[TowerTalk] Why did adding a radial make the input Z of my inverted L go up?

[Jim Smith]

I've found the reason for my puzzling results.

Just so you don't have to read the original post again, I had an 
inverted L on 160 with no radials (but with coax shield connected to 
ground so I actually had one radial, the coax).  I started adding 
radials and found the Z at the base of the insulated-from-ground tower 
went down when the 1st radial was installed and went up when the 2nd one 
was installed.  (Coax centre conductor not connected to anything.)  Z 
was now higher than with no radials at all.  This is better?

I had another go at the measurements.  I decided to find where the 
antenna was resonant as it had been purposely made too long.  I 
disconnected the coax centre conductor from the tower and the shield and 
the 2 radials from the ground plate (we use plates here, not rods) and 
adjusted the Autek RF1 analyzer frequency to find the freq where L=0 on 
the presumption that if there's no L we're at resonance. 

There was no frequency where L = 0.  L hit a minimum of 15.9 uH at 1775 
kHz.  Both the SWR and Z displayed by the RF1 were minimum at the same 
frequency, 6:1 and 179 ohms respectively.  For the rest of this, fo is 
the frequency at which L was minimum

1 Ground plate alone as described above fo= 1775, SWR= 6.0, Z= 179, L= 15.9
2 Coax braid connected to ground plate  fo= 1754, SWR= 5.3, Z= 130, L= 11.8
3 As in 1 plus one 12 ft radial         fo= 1750, SWR= 4.5, Z= 134, L= 12.2
4 As in 1 plus one 42 ft radial         fo= 1526, SWR= 4.7, Z= 112, L= 11.5
5 As in 1 plus both radials             fo= 1514, SWR= 5.5, Z= 112, L= 11.7

So, I have the answer to my question.  The 42 ft radial drastically 
altered the "resonant" frequency of the antenna.  Consequently, it is 
not surprising that impedance measurements all made at the same 
frequency would go up when the longer radial was connected.

What is surprising, at least to me, is that buried radials can affect 
the resonant length of the antenna so drastically.  (They're #14, 
stranded, insulated, 2-3" below the surface.)

For the sake of completeness:

6 As in 2 plus both radials             fo= 1510, SWR= 3.9, Z= 107, L= 11.2
7 As in 6 plus 10 more radials          fo= 1521, SWR= 4.4, Z=  82, L=  8.7

The 10 more radials range in length from 8 ft to 28 ft and their ends 
are 2 ft apart.  Why 2 ft instead of recommended 8 ft for 160?  I've 
found the L works better on 40 than my dipole so based the spacing on 40m.

I have 7 more radials to install.  At that point, my radial space will 
be full.  The new ones will all be about 28 ft. 

I'm gratified to see that Z is down to 82.  I'm hoping that the 
additional radials might bring it down to 60 or so.

Now I have 3 new puzzlements.

1.  Why can't I find a frequency where L = 0?
2.  Why did the the SWR go up when Z went down when comparing case 7 to 
case 6?
3.  When I get the rest of the radials in and cut the flat top to length 
how will I know when I've got the right length?

Just for fun I connected an unterminated wideband (10 MHz) true rms 
voltmeter (HP 3400A) to the shack end of the coax connected to the L.  
Got a rather surprising and very steady 12V.  Could be 60 Hz pickup, I 
suppose.  I tried again with a 50 ohm termination and now get a more 
reasonable 85 mV which is bouncing around 2-3 mV, presumably due to BC 
station modulation.  This may be the source of the SWR puzzlement and 
maybe even the "L doesn't go down to zero" puzzlement.

I suppose I could hump the TEK 535 scope out to the tower and measure V 
& I magnitude and phase of power fed from Tx to get around the BCI, if 
that's what my problem is.

I did make 99Qs/19 mults in the recent CQ 160 SSB LP with no radials, 
just the ground plate.  I'm looking forward to the beneficial effect of 
the radials.
 
Any and all comments appreciated.

73 de Jim Smith   VE7FO