Amps
[Top] [All Lists]

[AMPS] parasitic fantasia

To: <amps@contesting.com>
Subject: [AMPS] parasitic fantasia
From: W8JI@contesting.com (Tom Rauch)
Date: Mon, 7 Jun 1999 19:37:26 -0400
Hi Jon,

> The gird in a tube is essentially analogous to the gate in a FET and the
> Anode is analgous to the Drain and Cathode the source.  Now if I want to
> make an oscillator with a FET I will feedback the output signal to the
> input.  In other words, I will connect a low impedance connection between
> the Drain and gate.  Now even though the gate may be grounded at DC, it is
> not necessarily so at RF.

Thanks for making me feel old. ;-)
 
My age is showing. I always considered FET operation in terms of 
a tube, not the other way!!
  
> The same thing happens in a tube.  The feedback path is from the Anode to
> the Grid.  The grid is grounded, but we all know that it can be a few RF
> volts above ground.  So feedback can still occur.  Now if I want to make
> an oscillator, do I put a high impedance in the feedback path?  No, I put
> a low impedance.  If voltage potential between the anode and grid of a
> tube or the anode and ground was what caused oscillations, then and open
> circuit from anode to ground would cause all sorts of havoc because the
> highest voltage potential would be built up across an open circuit.

That's right, but you also have to consider Q and the shunt 
impedances. Consider a Hi-Fi amplifier. The best high frequency 
response occurs when the anode load resistance is low, but not 
zero. DC voltage gain is the highest with a near-open circuit in the 
anode, but power gain is low. Available power is limited, and so is 
high frequency response because the stage can not even drive the 
finite shunt anode circuit capacitance.

With any practical RF amplifier, there is an optimum value of load 
impedance at any given frequency for maximum gain.

For example, a 3-500Z with a few hundred ohms or a few meg 
ohms of anode load *resistance* would be an unlikely or 
impossible VHF oscillator. Zero ohms of *resistance* could 
actually make it more unstable if the anode resonance happened to 
it happened to coincide with a grid resonance. That's because Q 
would be high, and we are stuck with the internal lead series 
reactance that lets the anode move around in voltage. The external 
anode path would have nothing to dissipate power.

External anode tubes do not suffer the same level of anode path 
impedance problem, so they can be stabilized easier by adding a 
low impedance external path. In that case the anode resonance 
can be parked far away from the grid resonance. Tubes with long 
thin leads (572B's, 811A's) suffer it in spades. 
   
> Oscillation develops from the feedback of signals.  You want to eliminate
> that feedback.

Or change the phase. 180 degree phase shift requires a certain 
amount of Q. De-Qing the circuit not only loads the anode with a 
lossy resistance, it limits phase shift that occurs on the slope of 
resonance.

That phase shift is what you see in an un-neutralized amplifier, 
when the anode current dip or maximum grid current (in a GG amp) 
is outside the point of resonance.  
 
> Perhaps this isn't worded completely technically correct and the
> technically anal will, no doubt, correct me left and right.  But think
> about it and you'll see the general point I am making is correct.

It is. But there are limits. Optimum stability is a complex blend of 
interactions. There is no universal cure that works in every case. 
Every different case should be treated on its own merits. What 
works for a 3-500 on six meters, may not apply to a HF amp. What 
works for a 3CX1500, won't apply to a 3CX1200.

Knowing how something work is important.


73, Tom W8JI
w8ji@contesting.com

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm


<Prev in Thread] Current Thread [Next in Thread>