Ran some test numbers to sort of frame the discussion.. After all, we're not
really talking about a perfectly generalized matching situation, but with a 43
foot radiator, which is "special" (i.e. it's not a half wavelength, so the
feedpoint Z isn't way high).. Two different models for a 43 ft vertical give
feedpoint Z at 3.75 Mhz of 12.5-185j (perfect ground) or 50+156j (real
ground).. (didn't model radials or anything.. just a stick in the ground)
Trying several types of matching network, and assumed 1500W through the network
For the first case, a low pass L of 1470pF and 8.77uH does it
9.2A, 375V on C, 10.28A, 3kV on the L
For the second case, a low pass L of 6.6uH and 493pF, or a low pass pi of 260
pF, 6.9uH, 500 pF does it.
for the L, 5.4A, 1200V on the L, 10.3A, 1250V on the C
for the pi, 1.7A, 390V on C, 5.6A, 1300V on L, 10.5A, 1300V on C
Those are all RMS, so peak voltages are 1.4 times higher (and good design
practice would be to use 2.5 or 3 times RMS voltage)
And, obviously, a real antenna will be different, with different values, but I
think it gives an idea of the voltage and current ratings required for the
components.
For comparison, a typical RF silver dipped mica (e.g. CDE CDV16 type at 500pF,
1000V) is rated at about 2.2A max current. Lest one think you could parallel a
pair of 250pF caps, the rating on the 250pF units is half: 1.2A (the cap is
physically smaller, and it's basically a thermal dissipation thing)
So, you'd have to series parallel a whole raft of these babies to get the kind
of ratings you need. They're probably a $0.50 to $1 each, but you'd need a
fair number.
A more conventional ceramic doorknob type (e.g. a CDE 291/292 series) would be
500pF @ 6kV and rated at 6.8A at 3MHz. But those cost about $30 each.
Building an L to take 5-10A and stand off a couple kV from end to end is
trivial. A half dozen or so turns of AWG12 enameled wire will do nicely. (and
if you look inside those autotuners, that's what they do)
-----Original Message-----
From: Dennis OConnor
Sent: Nov 24, 2008 9:30 AM
To: jimlux
Cc: towertalk@contesting.com
Subject: Re: [TowerTalk] Was "43ft Vertical Feeding
HI Jim, thanks for the comments...
Yes, the relays will handle the voltage and current at 1.5kw... The ones I
use are ~ $2.50... I buy typically 2 dozen at a time for the price break..
You have to remember that the voltages are not that high at the base of a
43 foot vertical - and they are low on the 50 ohm coax side and not much more
at the antenna feed point, 1200-1500 volts maybe at worst case...
>>> A bit more voltage, I think.. a couple kV perhaps, depending on the
band. If you want to run 160, then a lot more voltage.
I build tuners, for converting 50 ohm coax to 600 ohm open wire line and
the voltages there are considerably more, and the relays work.. My tuners
handle 1500 watts without blinking... I have 7 out in the field (woods
actually) at the moment and #8 is on the bench...
We make our own condensors (motorized in my case) from aluminium sheets;
and 0.200" spacing handles 1500 watts in the tank circuit feeding 600 ohm line
with unknown (but high) impedences and reactances.. A conservative estimate is
that it will handle 5KW minimum, though I have no way to test that... The
dielectric breakdown for air is 3000v per millimeter... In a 1500 watt antenna
system you are not going to see those kind of voltages... A Tesla coil is a
another story..
>>>> Tesla coils see around 10kV or so in the primaries, which is where
you're building parts. Dielectric breakdown for air, in a uniform field is
3kV/mm, but you're hardly a uniform field. The issues would be breakdown along
the insulator path, and corona off the edges of the plates. If you're using
0.060 thick aluminum, and you nicely round the edges, the radius of curvature
is 0.03 inches. At 71kV/mil for air, when the voltage gets to around 2kV, it's
going to start to breakdown. In reality, around 1kV, you're probably getting
significant breakdown (OTOH, if the duty cycle is low, the ozone concentrations
might not get high, so you'll get some loss, but at least all the rubber parts
won't fall apart in a day). the 0.2" spacing works out to a breakdown of 14kV
in a uniform field, but along a creepage path (e.g. if you have a spacer
between the plates), you would probably get breakdowns at around 4-6kV. If the
spacer has a stack of different materials, it gets trickier
.
I have one I just whipped up to convert an 80 meter base fed vertical into
a halfwave base fed vertical for 40 for CQWW - so I needed a parallel tuned
tank circuit with a tap for the coax feed... ten feet of copper tubing and 2
pieces of aluminium from the scrap box to make a 71 pf condenser and I was in
business... Took all of an hour, with lots of coffee time and lolly gagging
around... I gave this one 0.300" spacing because the voltages across the
parallel circuit will be high at 1500 watts... No signs of arcing so far and
this is in a plastic box laying on the ground - and has a half inch of snow on
it as I type this......
Be aware that building a single L-match for one band, as I described, is a
trivial matter... And taken one band at a time, and later integrating it into
an enclosure with relays, etc., is not daunting when taken one step at a time -
over time...
Basically, your analysis says you don't want to build... That is fine -
your choice...
>>> well, the original poster said "if one's available I'd buy it".. so we
digressed into an exercise of what such a product would cost (and actually do
cost, since there are such things being made). Clearly, one can trade time for
money, as you say. Especially if you have a well stocked junk box and enjoy
fabrication and packaging, then it's a good way to go.
>>Jim, W6RMK
(excuse the weird formatting.. the webmail client is funky)
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|