it looks like going with ABS plastic is the way to go. A 3/16 ABS sheet is relatively stiff, though it will need supports along the way to make it stiff enough. OSH has Al sheeting that is 6" wide at $1 a foot. Scalling the 2145MHz AMOS to 1090MHs, the reflector should be
5.9". Is a tenth of an inch critical on a reflector?
Also. if you look at the assembly diagram, The PE support is at the top of the E segment. I assume this is to support the wire relative to the force of gravity. Now wouldn't you put all the supports at the top of the segments? For instance, the lower wire will eventually slip over time due to gravity pulling on it. Also, it still seems to me at least one of the wires, say segment A, should be supported, else the wire will twist. I can't see any reason why an extra support would be a problem.
This antenna looks easier to build than the biquad. As you know, soldering to the copper plate is difficult. I may build one for wifi too if the L band project works well.
Not stiff enough. If the spaceing between the wire element and the reflector changes, the antenna will detune.
L bracket stiffeners.
Where did the 2145 come from? Center of the band is about 2442 MHz. Also, I don't do antennas in feet, inches, cubits, yards, or in other non-metric standards. Think metric or wavelengths.
My guess(tm) is that you're not going to see any detuning caused by a change in reflector dimensions. There may be a slight change in pattern, but no detuning.
What assembly diagram?
Reminder:
However, I haven't run this through the optimizer. The dimensions shown are those from the original article. Close, but not perfect.
The NEC2 deck is in there somewhere. There are only a few numbers that need to be changed for 1090 MHz. Feed the result to a modeling program and see what you get for dimensions. Reminder.... all the dimensions on this model are in wavelengths. If you need help, ask.
You're correct. It will slide if the insulators aren't tight. Get a small hot melt glue gun and use it. You can also squash the copper wire with a pair of pliers to give the glue a better grip.
The insulators are critical in both material and dimensions. I once tried using nylon insulators for a 2.4GHz antenna. Major mistake. However, methinks coax dieletric is a bit tacky. PTFE will work best, but tends to cold flow. I have some polysulfone rod that will certainly work, but I don't wanna waste it. Whatever you pick, test it first in a microwave oven. If it gets hot or melts, it's the wrong stuff.
About the same.
Not with a propane torch heated giant soldering iron tip.
The article indicated the antenna was cut for 2145.
I found the NEC2 deck. I never got that fancy graphical version of NEC2 to install, though I could put it on an older 32 bit computer. NEC2 is substantially more cryptic than say Spice.
I can get thicker ABS, say 1/4 inch. I'm going to check the Al Rem center before making a decision
LT-SPICE is fine for me. 4NEC2 is the easist of the bunch to use. However, the install instructions need to be read and followed. Try again, it's worth the effort.
Ideal would be a 6" wide piece of aluminum C channel. You could also get a sheet metal shop to bend you one.
Scaling the original didn't work too well. The original antenna was not optimized and the results showed it. I scaled the NEC2 model to
1090MHz and then ran the resultant mess through the optimizer (for about 5 hours). See:
Some notes:
I changed just about everything, except the rear reflector.
The wire guage is 2mm dia (#12 AWG).
Everything is in wavelengths. Dimensions are to the wire center line.
Note that the characteristic impedance is 200 ohms, not 50 ohms due to the balun.
I had to run the center frequency plots seperately from the sweeps. Sweep default to the start frequency, which creates some rather odd plots. I've already complained on the 4NEC2 forum.
Thanks. I will try your optimized version. I caught the 200 ohm deal when the SWR was way off. I was playing with the optimizer too. I've done a lot of filter optimization, and it's been my experience you only change one or two parameters at a time until you are very close.
I assume you consider tuned to be lowest SWR, not necessarily highest gain. If I could remove some of the dips in the vertical response, I'd exchange that for slightly less gain.
There are some parts of the optimizer that are still a mystery to me. Initially, what I did was only allow it to change one item at a time under the same assumption. That worked but took literally forever. Then, I dragged home a faster machine, and decided to see what happened if I let it try to change everything at once. That worked every time I tried it. There was a tendency for the numbers to go insane on occasion. I would just hit stop, reset the values to the last sane optimized value, and then hit continue.
Incidentally, when I set it to optimize everything except the number of elements and the reflector dimensions, it took about 15 seconds per run on a P4/2.8GHz, 40 seconds on a PIII/1GHz, and 10 seconds on a P4/1.8GHz laptop under WINE on Ubuntu. Convergence was usually achieved in under 30 runs, so the days of all day optimizations are over. If yours takes longer, see below.
However, there's a problem, or rather a design error. If you try to optimize the antenna further using the supplied NEC2 deck, it's going to take 9 times longer than normal (the number of frequencies in the sweep). It will also optimize at 1050 instead of 1090MHz. That's because of the frequency sweep card. The work around is a bit clumsy, but is good enough for now.
Hit F4. This brings up the new NEC editor.
Hit the Freq/Gnd tab.
Save the frequency settings on a piece of paper.
Uncheck the "sweep" box, and set the frequency to 1090 MHz.
Run the optimization (F12). Be sure to hit "Update NEC file" button when done or you lose your results.
When the result finally display, you'll get everything except the frequency sweeps graphs.
Go back to the NEC editor and replace the frequency settings.
Run one calculation (F7) and the graphs should appear with 1090 at the center with minimum VSWR at 1090 MHz.
I've reported this to Arie on his 4NEC2 forum.
Nope. Best compromise between both. I gave both 100% weighing. I was more interested in the effective bandwidth of the antenna for Wi-Fi. Since you application is RX only, you could possibly squeeze another 0.5dB of gain out of it by tolerating a higher VSWR.
Yep. I ran the optimization for the AMOS-7 antenna for best gain and best looking vertical pattern, leaving VSWR with only 20% weighting. It was a waste of time. The problem was that the side lobes were so far down (-15dB) that dramatic changes in the side lobes only affected the main lobe gain slightly. If side lobes are an issue, I suggest you switch to the inverted-AMOS design, which has more gain, less side lobes, but a narrower horizontal beamwidth.
Please double check my calcs for the wire gauge (#12 AWG or 2mm dia). I'm not 100% sure of my numbers and need a sanity check. (The complex things I usually get right. It's the simple things that screw me up all the time).
I also found a possible construction problem. I have a different type of antenna on the bench, with a roughly similar construction in that the driven elements are suspended by dielectric insulators. My bench tests showed a resonance somewhat lower than expected due to the insulators. They are far too wide. The dielectric effects slow down the RF by the velocity factor of the dielectric material. After I drastically reduced the mounting rod diameter by filing the end, and converted to a notched and hot melt glue end mount, instead of a through hole, the antenna now tunes correctly. For Wi-Fi, this mistuning is probably too small to be significant, but for a narrow band transponder antenna, methinks it should be considered.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.