[Lf] Internal Variometer analysis, Oh Dear !!

[Andre' Kesteloot]

Rik Strobbe wrote:

> Hello Andy,
>
> The 250uH coil in the shack and the 1500pF of the coax will act as
> impedance matching :
>
>    -----/\/\/\----+-----
>           L       |
> R1             C ===     R2
>                   |
>    ---------------+-----
>
> The formulas are :
> XL = Q.R1
> XC = (R1^2+XL^2)/XL
> R2 = R1.(1+Q^2)
>
> But the limit for a 'stable' transformation is R2 > 2.XL
> In your case R2 = +/- 100 Ohm (antenna impedance) and XL = 214 Ohm (250uH
> at 136kHz), so this condition for a 'stable' transformation is not
> fullfilled and the slighest change of L or C will have an huge effect on R1
> (= impedance seen by the transmitter). With the values you have R1 changes
> from 50 Ohm at 135kHz to 5000 Ohm at 152kHz. I wouldn't worry too much
> about the 'notch' near 270kHz as the lowpass filter at the output of your
> TX will 'isolate' that.
> A lower L/C ratio (put some C in parallel to the coax) might give better
> results but will transform antenna impedance to lower values.
> But even if you get the transformation more 'stable', changing L (in order
> to adjust antenna resonance) will alway affect the impedance transformation.
> Maybe a low-Q 'double matching' network (with 2 L's and 2 C's) will give
> better results.
> Anyway, I suggest to simulate the matching before spending some more
> power-FETs. As long as you have a huge peak or notch close to 136kHz your
> FETs are not save.
>
> 73, Rik  ON7YD
>
> At 12:15 19/09/00 +0100, you wrote:
> >Just done a network analysis on the antenna matching system mentioned in
> >my previous Email and this sheds quite a bit of light on the exploding
> >PA problem.
> >
> >The network analysed was as follows :
> >
> >250uH variometer (series input element)
> >1500pF to ground (representing coax cable capacitance)
> >4.8mH loading coil,
> >270pF antenna capacitance
> >100 ohms loss / load resistance
> >(The last three in series, shunting the coax capacitance to ground)
> >
> >Results :
> >
> >135.1kHz       Zin   53 ohms resistive, close to its actual measured
> >value.
> >152kHz         A Very sharp resonant peak with the network appearing as
> >an OPEN circuit at the input
> >269kHz         Very sharp resonant peak and network Zin appears as a
> >SHORT circuit.
> >
> >You may recognise this situation as similar to the parallel / series
> >resonance of quartz crystals - the network topology is similar.
> >The 152kHz open circuit shouldn't give a problem as there should be no
> >energy at this frequency if the amplifier is unconditionally stable.
> >However, the 269kHz peak is very close to the second harmonic of the
> >drive frequency, meaning the Tx is terminated by a short circuit at 2.F.
> >
> >Now, this shouldn't be a problem for a current drive PA stage at its
> >fundamental, and a push pull PA  should cancel out second harmonic
> >energy anyway.   But instinct says this IS the problem --  circulating
> >currents within the PA perhaps which would be why they didn't show up on
> >the monitoring circuitry or on the scope.   Thinking about it, I did see
> >some even harmonics present in the RF current but didn't take any notice
> >at the time, and there was a significant visible 2nd harmonic component
> >to the output from the low power test amplifier too - but its
> >significance hasn't struck until now.
> >
> >It looks like the values for variometer and coax cable were just an
> >unlucky combination to give the second harmonic resonance, so I guess
> >some careful analysis and a more complex matching network are needed -
> >perhaps with a low pass filter at the PA output to isolate harmonic
> >impedances will make for a safer matching network.
> >
> >Also, the comment about a few percent increase in coax voltage / current
> >was a lie.  According to the above analysis, the values actually rise by
> >4.3 times in this network (which just happens to be the factor
> >SQRT(loading coil / variometer) so the coax feed needs replacing with
> >UR67 or better for high power.  The phase of the volatage / current on
> >the coax is -70 degrees so we have the classic tuned feeder situation
> >where feed losses contribute disproportionatly to overall loss.  Initial
> >calculations were obviously flawed !
> >
> >Alternatively, go back to the first design with a variometer built into
> >the loading coil, but it would be nice to be able to tune from indoors.
> >A motor on the outside coil variometer perhaps or remotely switched taps
> >?
> >
> >Andy  G4JNT
> >
> >
> >--
> >The Information contained in this E-Mail and any subsequent correspondence
> >is private and is intended solely for the intended recipient(s).
> >For those other than the recipient any disclosure, copying, distribution,
> >or any action taken or omitted to be taken in reliance on such information is
> >prohibited and may be unlawful.
> >
> >