Let's define our "trap Yagi" for this discussion as being a venerable Mosley TA-33: 3 elements each with two traps, one on each side of the boom.
Actually (having just repaired my TA-33jr) there are two traps in a single
housing on each side of the boom. The aluminum trap cover serves as the
radiator between the two traps as well as the capacitor plate for each of
Surely a TA-33 would perform somewhat better on say 15 meters if you replaced the trap director and reflector with straight aluminum tubing appropriately tuned for 15.
Possibly, but you have to be clear how much of the difference is due to
losses in the trap and how much is due to a different radiator length.
The basic yagi element spacing isn't too bad for 15m. The late W4RNL
showed that yagi elements can be shortened to 2/3 of full size with
little effect on efficiency (assuming that they have some type of loading
to maintain the proper reactance at the operating frequency.)
But the differences due to element length are NOT a factor of the trap
losses, which is the topic of discussion. You can model such a yagi
using lossless inductors, then replace them with a realistic model of a
trap to get some sense of the actual difference.
So there really are at least two sets of changes that we see when we
make a trapped yagi: the performance shift due to changes in the element
lengths and spacings, which are NOT a function of the trap efficiency,
(but rather a shift in the antenna pattern) and the loss of signal strength
due to actual power lost in warming the traps. Traps are often accused
of having low efficiency due to the former, but only the latter should be
That's why I said that the difference in performance between a trap yagi
and a monoband design is due to much more than the trap efficiency.