>
>Hello Hans,
>
>> If you short-circuit the grid to the cathode with a number of low
>> inductance caps ( chip caps? )in a super cathode tetrode amp - is this a
>> good start to get a parasitic-free linear? Without any supressors?
>
>Many amplifiers will be completely stable without suppressors if
>the grid connection can be made with a low impedance over a very
>wide frequency range.
>
This is not going to happen in the real world because every grid has L
and C. The result is a resonance that makes the grid virtually
transparent at that frequency. Thus, the output is no longer shielded
from the input, so oscillation is no longer impossible. For a
3-500Z/8008 with directly grounded grid-pins, the typical resonant freq.
is a bit below 90MHz.
>Such a connection is possible in tubes that use an internal cone-
>type connection to the grid with the leads brought out on a flange.
>
They also have a grid resonant frequency.
>One false assumption is that adding a small capacitor will reduce
>the impedance of the grid and improve stability. The problem with
>that approach is the impedance is reduced only at one narrow
>frequency, and greatly increased at all other frequencies.
>
The 600pF that is used to ground each 3-500Z/8008 grid in the TL-922 and
the SB-220 makes very little difference.
.
>That works well with a monoband amplifier, but as a general rule is
>a disaster with multiband amplifiers.
Grid-resonance is no different in monoband or multiband amps.
>
>The best policy in multiband amplifiers is to ground the grid with as
>short and the lowest possible impedance connection. This single
>thing does more for stability than any other single change.
>
Try measuring the grid-resonance with a dipmeter to see if this is the
case.
>> distorsion ( 48 volt final transistors ). I guess the internal structure
>> of the tube grid and its leads to the outside world will be resonant at
>> some high frequency where the tube gain maybe will be not high enough to
>> make the tube oscillate.
>
>Your statement is generally true, with a good layout.
What constitutes a "good layout", Mr. Rauch?
>If you
>measure the transmission loss through the cold tube in the actual
>chassis (no tank, no input. Connect analyzer directly to the input
>and anode), you will see a sudden loss in isolation at some
>frequency or frequencies.
>
>That is the potential trouble spot(s). Contrary to perceptions, anode
>resonances where isolation is good don't mean a thing (except for
>harmonic suppression issues).
If the anode-resonance is slightly above the grid-resonance, the input of
the amplifier is no longer shielded from the output, and a parasitic osc.
seems possible.
>
>As a matter of fact, resonance can be a very good thing if it makes
>the anode have a low impedance. Removing the resonance can
>actually deteriorate stability.
Nothing can remove anode-capacitance. As long as there is a wire between
the anode and the Tune-C, there will always be an anode resonance.
>
>Having fancy test equipment is nice, but if you understand the
>system an antenna analyzer and a regular receiver or even a diode
>detector can tell you everything you need to know. If you don't
>understand the system, all the measurements in the world won't
>mean a thing.
>
>> I guess you must make this short circuit very broadband? Must be equally
>> effective on short wave as on VHF/UHF? A number of 0.01 + 0.001 uF silver
>> micas and chip caps may be needed?
>
>Chip caps are excellent. They are the best thing that ever happen
>to power amplifiers. A 100nF multilayer chip capacitor can be a
>good bypass from a few hundred kilohertz all the way up to VHF
>and higher. No more need to parallel combos, although I still prefer
>to directly ground the element that is most critical to stability,
>either the control grid or the screen grid.
At RF, there is no such thing as a direct ground because one can not
escape the fact that even 1" copper strap has inductance.
>
cheers, Tom
--
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
|