Hello

We are in the process of installing a 2.4 GHz hotspot in a small harbour. In top season there might be as many as 500 ships in various sizes - all with alu-masts and other reflecting gear.

The idea is to give access to the internet for the many tourists bringing a laptop along on their holiday.

We have earlier made a number of successfull hotspots, and did not expect a lot of trouble with this solution - but we were wrong. With our "standard solution": High gain sector and some kind of Ubiquiti radio, the noise levels are so high (noise floor = -70 dB on most channels), that the wlan is unusable. We have been trying with other equipment with little or no improvement.

It is a local problem. In a tower less than 500 m away, we don't see the problem. It is not detectable interference - when scanning, there are several channels free. As far as we know, there are no Canopy installations in the area. Our theory is, that it is an unusually high amount of reflections. There are no radars at 2.4 GHz ?

Does anybody have experience with a similar scenario ? Which antenna configuration to use, high or low power radio etc etc. ?

Hi,

Well if you use a Big Gain sector, is possible that have a low Vertical Beam Width, if you use a Omni or antenna with less Gain, you have better Vertical Beam Width, but will need more points to connect your clients. Low Gain = less directional.

Regards,
Fran.

The idea of using a HG-sector is to shield off against a lot of wlan (the antenna has the back against land), while at the same time covering the area of intererst. If you us an omni, you will pick up noise form a much larger area and get interference.

Hi,

Yes, you are ok, but if you use a less power your client need connect more near of antenna, and in this case you will have lees problems.

Regards,
Fran.

I understand your point, but the thing is, that electronics and seawater does not mix well. Therefore, we want to have the gear located in one spot if possible at all. And the place is not that big...

If it is reflections causing your noise, have you tried backing off the power of your radio? Maybe use a 2-antenna configuration with a low-gain tx + high-gain rx if you want to shield it from the land side...

You say you're using a Ubiquity radio + high-gain sector=very high power covering a small area, as you explain.

You're probably making your own noise in a 'virtual' confined space caused by reflection off the masts - I'd say this is, in the words of the great Mythbusters - 'Plausable'!

OK Gwisa, it may be a good idea with two antennas. I didn't try that here.

But we need enough power to let the tourists connect with their laptops directly - and not necessarily with direct LOS. Our current solution works wery well on a number of other sites.

Would it be an idea to use an AP with diversity mode enabled and two sector antennas ? Not a MT, but maybe a Colubris or similar ?

Also, I noticed that the ack-timeout went skyhigh. This is probably caused by some wlan on "the other side" - maybe some 50 - 60 km away. I am going to fix the ack at 56.

Problem solved ! The reason was noise from an unknown source in the building where we installed the hotspot. Our signal was jammed, and even amplified by the antenna.

We moved the hotspot some 100 m away, changed the sector antenna to an even narrower (90/17) type and increased the tilt to 10 deg. Then everything worked as expected. You can connect with a laptop and get 8 Mbps or better anywhere in the harbour now - both outside and inside the ships and inside most of the buildings surrounding the harbour.

The conclusion is, that you almost always will find a very simple solution to a theoretical complex problem, when you are working with RF - as long as you keep trying. And just forget all about low power !

The second conclusion is, that we once again missed a built-in tool for general purpose scanning. Not many own a spectrumanalyzer. I hope, that Mikrotik once will find a way to do this !