D-Link DWL-M60AT. We discussed those details in my June 11 message and your reply on June 12. I should have used the word "rotate" and of course I really mean degrees and not inches. Once I found the best location at the end of my wooden desk the reported signal strength average would vary by a couple of db with a +- 10 degree rotation. Meaning that if it is rotated 10 degrees from its current "best" location my signal drops a couple of db. And by the way... the D-Link antenna came with a RP-SMA to RP-TNC adapter that I didn't need but refer to below.
My Linksys PCI card antenna is 17.5 cm from hinge to tip.
My Linksys WRT54GS antennas are 11 cm from hinge to tip.
I didn't add an antenna... I replaced one with another. See below.
Regarding the longer / shorter antennas. I'm a bit confused and I don't think I worded my question succinctly. My leftover Linksys antennas from PCI cards are longer than those on the router.
17.5 cm to 11 cm. (from the hinge) I had one RP-SMA to RP-TNC adapter laying around so I mounted a 17.5 cm antenna on the router replacing ONE OF THE 11 cm antennas. So the router has one 17.5 cm antenna and one 11 cm antenna presently. Might I get better overall performance from the WRT54GS router this way? If so... with another SMA TO TNC adaptor I could use the other leftover 17.5 cm antenna and be even better off. You said "The extra length also moves the radiating part of the antenna well above the top of the router box for improved antenna pattern (and clearance). These are much better than the shorter antennas." ... so that's why I'm asking for clarification. The router ships with the shorter antennas.
Sorry. The only way to answer questions and remain sane is to immediately forget everything and move on to the next question. I sometime re-read the previous history, but that takes too much time. If I tried to remember all the previous information, I would soon become... well, let's not go there.
It's difficult to guess what's under all that plastic. (Hint: tear it apart and take photos). According to the data sheet, it has a gain of
6dBi and a -3dB beamwidth of 80 degrees in both vertical and horizontal directions. It's also vertically polarized. With an 80 degree beamwidth, you should have been able to rotate the antenna 40 degrees off axis, and still have a reasonable signal.
I'm not sure what the extra narrow beamwidth means. My guess(tm) is that either the antenna doesn't meet spec or that you are creating a multipath problem. When you move the antenna off axis, there's still a component of the signal going direction to the intended target. However, you now have most of the signal going 10 degrees off axis which will tend to bounce off something in the room and eventually also hit the target. The problem is that there's a slight propagation delay in the bounce path. These can cancel or add with the direct path when they arrive. Welcome to nulls and fades. Try rotating past the 10 degree point and see if the signal comes back. If it does, you have a multipath problem. Incidentally, this effect is why you don't want to test antenna patterns in an enclosed reflective area.
Yep. They do that on the PCI cards to get away from all that metal in the case.
The two antennas on the router do not need to be matched or identical. They can be literally anything. Since only one antenna is active at a time, there is little or no interaction between antennas. If ridiculously long, I would make some effort to not allow them to tilt towards each other, or there will be some interaction.
The "overall" performance is largely dependent upon the antenna pattern of each antenna. It will be no better than that of the best (longer?) antenna. If the differences in pattern between the two antennas is substantial, then there may be areas of the house covered by one antenna, and not the other. This is good as it will fill in the nulls and offer additional coverage. I've taken advantage of this effect by installing a biquad directional antenna and pointed it at the neighbors on one port, while leaving the stock omnidirectional antenna on the other port, to take care of my local, in house coverage. The resultant overall pattern is something like a keyhole with a big lump in the direction of the biquad.
Using and adapter will work just fine. However, without tearing into the antenna, I can't tell how well it will work or if it will constitute an improvement.
I tore apart one antenna to satisfy my curiosity. I would rather not mention the vendor. The antenna is about 24cm long (from the hinge) so I assumed that it had some form of vertical collinear arrangement inside. The gain specs were rather vague which made me curious. I tore it apart and found that it was just a vertical coaxial dipole, with about 2dBi gain. The rest of the length was just coax cable. It's major benefit is that it elevated the antenna above the access point or away from the metal case when used with a PCI card. These are real advantages.
Meanwhile, I did a dumb thing. I built a simple vertical collinear antenna out of a single piece of #14AWG copper wire. On paper and in the computer model, it looked like an 8dBi omni antenna, about 60cm long. It proved totally worthless when attached directly to the back of an access point. What happens is that a vertical collinear radiates 1/2 of its power in the lower 1/2 wave of the antenna. The next half wave radiates 1/4th. The next does 1/8th and so on. Most of the RF was going into the access point case at the base and very little was making it to the tip of the antenna. The radiating part of the antenna should be away from any metal that will affect the antenna or it just won't work well. So I added another 13cm of coax at the bottom, some ground radials, and it sorta worked as expected.
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.