> > << The fix that seems to work is: reduce VHF
> > amplification. >>
> > Rich it seems to me that one way to reduce vhf gain would be not to use
> > tubes ( very expensive ones) that have max. ratings to 500 mhz on
> > frequencies of 30 mhz or below. HANK
>
> Sound practice!
Actually not. The maximum usable frequency of the device has
nothing to do with stability.
It isn't difficult to understand what the problem is. The problem is at
some frequency the grid no longer functions as a shield between
the anode and cathode. The grid, because of series inductance and
distributed capacitance INSIDE the tube has a very high
impedance.
This allows the anode to grid capacitance to excite the grid, and
form a TPTG oscillator IF the grid resonance happens to occur at
the same frequency as a high anode impedance.
> 1. Use power device with lowest transfer frequency to achieve the power
> required.
Bad idea! That's because a low transfer frequency generally
indicates the device has large amounts of internal inductance and
capacitance.
That means the device is unstable closer to the operating
frequency because the grid can be less effectively grounded at
frequencies where the anode may have a high load impedance.
2. Pay utmost attention to isolating the input from the output.
That helps for oscillations near the operating frequency, it is a non-
issue at HF in most cases.
> 3. Pay utmost attention to RF bypassing.
Near the grid, that is true. It is also true for VLF or LF parasitics.
4. Pay utmost attention to layout
That's true, because long anode leads force the designer to use
large values of inductance in the suppressor, if one is required.
There are cases, however, where long anode leads can help
stabilize the PA by moving the anode resonance below the grid
resonance, but it is generally not a good idea.
Grid leads are critical, they should be very short and direct to the
chassis. That's the bulk of the problem with tubes like the 811 and
572. Both have long thing grid leads and lots of grid capacitance.
On the other hand, tubes like the 8877, 8875 and 3CX5000 have
very short wide grid leads brought out on a flange. It is very easy to
stabilize those tubes, as a matter of fact with the flange directly
grounded the tubes are unconditionally stable in almost any layout.
> in input and output compartments to reduce stray coupling. 5. Don't assume
> that power supply wiring cannot be part of the oscillation (RF bypass
> these components properly).
Things downsteam of large chokes are only problems for VLF and
LF stability.
6. VHF components generated in the input
> circuit will appear amplified on the output circuit.
There are no VHF components generated in input systems, unless
you are talking about rectification in the grid-cathode area.
> output, to reduce the gain. 8. Pay attention to input and output SWR to
> ensure that reactive components do not excite the PA.
Reactive components don't "excite". They simply shift phase.
9. Check that the
> harmonic content of the drive source does not contain unwanted
> frequencies.
If it does, you better junk or repair the exciter! Even if it does
contain some spurious frequencies, only those within the passband
of the input circuit will excite the tube. With any decent PA design,
they won't do anything nastier than when the exciter is barefoot.
73, Tom W8JI
w8ji@contesting.com
--
FAQ on WWW: http://www.contesting.com/FAQ/amps
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
|