[Elecraft] K3 and Kenwood MC-50 mike

[David Woolley]

The purpose of the transformer isn't to get anything close to a 
conjugate match.  Doing so would produce a very inefficient PA, and, I 
suspect, would produce unacceptable levels of distortion 
(non-linearities in the output impedance would be visible in the signal).

The purpose is reduce the design load seen by the PA to a point where 
the design power is achievable with the available power supply voltage.

The design load is determined by P = a * V^2/R, where a is a constant to 
allow for things like not being able to actually reach +/- V, RMS 
correction, the output impedance not being truly zero, and a safety 
margin, to allow for the presented R not being no higher than the design 
value (SWR not 1:1, with load on the high side).  I'm assuming the 
reactive part is being tuned out.

For an ideal, class B, PA, with zero output impedance, and full rail 
swing, and no safety margin, a would, I believe, be 0.5.  I don't know 
what the actual design rule value is, for typical, real world, designs.

You can see this in the alternative winding configuration for the K2, to 
give more efficiency at reduce power.  That increases the presented 
load, but doesn't change the effective output impedance.

-- 
David Woolley

On 11/04/2022 16:37, Joe Subich, W4TV wrote:
> 
> 
> 
> On 2022-04-11 10:55 AM, David Woolley wrote:
>> RF PAs are also, typically, much lower impedance than their design load.
>>
> 
> 
> Except they *DO HAVE* impedance matching - typically a broadband output
> transformer (4:1, 9:1, 16:1) to 'get close' followed by some kind of
> "antenna tuner" (automatic or manual), i.e. matching network to match
> the working load.
>