[TowerTalk] Tuners

[jimlux]

Kevin Normoyle wrote:
> On 11/29/2010 8:49 AM, Steve Hunt wrote:
>> That's a useful tuner simulation; but **DO** remember it assumes a
>> constant Q for the inductor. Here's how a real roller inductor Q
>> measures at different inductance settings on different bands:
>>
>> http://www.karinya.net/g3txq/temp/rollercoaster_q.png
>>
>> If you want to get meaningful loss results with that simulator, be
>> prepared to manually change the default Q value as you change bands and
>> change inductance values.
>>
>> 73,
>> Steve G3TXQ
> thanks steve. Interesting.
> 
> AH! real data? so for 160/80, Q increases for more turns (more uH). But the 
> higher frequencies are more complicated.
> 
> (note the arrl article's simulated numbers used fixed Q of 200 for the 
> inductance. That's typical in most simulated numbers I've seen)
> 
> In steve's picture, it's interesting how the Q decreases with larger uH for the 
> higher frequencies. That's counterintuitive.
> In fact it decreases for everything except 40M and 80M where it increases.
> 
> Is this because of coupling/capacitance in the coil? Is this a special behavior 
> for roller inductors vs air coil (the equations for air coil were posted earlier)
> 
> It also seems that using a single Q=80 or so would be better than using Q=200, 
> if you only had one choice.
> 
> Wondering about the picture. Could there be measurement error due to stray 
> capacitances?


Yes.. the data raises many questions.

A couple years ago, I started out trying to measure/characterize a LDG 
AT200PC with a TenTec VNA.. Never got very far: mostly because I was 
tired of trying to figure out how to automate the whole process of {"set 
tuner" "capture S11/S21 data", repeat} with two applications that did 
not lend themselves to scripting.  I'd love a pair of programs 
presenting a command line interface to the VNA and the AT200PC.  the 
former has a fairly complex user interface (what with calibration and 
all), the latter has a funky RS232 command sequence requiring 
raising/lowering RTS for a few milliseconds to wakeup the processor.

Maybe if I get ambitious over the holidays..
Maybe someone has already done the commanding interface..

BUT.. what I did discover in my limited experiments is that though the 
LDG (a switched L network topology) should have a fairly linear L as a 
function of the binary number, it does not.  That is, even though the 
inductances follow a 1,2,4,8, sort of sequence, the measured L (or, more 
correctly, the calculated L from the S11 and S21 looking into a short, 
for instance) was nothing like that.

So, Steve's seemingly anomalous data is probably pretty typical, and all 
of our speculating by dialectical means on tuner losses is probably 
bogus from the start.  That is, the models (conceptual or otherwise) we 
use are sufficiently different from reality that we can't make blanket 
rule-of-thumb type statements based on them.


Maybe what I need to do is work up a fixture for a real 2 port VNA that 
can do all 4 measurements.