[AMPS] Re: Conjugate match

Thu, 4 Sep 1997 08:33:21 +0000

> From:          Peter Chadwick <Peter.Chadwick@gpsemi.com>
> To:            "'amps'" <amps@contesting.com>
> Date:          Thu, 4 Sep 97 08:03:19 +0000

> >I'm not sure what you mean by the "output impedance of the 
> >transistor". How was that parameter determined? 
> 
> In the same way as ra  is for a pentode - from delta Vc over delta Ic.

I thought so. 

That parameter is NOT what we measure looking back into 
the PA. It would be like  matching the resistance change of a tube or 
voltage in a power source by holding the grid voltages steady for an 
instant of time and perturbing the anode  voltage. In a tube it is a 
measure of the stiffness of the anode to cathode resistance change 
for FIXED grid bias voltages as the anode voltage is moved.

Bruene "assumed" the plate resistance of the output device "looked 
like" the source impedance, but if you look at his graphs in QST 
you'll see the ETO PA crossed ZERO reverse generator SWR at about 
1100 watts. That would be the point of conjugate match, and if you 
measured anode efficiency at that point it would be highest, and 
larger than 50%.

By the way triodes, tetrode and pentodes...and even my solid state 
amateur rigs, have a higher source Z than the conjugate of the load Z 
when drive is reduced. Measuring efficiency reveals efficiency peaks 
at one level of drive ( in solid state fixed tuned rigs it is 
sometimes beyond the rating of the devices, but an external tuner 
will bring it back in a safe area) and decreases either side of that 
value. Source impedance always increases (in non-class A PA's) as 
drive is reduced, but if a reverse power generator is applied and 
the tank adjusted for minimum reflected power while the PA 
is driven normally with a fixed level of drive, maximum efficiency 
will be obtained.

If drive is reduced until the PA device moves into class A, the 
source impedance  simply stops changing. At least it did this in 
about 15 PA's I tested, and a few PA's Maxwell tested.

> I gather that's the way Spice does it as well.

No, SPICE works perfectly if you do a correct model. There is at 
least one paper published on that subject.

73, Tom W8JI

--
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