[TowerTalk] Traps and Losses

[Tom Rauch]

It's amazing what a marketing tool this trap stuff has become. Now 
we even have pie-in-the-sky rules-of-thumb like "obe dB loss per 
trap.

How silly can we get!

That means loss in a trap is about 20% of applied power for a 
single trap. That means if you had a 40/20 meter vertical handling 
1000 watts, the trap would be dissipating 200 watts.

Now try this test. Find a 200 watt light bulb with the same surface 
area as the trap. Run a kilowatt for one minute with the trap, shut 
off power, and grab it. Do the same for the light bulb. Look and see 
which one your skin is stuck to.

The loss in a trap is mainly confined to the inductor, so that is the 
heat source. The plastic form in the inductor would be history in 
short order if loss was anywhere remotely close to 1 dB!  Get real 
people!  

Jim posted this.

> Trap losses have been a popular subject since the 1960s, 
> when one manufacturer with an excellent test range (I 
> have been there) measured some competitive beams. 
> The beams of one competitor showed little, if any
> gain, over that of a dipole, but still gave very good 
> front-to-back ratios. But trap design has advanced (I think) 
> in 30 years.

Every measurement has error, no matter how much we pretend it 
doesn't. Gain is a tough thing to measure outside of a real test 
range. We may not like it, we may protest it and argue against it, 
but that's the way life is. 
 
> The loss of gain due to the use of traps or other forms of 
> loading elements does not necessarily mean loss of power 
> in the sense of conversion into heat. Hence, a low-Q trap 
> does not turn it into a resistor.

The statement above is not correct. A low Q trap behaves exactly 
like it contains a conventional dissipative resistor!! 

At resonance, it is a pure dissipative resistance of some easily 
measured value. At other frequencies, it is a resistance in series 
with a reactance or in parallel with a reactance.  It has ESR and 
EPR, and the resistance is dissipative just like in an Ohmite 
resistor.

Every bit of energy that is lost is lost only because it turns to heat. 
Otherwise, it is radiated as a signal!

Now the pattern might be less than optimum, but that would 
certainly show up as a loss of directivity.

 >Rather, it turns it into an 
> inefficient trap, which allows significant power beyond the trap 
> point. Low Q will also mean a higher resistance, but in 
> relationship to the reactance of the components, and
> this may also create a higher power loss, but usually not to 
> the point of self-destruction. The reduction of gain on 20 
> meters of a 20-15-10 meter trap beam is in part due to the 
> fact that at 20 meters, the traps act as inductive loads in the 
> elements, reducing effective radiation from the element to the 
> degree that coil loads can be considered to be almost
> non-radiating substitutes for what would otherwise have been 
> at that point a linear radiating element segment.

While it may be convenient to explain away discrepancies, energy 
just doesn't "disappear" without a trace.

A traps ONLY loss mechanism, besides desired radiation, is 
undesired heat. Heat is a result of current passing through the ESR 
of the trap when it is operated below resonance. The trap is maybe 
2/3 out on the element on twenty meters and current at that point 
is low. Loss in the trap, just like in a resistor, is current squared 
times ESR. What is lost turns directly into heat.

If the trap isn't getting hot, it isn't causing a power loss. If a trap 
reduces field strength, and doesn't destroy F/B ratio, it can do so 
only by converting RF energy into heat.

You can absolutely bet the ranch that if pattern remains the same 
shape but FS is reduced the trap will be getting very very warm at 
high power levels.


73, Tom W8JI
w8ji@contesting.com

--
FAQ on WWW:               http://www.contesting.com/FAQ/towertalk
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com