I addressed Don.

IM or distortion products caused by amplifying signals with
changing amplitude can't be filtered with any lowpass
filter. That's because they all fall fairly close to the
desired signal.

First, harmonics are NOT amplified by the PA Don. The only
exception would be if the PA is very poorly designed.

Second, virtually all modern exciters are very clean for
harmonics. Even if the exciter had harmonics, a PA of good
design would have very negative gain on the harmonic
frequencies.

> Perhaps operating an amplifier ONLY within the linear
portion of its
> curve would cure most of the spurious frequencies and
limit the need for
> a LP filter post-amp.

No, not at all. The type of linearity you are referring to
is a transfer function of the input power to the output
power. It is gain linearity, and mostly affects in-band
distortion like keyclicks and splatter.

Harmonics have nothing to do with transfer function or gain
linearity as the input signal level varies. Harmonics are
related to conduction angle of the amplifying devices. They
are generated in the output device because it conducts less
than a full RF cycle.

Placing a lowpass filter on the input can't reduce harmonics
at the output of the PA unless the PA has some very major
design issues.

> As for the statement that 'amps do not amplify harmonics'
that was
> posted by another submitter .. I defer to the "experts" as
to why
> harmonics ARE amplified by a well-designed amplifier .. a
FACT that  can
> be easily measured.

I measure a lot of PA's, and I can assure you any
type-accepted PA or PA of adequate design will have a very
large value of *negative* gain (attenuation) when driven at
harmonics of the input frequency. The PA will reduce level
of harmonics passed through from an exciter, not increase
them.

You can test this by driving your amplifier with a 20 meter
signal while it is tuned to 40 meters. You will see with a
few watts of 20 meter drive you'll have virtually no power
output when the amp is on 40 meters. Typically the loss is
25 or more dB at the second harmonic, and even more loss at
higher harmonics.

I think what is confusing you is the harmonic levels are
often higher after a high power PA than from the exciter
alone. That is not because the PA amplifies harmonics. It is
because the output device or other non-linear devices in the
amplifier *generate* harmonics. This is why placing a filter
on the input of the PA generally does no good.

The sole exception to this would be if the PA has no input
or output filtering at all (some CB amps are built that way,
but no type-accepted amateur amps are), or if the input
circuit of the PA or output circuit of the exciter is so
poorly constructed it allows substantial direct radiation to
the TV system from the input circuit of the PA. In almost
any PA harmonics generated in the amplifying device totally
dominate any exciter harmonics.

Harmonics are caused by fractional cycle distortion, not by
amplification linearity issues that cause keyclicks or
splatter. When you drive a PA harder harmonic levels can
increase (or decrease), but that's because waveshape over a
fraction of the RF cycle changes. It is a different issue
than amplitude or gain linearity.

73 Tom



More information about the TowerTalk mailing list