[RFI] Ultrasonic Dish

[N1BUG]

I have been working on the "dish problem". It is not as simple as it 
may seem, if one is looking for optimum performance. Aren't we all? ;-)

It is very important to match the shape (depth) of the dish to the 
detector used. Virtually every available dish, including the Edmund 
Scientific ones, are too deep for optimum capture, given the 
beamwidth of commonly available transducers. If you study the 
pattern of various transducers you will find there are some that 
have a beamwidth of about 50 to 60 degrees in one "solid" lobe with 
no nulls or side lobes. Others have a narrower front lobe with deep 
nulls on either side, then another lobe at each side, for a total 
beamwidth around 70 to 80 degrees. The latter type would probably be 
best for the deep dishes commonly available, but will not make use 
of the entire dish surface no matter what you do. The now obsolete 
transducer used in the W1TRC design was of this type.

The type with one 50 to 60 degree lobe approaches ideal use of the 
dish **if** you use a flat enough dish. You want a dish with f/D in 
0.7 to 0.75 range for these. That amounts to about one inch depth 
(rim to center) for a 12 inch diameter dish, or two inches for a 24 
inch dish. Most dishes are not anywhere near that flat! One 
exception is the 12 inch dish sold by Midnight Science. It is 
optimized for this application but for the price I don't think much 
of its quality or surface accuracy. Poor surface accuracy (not 
adhering to strict parabolic shape) reduces efficiency and can 
create unwanted pick up from directions off to the side of where it 
is being aimed.

One final comment on the two variations of transducer. Typically the 
larger diameter transducers have the 50 t0 60 degree pattern. The 
smaller ones tend to have the split lobe with overall wider 
beamwidth. Note there is another trade off with smaller diameter 
transducers: it is more critical to get them exactly at the focal 
point of the dish, and dish accuracy needs to be better to focus the 
collected ultrasound into a smaller "spot".

The fact it results in a bulkier unit aside (due to the transducer 
needing to be mounted farther out in front of the dish), I believe 
the larger transducers with solid 50 to 60 degree lobe are the 
better bet, if a truly suitable dish can be found.

Green Power Science has dishes that are actually too flat at around 
0.9 f/D! These might not be too bad, as it would mean the outer 
portion of the dish would be the area not effectively used. That 
would provide better rejection of noise from behind the dish (bugs, 
traffic, etc.). However, from looking at videos of these they appear 
to be very flexible and would probably need a solid rim support 
added. They are also coated with a highly reflective surface, as 
they are intended to be used as solar collectors. That would have to 
be removed or the dish painted. I have not worked out the math to 
see how much the effective diameter would be reduced by the too flat 
shape. Bear in mind the transducer would be a *long* way out in 
front of the dish (it has to be at the focal point, which gets 
further away from dish as the dish gets flatter or higher f/D).

I'm still looking for a dish 18 to 24 inches in diameter with f/D 
around 0.75 and good rigidity. I have not found anything. I may get 
frustrated enough to try spin casting a parabolic mold and making my 
own dish. It's not a trivial project. Maintaining parabolic shape is 
very important (else we are back to the same problem of poor 
efficiency). Spin casting is about the only practical method of 
ensuring shape accuracy I can think of, but I am open to other 
suggestions!

Obviously receiver sensitivity is another important factor. I bought 
the Midnight Science RX2 and was not happy. The new RX3 is much more 
sensitive. I have no idea how it compares with the expensive 
commercial units. I bought all the parts to build the W1TRC 
receiver, intending to compare it with my RX3 on the test range, but 
simply haven't had time and energy to do it (yet). I wish I could 
get my hands on a Radar Engineers 250 for side by side comparison on 
the range, but that seems highly unlikely.

73,
Paul, N1BUG
RFI Committee chair,
Piscataquis Amateur Radio Club