[Amps] Re: [Amps] various parasitic sources... chassis, tubes, etc...

[Ian White, G3SEK]

Rich wrote:

>>>The 8877 has about 2/3 as much feedback-C as the
>>>3-500Z, but it has about 2x as much gain at 100 - 150MHz.  3-500Zs have
>>>a history of vhf parasite problems, so the 8877 should not be
>>>absolutely immune to oscillation at vhf/uhf.
>>
>>The 8877 is well known as a *stable* tube for amplifiers at 50MHz and
>>144MHz.
>>
>>With the input and output both tuned to the same frequency, and
>>typically 15-18dB of on-frequency gain, it's stable. If you provoke it
>>by removing both the input and output loads and swinging the input and
>>output tuning, a well-constructed 8877 VHF amp will still not oscillate.
>>
>**  Agreed, Ian.  However, VHF amplifiers do not have two resonances to
>deal with in the anode circuit - wherein the tube exhibits substantial
>gain.  HF/MF amplifiers uniquely do -- one at the operating frequency,
>and one at the VHF resonance formed by the anode-C, the anode leads, the
>DC blocker cap's reactance, and the Tune-C.
>

On the contrary, every VHF/UHF amp I can think of has at least one 
parasitic resonance. For example...

50MHz pi- or pi-l: parasitic resonance above the operating frequency, 
for exactly the same reasons as an HF amp.

144MHz quarter-wave stripline: three-quarter wave (and higher) parasitic 
modes from the stripline; and a parallel resonance between the DC 
blocking capacitance and the RF choke that provides the DC feed (this 
one will typically be below the operating frequency, where the tube gain 
is even higher).

432MHz half-wave stripline: similar higher-order resonances of the 
stripline; and a VHF parallel resonance between the RFC and the 
distributed capacitance to ground and/or the anode compartment.

In addition there are parasitic resonances at the input side. And all 
these resonances are unloaded.

But who cares? The existence of parasitic resonances doesn't matter. 
What matters is the loop gain through the tube at those frequencies.

If the loss for reverse feedback is greater than the forward gain at all 
frequencies, the tube will not oscillate at the parasitic frequency, the 
operating frequency or any other. And that brings us right back to 
grounding the grid in a way that's effective from LF through VHF.


>>Why? Because the 8877 has a grid ring that makes a very direct,
>>low-inductance connection to the actual grid inside (a sheet-metal cone
>>which very effectively shields the input from the output). If you ground
>>the grid VHF-style, with finger-stock contacts directly on to the grid
>>ring and fixed directly to the chassis, then the tube will not oscillate
>>at VHF. This applies equally to an 8877 in an HF or a VHF amplifier.
>>
>**  According to Eimac's W. B. Foote, it was discovered that the 8877
>could oscillate at what appeared to be in the UHF region.

If you don't ground the grid effectively at those frequencies, that is 
no doubt true.

>>For glass tubes without a grid ring, such direct grid grounding is not
>>possible, which is why we have to use Plan B - parasitic suppressors -
>>to reduce gain at VHF.
>>
>**  the grid-resonance of a g-g 3-500Z is in the mid-80MHz range.  What
>is the grid-resonance of a g-g 8877 ?

Much higher (a figure was quoted once on this list, so it's somewhere in 
the archives). You can see this simply by looking at the path from the 
chassis to the actual grid cage inside each type of tube.

For the 3-500Z, it's several inches of thin wire (3 in parallel) going 
from the chassis to the socket tag, and from there through the pins to 
the stamped shield that supports the grid cage. That's a lot of 
inductance at VHF.

For the 8877, the photograph at
http://www.vcnet.com/measures/8877.gs2.JPEG
clearly shows the grid cage supported on a stamped metal shell (appears 
dark brown). This shell is directly attached to the inside of the grid 
ring, and the outside of the grid ring can be nailed firmly to chassis 
by multiple finger-stock contacts. The total path inductance from 
chassis to grid cage is much lower than for the 3-500Z, and the grid 
resonant frequency of the 8877 is accordingly much higher.


-- 
73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)
                            Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek