John, this is a quality reply to my question about the effect (with the
specific reference to harmonics from ham rigs) of harmonics causing
oscillation in PAs.
The powers that be pooh-poohed this possibility.
My experience with signal amplifiers, particularly X-band GaAsFETS and
similar devices, is completely in parallel with what you have outlined
below. One can make up a circuit, and under test conditions with the
correct loads, they look good, but connect them to the real world,
simply an antenna, and they take off.
Not only that, they appear to achieve an output power in excess of what
one would expect from the gain available.
But PA tubes do not have much more than about 11 dB of gain, which is a
bit low for real oscillatory tendencies to manifest, therefore the
layout in the PA has to be poor for the mechanism to show up. Most ham
PAs are pretty bad, particularly the old green monsters.
Sure ways to control oscillation is by resistive loading of the output
circuit (at the expense of gain and heating in the loss), deliberate
mistuning of the input and/or output circuits to reduce gain, and by
careful bypassing.
Loading the output circuit, either in series or parallel (difficult to
implement), seems to be the mode in ham PAs, but even carbon resistors
in series look like lumped constants at VHF. A piece of "lossy" wire has
few lumped constants, other than it's intrinsic inductance and coupling
to the capacitance of it's surroundings.
Any lossy wire will work as a suppressor, methinks, even a piece of blue
fence wire.
What intrigues me, is that certain amps appear to work for a while, even
in hygienic surroundings, then suddenly say "good bye", apparently
without good reason.
But your reference to the effect of reactance on oscillation might be
the clue, especially when one reads reports about deteriorating
connectors and water running onto Joe's feet down the coax...
Cheers,
Ian ZS6BTE
"John T. M. Lyles" wrote:
>
> Designing a cylindrical power amplifier cavity at 100 MHz, using the
> 4CX3500A tetrode series (also tried the 3000A and the 5000A tubes), I
> 'picked' a geometry which supported a mode near the 2nd harmonic.
> Computer models didn't show this, as they were using pure TEM mode
> calculations, such as ladder analysis using ABCD matrix transmission
> line approximations. There was apparently enough feedback through the
> tube and gain at 210 MHz that it could come back in phase to make an
> oscillator. This essentially made a parasite which would be excited
> when driven by the 2nd harmonic. The amplifer was class C. The
> parasite wasn't always there, and when it came on strong, the
> harmonic filter would blow one of the capacitors (it was a Bird 5 kW
> shoebox style filter) inside. The parasite was not an integral freq
> of the fundamental, thats how we knew it was parasitic. We could see
> F2, and then the parasite nearby, and then there was smoke. Of couse,
> the presence of the parasite made a lot of mixed products in the PA,
> which complicated measurements.
>
> When the transmitter was terminated in a pure 50 Ohm water cooled
> dummy load, there was no parasitic. When terminated in a harmonic
> filter, either the Bird L-C LPF network, or a stub line tuned for 2nd
> harmonic, it would become instable. The reactive termination in the
> 200 MHz band was essential to make it unstable.
>
> Internally inside the amplifier cavity, we had a patented 2nd
> harmonic suppressor.
> US patent # 4334203
> It was a high Q device, coupled into the electric field along the 1/2
> wave amplifier circuit.
> I am talking about lumped and distributed ciruits such as coaxial
> cavities now, not LC discrete component circuits. But parts of the
> parasite suppression scheme are somewhat similar to what I have been
> hearing about.
> It was coupled into the field of the main TEM mode, not directly
> inline with the plate connection of the tube.
>
> Anyhow, my parasitic would eat this suppressor circuit as well. The
> solution, found after about 6 weeks of empirical testing and
> modifications, was to change the copper strap on this suppressor to a
> nichrome coil of wire. This device had lower Q, and squelched the
> parasitic, as well as continuing to trap out the F2 component in the
> tank circuit. At this time, I was convinced that Nichrome or other
> resistive alloys had merit in high power RF circuits. Also, I played
> with Carborundum resistors, and ferrite tiles and tubes. Nothing
> worked as well in my application as the wire. The nice thing with
> lower Q was that it was no where as sensitive in position as it was
> with a copper strap and a paddle.
> I am not vouching that resistive wire is needed in 3-500Z circuits,
> or that it will solve all problems with parasites. Just that it IS
> another way to skin the cat, so to speak. Also, yes, second harmonic
> can pump a nearby parasite. I have seen it, recorded it, and worked
> around it with a suppressor.
>
> John
> K5PRO
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|