CONNECTOR ADAPTERS

I am loking for RP TNC to standard SMA adapters.

The reason being I just cludge together a 1/2wl coial antenna made of .

141" hardline and hobby brass tubing for the sleeve. Right off I noticed this antenna had about 3db more gain than the plastic coated antenna on that came with the router although these are supposedly the same type antenna.

Jimmie

Reply to
jimmie68
Loading thread data ...

Oh I can tell I wrote this in a rush. Coial =3D coaxial

and I am trying to build a little less kludged design now that I know it works. I thought the conenctors world help me dress it up a bit.

Sorry been in the hospital for a little outpatient surgery yesterday, still a little out of it.

Jimmie

Reply to
jimmie68

Well, your spelling is having a bad day, but everything else seems intact. Good luck on the recovery thing.

I found plenty of [RP-TNC to RP-SMA] but no [RP-TNC to SMA] with various Google searches. I'm fairly sure someone makes it, but I sure couldn't find it. You can do it with two adapters, but at that point, you might as well buy a pair of crimp connectors, and make a pigtail. That's what I do for my wi-fi antennas.

Oh, found one:

The picture is all wrong and the price is outrageous. Use the adapter wizard if I picked the wrong sex. Otherwise, I suggest that you make your own "pigtail".

Incidentally, crimp tools aren't that expensive if you know what to buy:

These are about $35/ea (or less). 2 of them cover everything from LMR-100 to LMR-400 (SMA to Type N). I think the numbers are: HT-336K (yellow handles) HT-336G (green handles)

While I've got you drugged and cooperative, I suggest you build something better than a coaxial antenna. That's what's inside the typical wireless access point antenna:

If you need a head scratcher, notice that the exposed center conductor is a 1/4 wavelength long, but the coaxial sleeve is considerably shorter.

There are plenty of better antennas worth building:

I recommend an AMOS or Franklin antenna:

The hard part is making the balun. Use semi-rigid solderable coax and you'll be fine. A biquad is also a good wi-fi antenna, but the AMOS/Franklin has more gain and a better (sector antenna) pattern.

Reply to
Jeff Liebermann

-

I guess it must be the drugs because I have a brass door kick here and I am invisioning it as the groundplane for the AMOS/Franklin.

Whats the practical length limit you could make the AMOS antenna at

2.4Ghz? I have a 10ft piece of 3 inch fiberglass tubing.

Jimmie

Reply to
jimmie68

There's very little soldering that needs to be done to the reflector. I suggest you use cheaper aluminum angle or "C" channel. Kinda like this:

My first version was a wood 2x4 with a strip of aluminum foil duct tape attached. Cheap, crude, and simple.

With larger than 7 or 9 dipoles, the operating bandwidth decreases and the VSWR climbs at the band edges. See the reflection coefficient display on the network analyzer near the bottom of:

It's shown much better in the article on the inverted AMOS antenna at:

See the 4NEC2 output for VSWR for a 7 dipole antenna. My guess(tm) is that 9 dipoles is about the limit of practical construction. I need to build yet another NEC2 model to be sure.

Reply to
Jeff Liebermann

One of the problems with building and designing AMOS/Franklin antennas is that nobody seems to have posted an NEC2 deck suitable for tinkering. 5 frustrating hours later, I threw together a 5 dipole AMOS antenna model. See:

The NEC2 file is stuck to the bottom of:

The design is from dimensions stolen unchanged from the chart near the bottom of:

Most (not all) of the letter designations in the above article were used in the model. Everything is in wavelengths with results scaled for 2.4GHz. That will allow the antenna to be easily scaled for other frequencies without recalculating all the dimensions.

I did not model the required 4:1 (impedance) balun. I cheated and just set the working impedance to 200 ohms. I'll fix it later (after I figure out how to model a balun).

The two halves of the GW dipole wires are un-necessarily duplicated in the model. I'll fix it later with a GX card.

The 4:1 VSWR is also far too high. The dimensions in the article are close but not correct and will need to be tweaked. I'll run the optimizer later and post corrected dimensions (time permitting) and when I move it to a faster machine.

Reply to
Jeff Liebermann

I am a little bit familar with these antennas, though I didnt know them by this name, because they are sometimes used with secondary radar systems. I was thinking of placing two antennas in my fiberglass tube to create an omni pattern.

This would probably two antennas mounted on opposite sides of the same reflector.

Please keep in mind I like to try a lot of things for which I really dont have a practical use. I have also done quite a bit of learning by destroying.

Jimmie

Reply to
jimmie68

Argh. I just found a big mistake. I'll fix it sometime today and add more models as time permits. (Still, not bad for after midnight hacking).

There are photos and models on:

that do exactly that. Search for: "Omnidirectional antenna built from two Amos antennas".

Yep. You'll need a power splitter. Note that the tx power is split equally between the two antennas, thus reducing your EIRP to half from each antenna. However, the receive sensitivity is unaffected by the presence of the 2nd antenna and power divider. The bad news is that omni type antennas pickup interference equally well in all directions. Think carefully before using or building an omni.

On the other foot, everything I do has a dollar sign attached.

Reply to
Jeff Liebermann

Ok, I found some connectors that will go from RP male TNC directly to .

141 hardline for about $9 each plus shipping. This seem to be the way for me to go on some of my client PCs as it has a little less loss than the factory antenna and by making the .141 hardline long enough I can make it peak from behind the PC giving it a better LOS view of the AP. To clear up my confusing ramblings instead of rewriting. The antenna for the AP will be made use RP male TNC connectors and the client PCs will use RP sma connectors both connected to .141 hardline. The antennas for the AP may wind up being colinear arrays. I will at least build 1 for play. The antennas for the client PCs will be 1/2 wl coaxial dipoles. I was thinking of building a 2.4Ghz version of the old ISOPOLE antenna but couldnt figure out how to build the impedance matching network.

JImmie

Reply to
jimmie68

Nope. .141 hard line is too small and lossy to be useful beyond about a meter in length. It's also difficult to handle. I suggest you get a length of LMR-240 coax (about 0.240" dia) and matching RF Industries crimp connectors, which should cost about $6/ea from various online vendors. Also get the crimping tools I mentioned.

Reply to
Jeff Liebermann

Ok. Mistakes and screwups are now mostly fixed. See:

I'm still using 200 ohms instead of 50 ohms until I figure out how to model a balun. I also got a surprise in that the gain of the 7 dipole version is only about 1 dB more than the 5 dipole version. That probably means I goofed somewhere. The NEC2 deck is in there somewhere.

Reply to
Jeff Liebermann

I don't follow why the receiver is not effected by the power divider. The receiver has a minimum SNR where it can function, and you have reduced the signal fed to the receiver, so you would think it's performance will suffer.

Reply to
miso

Another thing came to mind. If your wifi box is MIMO, does that mean you can skip the power divider and just connect an antenna to each port?

Reply to
miso

For my use I would only be using a few inches per antenna, a foot at the most. Im talking about an indoor antenna mounted on the back of a PC or wireless router. I want the feedline to be stiff because I intend for it to be the support for the antenna rather than encasing the antenna in plastic.

Reply to
jimmie68

Well, it requires that you understand how a Wilkinson power divider/combiner works. Let's name the 3 ports A, B, and C, where A and B are the two ports going to seperate antennas, and C is the "common" I/O port. Also, to make things simple, let's ignore the usual 0.5dB internal loss and pretent there's perfect isolation between A and B.

If you send some RF into port C, the power gets equally divided between ports A and B.

If you send some RF into port A, all the power appears at port C and none appears at port B.

If you send some RF into port B, all the power appears at port C and none appears at port C.

If you connect a receiver to port C and two antennas to A and B, there's no loss in sensitivity. All the power received by each antenna goes to the receiver.

If you connect a transmitter to port C and two antennas to A and B, then the TX power is divided equally between the two antennas with a corresponding reduction in signal strength.

Reply to
Jeff Liebermann

Yes, with limitations. For spatial diversity MIMO to work, you need to have a separate path between each antenna from the access point to each antenna on the client. That usually means omni antennas for everything. So far so good.

However, what the OP wanted was to take two sector antennas, place them back to back, and create the equivalent of an omni antenna pattern. The problem here is that the antenna patterns do not overlap and therefore there's no common path between a client and BOTH antennas at the AP. For MIMO, that's just not going to work.

Also, you generally need to have a reflective environment for spatial diversity MIMO to work. That's easy indoors, but not so easy outdoors. If you use two sets of directional antennas at both ends, it's highly likely that the signals will follow the exact same path, have the same exact propagation delays, and be totally useless for spatial diversity MIMO.

Reply to
Jeff Liebermann

I assume you mean none appears at port A. If there are no losses from port A to port C and no losses from port B to port C why don't you get twice the output at port C , assuming the injected signals are identical?

I was under the impression that with a Wilkinson combiner half the power was dissipated in the internal resistor and half in the output load. i.e. 10mw at port A would give 5mw at port C. 10mw at port B would give 5mw at port C. If the signals are identical output at port C would be 10mw or am I wrong as usual.

Reply to
LR

Oops. Yes, it should be port A. Nothing goes from A to B or B to A. (I'm a bit distracted due to current medical issues. My appologies).

You do if both antennas hear the same signal. However the original question was about using to opposing 180 degree sector antennas. If these antenna work as expected, then only one antenna will hear the remote client radio.

Zero in the resistor. If you feed a signal into the common port C, the phase angle on both sides of the 200 ohm resistor is exactly the same. Since they're in phase, they cancel and no power is dissipated in the resistor.

Wrongo. I made a total ass out of myself on the NEC mailing list exponding exactly the same misinterpretation of how a Wilkson combiner/divider works. 10 mw into A gives 10 mw into C because there's usually about 20-30 dB of isolation from A to B. As long as B is terminated with 50 ohms, the isolation is sufficiently high that no power gets dissipated in port B.

Y'er wrong, but don't feel bad about it. It's a common misconception that I also managed to suggest. I can supply more detail if you want, but I gotta do some fast reading as I just realized I don't completely understand exactly how the A to B isolation works.

Reply to
Jeff Liebermann

As long as the inputs/ outputs are balance the power in the internal resistor is 0.

Jimmie

Reply to
jimmie68

Oops (this is becoming a bad habit). I meant 100 ohms.

This looks like a good workable explanation:

See the first paragraph which ends in: "The resistor adds no resistive loss to the power split, so an ideal Wilkinson splitter is 100% efficient."

For what's inside, see:

Scroll down to: "Simple 2.4 GHz splitter for two antennas" Nice photos and drawings:

Note that the 100 ohm resistor is missing. That works only if you don't care about the isolation between ports A and B. Since the antenna impedances are 50 ohms only on the data sheet, the isolation would be lousy even if the 100 ohm resistor were included.

Reply to
Jeff Liebermann

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.