On Mon, 2008-05-19 at 15:29 -0400, Bob Close wrote:
> Jim, it has been a long time since I got to ask a question of a PhD
> engineer type!
>
> IF the transceiver had harmonics and those were transferred to the linear,
> would they not be amplified by a factor of ten? (at least where the
> harmonics fell into the range that the amp might be able to work on them)
> Then, if they were filtered they might not be so quiet as to be
> insignificant? A filter isn't a brick wall for those frequencies outside of
> the pass band. If filtered before the amp, the linear amplification would
> only be done on the good stuff, right?
Except that the class B operation of the PA device (for much greater
efficiency than class A which is most linear) involves half cycle device
conduction and that creates harmonics only a couple dB down from the
fundamental. So the PA requires its own low pass (starting with the
output matching network in a tube amplifier or the set of low pass
filters by band) to cut back on those harmonics generated in the PA. And
the power in the output represented by the harmonics from the exciter
which should have been attenuated to 40 or 50 dB below the signal aren't
significant compared to the harmonics generated BY the PA active device.
> I would think the overall effect
> would be a cleaner signal out of the linear, and a lower overall content of
> harmonics to filter (only those generated by the linear amp (which would
> mean it was not so linear)) .
Won't be a measurable difference in harmonics from the PA. And adding
the low pass filter between the driver and the PA can add some more
mismatch when the tuned input of the PA isn't perfect (and it can't be
perfect for different power levels). A grounded grid PA without tuned
input NEEDs the tuned output of the driver to have decent in band
linearity, but I don't think that will affect its harmonic generation.
>
> I guess I am, for the sake of discussion, assuming the amp was being
> driven by equipment inferior to TenTec, (or perhaps Elecraft to keep those
> guys happy) with it's resulting rice-induced harmonic content. :) Please
> ignore the dumb presentation, but I am seriously asking the question: Might
> not there be benefit to filtering between the exciter and the linear
> amplifier if the exciter was not clean?
Any exciter without at least 40 dB harmonic suppression (both HF and
VHF) should be used on the air. And should not have passed FCC tests.
>
> --------------------------------------------------
> From: "JAMES HANLON" <knjhanlon@msn.com>
> Sent: Monday, May 19, 2008 1:55 PM
> To: "Discussion of Ten-Tec Equipment" <tentec@contesting.com>
> Subject: Re: [TenTec] Low Pass Filter
>
> > Ed,
> >
> > A properly designed filter should be terminated in its characteristic
> > impedance at your operating frequency, probably 50 ohms, and will present
> > a 50 ohm impedance at its input at your operating frequency as well.
That's only true in the ideal case. Many a fancy modern filter raises
the input SWR. The rise is greater the steeper the cutoff and the more
the pass band ripple. Often a filter acts as a lumped transmission line
and changes the input impedance with frequency.
> So
> > putting a filter in between the output of your transceiver and the input
> > of your amplifier should not upset the load impedance for the transceiver
> > as long as the amplifier input looks like 50 ohms as well.
But at best the fix tuned PA input has only an approximation to 50 ohms
when the PA is operating at exactly its designed parameters. Different
loading or drive and that input Z can be different. Remember in the
grounded grid PA that the input has the same current as the output and
that makes tuning more difficult.
> >
> > However, it is probably overkill to put a filter there. One filter at the
> > output of the amplifier will sufficiently attenuate all of your harmonics
> > and other unwanted signals above the filter cutoff frequency that leave
> > via the amplifier output coax lead. Decent shielding of the transceiver
> > and amplifier and proper filtering of the wires coming out of them, like
> > power lines, microphone, key and headphone leads, etc, will keep any
> > unwanted signals from getting out through open holes or on those wires.
> > Just about all modern equipment has adequate shielding and filtering, so
> > you normally shouldn't have any problems there. So one low pass filter in
> > the coax at the amplifier output should be sufficient. Also, I see no
> > technical reason for keeping the coax cable between the amplifier output
> > and the filter short, other than the possibility that the shielding on
> > that lead is less than perfect and that would allow an unwanted signal to
> > leak out.
That concept is not well known. The reason for keeping the coax between
the PA and the low pass filter can be explained. I thought I did, I'll
try again. Most of the classic ham low pass filters were designed at the
beginning of TV and use a constant K mid section or two with m derived
end sections. Invariably these end sections involve a series resonant
circuit shunting the coax connector to make a notch at 42 (TV IF) or 54
MHz (channel 2). That series resonant circuit has a low impedance its
resonant frequency, limited only by the loss in the coil (and
capacitor). If the coax between the filter and the PA is an odd multiple
of a quarter wave long the coax transforms that low impedance to a high
impedance seen at the PA end. That high impedance causes those harmonics
to be transferred from the PA through the low pass filter and on to the
antenna. That's why that coax needs to be short. Otherwise the low pass
filter can enhance the harmonics rather than reduce them.
I have found that the classic constant K filter (and m-derived) has the
greatest tolerance for component variations and also the lowest shunt
capacitor RF currents which means for a capacitor with a limited current
carrying capability, the constant K filter allows more power than any
more modern filter design. That's because the constant K filter uses
smaller shunt capacitors than the more modern filters for the same
cutoff frequency.
> As far as the impedance presented
> > to the amplifier output, that won't be changed by the length of the cable
> > to the filter as long as the filter works into a matched, usually 50 ohm
> > load.
Feeding an antenna, the filter works into a matched load only at the
fundamental frequency, rarely at any harmonics. At the harmonics that
are not to be passed through the low pass filter, the input impedance is
not resistive, its reactive and constantly changing with frequency.
> >
> > Jim, W8KGI (PhD, EE)
73, Jerry, K0CQ, (PhD, EE, ham since 1955, registered PE in Iowa)
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|