On Mon, 11 Jun 2018 11:11:57 -0500, amdx wrote:
> On 6/11/2018 10:09 AM,
mako...@yahoo.com wrote:
>> op,
>> i think someone else already said this but your issue is not the coax.
>>
>> Your issue is the input circuit to the radio.
>>
>> Make sure the input terminal that you connect the shield of the coax to
>> is a good ground for the radio, I don't mean you need a ground rod
>> there I mean it must be the RADIO ground.
>>
>> Is it a good connection to the radio chassis?
>>
>>
>>
> There has been a lot of talk about my radio.
> I don't believe the radio is the problem.
> I'm using an Icom R-71A, driving the 50 ohm input.
>>
https://www.eham.net/reviews/detail/496
Very impressive reviews, including a few by operators who rated it
better than the professional/commercial communications receivers they
were using alongside of it. The pictures suggest the use of a steel case
so any thoughts of it allowing stray high field strength signals leaking
past the input screening can be pretty well set aside.
However, that doesn't preclude the possibility of mains borne
interference being conducted into the receiver via its power supply
connections. Testing with a 12v SLA is an effective way to remove such
potential interference sources from the equation. Not only does it
eliminate ingress via the power supply connections, it also eliminates
the possibility of interference being conducted onto the outer braid of
the co-axial feeder where it can then reach the antenna feed point and so
appear as an interference signal received by the antenna.
>
> Connecting a 75 ohm cable to a 50 ohm input for receiving
> is not perfect, but in most cases it won't be noticed.
Very true, even in the case of a transceiver where it is of some
significance to the Tx output filter, feeding a 75Ω line from a 50Ω
filtered Tx output is normally not that much of an issue in practice
(especially if a high vswr protection circuit has been included).
> I'll be running a twist pair maybe today, That is pretty close to
> 100 ohms, I will try that with and without a 2 to 1 matching
> transformer.
If you're referring to the turns ratio of the matching transformer,
don't forget that a 2:1 turns ratio results in a 4:1 impedance
transformation ratio.
Conventional transformer winding construction practice for mains and
audio frequencies doesn't work so well at MF and HF frequencies due to
the leakage inductance issue so such transformers are normally made up
using bifilar and trifilar windings (perhaps even quadrifilar and
quintifilar in more extreme cases) to minimise leakage inductance. The
consequence of this is you rarely see such MF/HF transformers used to
obtain impedance ratios outside of the common 1:1 and 4:1 options.
The classic balun transformer used to connect a 75Ω co-axial feeder
(orthogonally) to the 72Ω centre fed half wave antenna is a trifilar
wound 1:1:1 auto-transformer with all three windings in series phase
aiding. The result is an auto transformer with two taps at one third and
two thirds the way along from either end.
Working from left to right, the connections used for the antenna are LHS
and Tap#2 with the feeder braid connected to Tap#1 and the co-ax inner to
the RHS end connection. This provides antenna connections that are
balance with respect to the braid connection (centre tap of the section
between the LHS and Tap#2) and a 1:1 turns ratio between the feeder
connection (Tap#1 and the RHS connection).
If you're using a centre fed folded dipole with a 300Ω feedpoint
impedance, you could use an even simpler bifilar wound balun with your
75Ω co-axial feeder.
A better option may be to replace that 100Ω speaker wire with a 300Ω
balanced feeder connected to a bifilar isolating transformer which
provides a galvanically isolated link to a conventional 2:1 bifilar wound
balun feeding the Rx antenna socket. If you can arrange for the feeder
impedance to be 200Ω, the Rx will see a 50Ω impedance instead of the 75Ω
in the 300Ω feeder case. However, the exact impedance value as you stated
before, is far from critical for an Rx only setup.
The real benefit of using a balun and isolating transformer in this case
will be to remove galvanic connections between the feeder and the station
and antenna grounds and thus attenuate any interfering signals which may
be polluting the station's local grounding point from travelling along
the feeder as a common mode signal voltage to be injected into the
antenna connection.
Mirroring the 1:1 isolation transformer and 2:1 balun setup at the
station end will provide even greater attenuation to any such
interference arising on the station's local grounding point (a balun will
be required regardless). In this case though, you'll need your balun's
unbalanced port to be a much higher impedance than the 50Ω or 75Ω
impedances used at the station end of the circuit.
It's *so* not easy to achieve all of that in one step that I'd recommend
the use of a third transformer to translate the notional 75Ω output of
the balun into either a 300Ω (simple 2:1 bifilar wound auto-transformer)
or else a 675Ω (3:1 trifilar) impedance match.
Attempting to get a better impedance match with non-standard broadband
matching transformer ratios is unlikely to improve the situation. As is
so often the case, compromise rules the day. Impedance matching in this
case is secondary to achieving a galvanically isolated balanced
transmission line setup to minimise unwanted signal pick up. Since the
impedance of the Beverage antenna is expected to lie somewhere between
300Ω and 800Ω, it seems to me that you'd be better off testing both 2:1
bifilar and 3:1 trifilar transformers and choosing the least worst one.
Keep in mind that if you're going to replace a co-axial feeder with a
balanced line feeder, unless you're using shielded twisted pair cable,
you'll need to support it clear of any ground influences that might upset
the balance, particularly true with the classic 300Ω ribbon cable
transmission line. Regularly spaced support posts with insulated carriers
to keep the ribbon cable at least a foot or so clear of the ground and
any patches of long grass should suffice, not forgetting to transpose at
every supporting point through a 180 degree rotation for good measure.
If you're using a high enough number of supporting posts along the line
route, it might be better to transpose through either a consistent
clockwise (or anti-clockwise - pick one or the other and stick to it) 90
degrees rotation at each support point along the route. If eliminating
unwanted signal pickup is the name of the game, then transposition isn't
optional, it's demanded (at least in the case of open wire or ribbon
cable feeders).
--
Johnny B Good