[TowerTalk] RG-149: 50 ohm/70 ohm - does it matter?

[Michael Tope]

On 12/14/2013 2:24 PM, Jim Lux wrote:
> No, they use some sort of transformer SWR bridge, with diode detectors 
> (and some use a phase detector too(SGC), to know which direction to 
> tune).  It's not a directional coupler in the sense that there's a 
> sufficiently long line in terms of wavelengths and you couple off the 
> traveling wave like you do in microwaves.

Yes, I know. I sometimes get sloppy with my terminology (mea culpa). In 
this case, it would be incidental to the function of the auto tuner 
whether you use an SWR bridge (Bruene circuit) or a true traveling wave 
directional coupler to determine VSWR. The results would be the same.

> More like a clever transformer scheme as seen in a raft of SWR meters. 
>  These things measure forward and reverse power, but don't really have 
> a characteristic impedance in and of themselves.  What's 50 ohms about 
> them is that the "calibration" is correct for 50 ohms.

Yes, the calibration Zo is determined by the ratio of the capacitive 
coupling (voltage sample) to the inductive coupling (current sample).

> Think about it.. how are you going to build a real directional coupler 
> that's 1" long for 160 meters?  What you can do is build a form of 
> hybrid which has two outputs which are related to the forward and 
> reverse power.

That depends on what you consider a "real" directional coupler. You can 
build very small "directional couplers" that work across a broad 
frequency range with ferrite transformers. If I gave you the 
S-parameters for one of those and S-parameters for a parallel 
transmission line coupler, you would be hard pressed to tell which one 
was "real" and which one was "fake".

> Sort of.. They have a routine that takes the ADC values for Fwd and 
> Rev and turns that into a number that is similar to VSWR, and they 
> minimize that. I don't have the listing handy, but it's soemthing like 
> (Vfwd-Vreverse)/fwd, and drive it to zero.  The scaling from Vfwd to 
> actual forward power may change, but optimizing this will get you to a 
> match.
>
>  Since they never need to display actual VSWR, they don't bother to 
> normalize it or anything.  They just have a known threshold for when 
> to stop adjusting.

Yes, but it is normalized to a specific impedance in the sense that the 
VSWR bridge is calibrated for a certain Zo. If I put 100 ohm 
transmission lines on either side of a VSWR bridge calibrated for 50 
ohms and I terminate the load side 100 ohm transmission line with 100 
ohms and drive the source side with a 100 ohm source, the bridge will 
still read 2:1 VSWR even though both the input and output lines are 
matched (i.e. no standing wave on the line).  A travelling wave 50 ohm 
coupler would do the same (i.e. show a ~10% power reflection on the 
reverse port) if I installed it in place of the VSWR bridge in the 100 
ohm matched system.

>> If the auto tuner were part of the amplifier, then it would be possible
>> for the tuner to using additional information (like drive power and DC
>> current) to account for the changing amplifier output impedance.
>>
>> 73, Mike W4EF............
>>