[Amps] Tube gain vs frequency

[John Lyles]

Alex and Gerald pretty much gave you replies that explained this. Here is more 'free' information. Usually the output circuit is the most difficult to design as you push tubes in Power x Freq. Output being taken between the following terminals, screen grid and anode (plate) for tetrodes and between control grid and anode for triodes (no cathode followers or inverted amplifiers in this discussion).  The transit time is influenced by the electric field inside the tube between these elements, which is of course determined by the spacing of the electrodes and the HV applied (and more complicated, by the plate high freq AC voltage component superimposed on the DC). Higher the voltage gives faster electron movement, but this cannot be continued as breakdown between elements occurs in vacuum and on the insulators outside the tube (especially there). So you can reduce the spacing. This is contrary to raising the HV, so there is a trade off of parameters there. As the operating frequency and
voltage causes traditional current flowing back and forth (electrons actually), eventually will get to a point where the physical transit time is too long and the gain drops off. I know this is simplified, but there are plenty of texts out there that explain in great detail. Both ARRL and other professional texts like Terman explained this. 

The output capacitance is fixed, from anode to the other elements, so that it is accomidated in the circuit design, as others pointed out. At lower frequencies, it is absorbed into the tuning and coupling network parameters by design. At some point, in a transmission line UHF circuit, the first voltage standing wave minimum will not be far enough away from the peak standing voltage inside the interelement spacing, so designers will use higher fractional-to-full wavelengths in the lines, such as a 3/4 lambda circuit, which moves a tuning short further away from the tube connections. Otherwise tuning contacts would be right against the tube surfaces! You'd be surprised but the stray capacitance of a cylinder or ring contact above another ring contact can be minimized to a degree, but as you pointed out, short fat tubes will have some capacitance to deal with. In older glass tubes, the capacitance was mostly from the elements inside, and less dominated by the connecting cylinders. But
this can be dealt with, as discussed above. 

The lead inductance in tubes is a considerable difference in old tube designs and modern ceramic metal tubes. In older designs tubes were long with leads coming up from the socket to the elements. These leads become part of the circuit at higher freq. This is directly related to the ongoing discussion here on GG and GK amplifier topologies. In reality, as the frequency goes up, neither is correct terminology, as both the grid and cathode are above ground by this series inductance, so the actual voltage across the interelectrode spacing in the tube is not what is measured or calculated at the socket. At VHF and UHF, this can be a real problem, leading to amplifiers that have poor gain & efficiency. Fat wide straps and cylinders, of course, will improve this. Or resonating with these inductances to reflect a ground inside the tube, via a standing wave in the circuit. All sorts of tricks can be played but at some point, lead inductance may be the limiting factor in a particular tube. Of
course, as others and I have alluded to on this forum, they can also cause parasitic VHF resonances in the circuit/tube combination that should be accounted for by design. 

A practical limitation of high power gridded tubes is the power dissipated in the screen grid or control grid if it is part of the output circuit. The circulating currents are high, and due to skin effect, concentrated in the lead-in to the tube elements. At some point, this will cause problems, limiting how high in Freq and Power they can go. Water cooling, forced air, and other techniques apply here. 

Finally, when tubes get bigger, to handle high power, there are internal waveguide modes that can happen in the large spaces of the output elements (remember, anode to screen or grid) that will restrict performance unless accounted for in the design of the circuit. 
 
73
John 
K5PRO

> Date: Wed, 26 Aug 2009 19:20:37 -0700 (PDT)
> From: Stirling Schmidt <kc0nxm at sbcglobal.net>
> Subject: [Amps] Tube gain vs frequency
> Hi all:
> ??? I'll try to keep this short, but have several questions.? First, what is/are the factor(s) that determine a tube's upper frequency response?? I tried thinking (uh-oh!)?but only deduced that ceramic tubes, with their much larger plate structure, should only have more stray capacitance to a nearby ground plane than a glass tube, and therefore much less usefullness into the vhf-uhf region - obviously not the case, so what gives?? Second, wouldn't it be beneficial to employ a tube for an HF amp that naturally begins to lose efficiency above HF (seems like the 833 would be a prime candidate)???Third, if RF flows mainly on the surface of conductors, why don't ceramic tubes have an insulator at the top?(it seems as though all the RF flowing around the bottom edge of the plate would concentrate heat exactly where it's least welcome - at the seal - a top insulator would at least divide the current flow in half)?? Bear in mind these are beginners' questions - Thanks in advance for your
consideration.