[Top] [All Lists]

[AMPS] MFJ-259 and matching circuits

To: <>
Subject: [AMPS] MFJ-259 and matching circuits
From: (Michael Tope)
Date: Fri, 5 Feb 1999 07:38:09 -0000
Hi Jon,

The non-linear component of the tube's input impedance represents 
energy at harmonics of the drive signals fundamental or a DC term. If you 
consider that the resulting voltage and current waveforms present in the
non-linear circuit can be decomposed into a linear superposition
of sine waves (fourier series), then determining the "average impedance" 
is just a matter of doing a linear analysis on the resulting first order 
terms. The higher order terms shouldn't matter since these are represenative 
of energy which is no longer at the fundamental. Unless 
a significant amount of this harmonic energy appears across the source 
impedance, then you can ignore it in terms of its contribution to VSWR. 
If the harmonic energy present at the source impedance is significant 
compared to the first order component, then the reflected power component 
of the VSWR reading could be in error somewhat, depending on how you 
look at it. 

An interesting test would be to plot the input VSWR of the amp as a 
function of drive level. If the departure from linear behavior is as 
significant as everyone seems to be claiming, then you would expect 
to see a change in VSWR as a function of drive level. 

This is all classroom B.S. of course. The proof is in the pudding, and
you guys are the ones out there cutting metal. My first amp is still a
pipe dream pile of parts in the garage waiting for some attention.

73 de Mike, W4EF.............

From:   Jon Ogden[]
Sent:   Wednesday, February 03, 1999 3:46 PM
To:     Michael Tope; 'AMPS'
Subject:        RE: [AMPS] MFJ-259 and matching circuits

>Would it not be possible to tune the circuit for min VSWR
>using the exciters VSWR meter, as Rich suggested (real 
>world operating conditions), then turn off the amp, terminate
>its input port in 50 ohms and measure the impedance of the filter
>looking back from the cathode with the MFJ-259B (cathode disconnected
>of course)? This impedance should be approximately equal to the 
>complex conjugate of the large signal cathode impedance. It 
>would be interesting to see if this number comes close to Jon's
>estimate of 110 ohms -j (2*pi*f*27pf)^-1.

Theoretically, yes.  However, interfacing the MFJ-259 or any other 
impedance measuring device to the cathode pins is a challenge and error 
will be introduced there.  Rather, one could set the impedance as Rich 
suggests, remove the tube and put the 110 Ohm//27 pF network in place and 
check the impedance.  That is a little easier.

But you have your network theory correct.  Again, though we (you and I) 
are sort of applying a linear network analysis to a very non-linear 
system.  As others in the group have instructed me, a tube running in 
class AB does not have a constant input impedance.  Rather it varies over 
the drive cycle as the tube is cycled on and off.  But I will still argue 
an "average" value can be used to simulate a more linear system.

If folks doubt me on the "average" bit, please define then how your AC 
wall outlet can be "110 Volts" or "220 Volts."  Actually, these are RMS 
values if I remember correctly, but my point is that AC voltages vary 
wildly over their cycle as well, yet we call them by specific values.



Jon Ogden

"A life lived in fear is a life half lived."

FAQ on WWW:     
Administrative requests:

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