My latest creation

[dkii]

dual WRAP boards 4 chans


http://www.uwcorp.net/box1.jpg

[georgew]

Pretty!

Does anyone want to buy a 5 gallon pail of dessicant pacs? I mention it because I saw you were using them...

I bought them intending to put them in every AP, before I decided to vent them instead. In colder climates a sealed box is going to be the thing to do, and the dessicant is a good way to avoid moisture problems.

These are a clay-based dessicant, and they can be recharged in an oven, so if you change them every time you open a box, then bake the used ones, they will last forever.

I would keep them, but it is starting to look unlikely that a sealed box is viable in a central Texas summer.

[go.fast]

Wrap Boards? Where do they come from and how much do they costs.
Looking at the pictures, I can't make out the wireless cards that are in there. What cards are they?
Thanks
George

[dkii]

My first AP I placed was vented, a nice fan and all, which worked great for 27 days. But for some reason, from the 28th day on, every time it rained, the board would lock up, and I would have to go reset it. Then it would come back up, maybe lock up once or twice more during that shower, then it would be fine until the next storm. Then it locked up once, and when I rebooted it, it only detected 1 of the 2 pci-pcmcia adapters(it detected the second, but showed no cards present) and I coul dnever get it to come back online correctly. Pulled it down, sealed up the holes, put it back, and it's been fine since. I guess it's just too humid down here to do that. So I decided on sealed boxes from now on, I've got 2 sealed boxes that have been running for almost a year without any heat problems, one is stainless and sits on the south face of a tower in direct sunlight. Then I started thinkin about humidity inside the box and what would happen when winter comes around and it gets to the low 30's here, so I figured a little dessicant wouldn't be a bad idea.


The WRAP board is from http://www.pcengines.ch/wrap.htm I paid $125 for it, but they are the alpha samples, the beta boards are $150, he's out of stock as of tonight, but expecting more very soon I hear. The cards are Agere Ruby Mini-PCI cards from YDI.

Does anyone know where I can buy Hermes mini-pci cards?

[dkii]

looks like Pascal has a few in stock now.. get em while they're hot

[georgew]

My first AP I placed was vented, a nice fan and all, which worked great for 27 days. But for some reason, from the 28th day on, every time it rained, the board would lock up...

Yeah, I anticipated rain problems. I designed a venting system that had too low air velocity to suck rain in. We have had the rainiest summer of my life this year, and I had zero moisture issues.

I used an intake 8 times larger in cross sectional area than the exaust, then I chose a fan that barely spun untill it got over 100*. We had flash floods as well as multiple days of fine mist.

I have a sealed unit in the shade, and it has to be rebooted frequently. The vented unit is in direct sunlight, when it's not raining. You can see the construction details of the vented enclosure at http://forums.star-os.com/viewtopic.php?t=1138

[Davader]

I noticed your using amps. Are the units in the box, the injectors?

How many feet to your amp/antenna?

[dkii]

Those are the indoor versions of the amps, the box is tower-mounted, with 10' 1/2" superflex jumpers to the antennas.

[bobbyc]

Hi,
do the ruby mini-pci cards have the same antenna connectors as the pcmcia cards? Also do the minipci cards have a internal antenna?
Thanks,
Bob C

[dkii]

no internal antenna, but they do have diversity support and they have some screwy connectors, I haven't been able to find a pigtail for it, so I remove the connector and solder a pigtail directly to the card.

[georgew]

all of the minipci cards I have seen use the U.FL/MHF/Hirose connectors, the pigtail at the top of this page usually works!

http://www.demarctech.com/products/reliawave-mnt/mount-cable-assemblies.html

[John Huszar]

To DKII

This is a very clean looking setup you have there! Are you running 4 sector antennas off this box?

I am curious about something, maybe you can answer my question..

I have been talking to a lot of people about wireless lately. Half of them say use an amp, the other half say dont use an amp. Which is correct? What model amps are you using in the this box you show here? How is it performing for you? Would you recommend using an amp, and if so what power output? 500mW? 1 Watt?

Thanks in advance. I was going to email you off list but dont see your email address anywhere.

John

[lonnie]

Look just under the posting and you see the icons for website, email, private message, etc.

[dkii]

Thank you. I am running 4 sector antennas off this box, however it isn't working well. I heard mixed reports about using 4 channels, but I can attest that if they are all in the same box, you will have trouble. I have one antenna disabled and the other 3 on 1/6/11, and all is working well. The 4 channels probably would have worked if I didn't stick all 4 of them in a metal can with amplifiers. I should have separated them in to 2 seperate boxes, maybe provided a little vertical separation as well.

To amp, or not to amp.. that is the question :)

It's really going to depend on your area, terrain, business plan, pocketbook, and equipment. I'm in a fairly flat terrain(central florida) with lots and lots of trees. I won't put up anything but a microcell without amping it, and I usually amp them as well. On all of my longer shots, it seems the amps give it that little extra bit of signal that it needs to be stable. If your plan is to only cover a 5 mile radius, you have no trees, and a 5000' mountain nearby, then I wouldn't use amps. You won't need them, they're another piece of equipment to break, and will ultimatly cause problems.

There is a problem using amps with lucent/orinoco/buffalo/avaya/proxim(hermes and ruby chipset) AP's when the client is also a hermes/ruby chipset. It is difficult to explain, and I have a very detailed email I can send you if you are interested.

The power output you want will depend on a few things also. Your regulatory agency(fcc, etsi, etc...) dictates how much power you can feed an antenna, and how much total(EIRP) power you can radiate. This will be the limiting factor on power output. I try to use hermes/ruby chipset cards, YDI amps, and YDI antennas, so that my entire RF section is FCC certified. YDI sells 3 amps. The indoor version has 10db of TX gain, and 12db of RX gain. Throwing out cable and connector loss, connecting this amp to a hermes/ruby card will give you 25dbm(~315mw) of TX power. They sell a full-power indoor version that has a 14db TX gain and 16db RX gain. This will bring your power output to 29dbm(~800mw) of tx power, but the amp will start distorting your signal at this level. Their outdoor/pole mounted amp has the same gain as the indoor full power, but they recommend(require if it is to be fcc certified) at least 3db of line loss between the injector and the amplifier. That would give you 26dbm(~400mw) of tx power. Anything over 500mw hasn't seemed to work any better.

If you plan on using omni's, don't. If you insist on using omni's, don't amp them, they will cause you many sleepless nights.

[WISP]

What issues are you seeing with your hermes/ruby ap - client setups. I wouldn't mind seeing this email you are referring to.

I exclusivly use lucent card on both the ap and client end without a bit of trouble.

Wayne

[John Huszar]

If you plan on using omni's, don't. If you insist on using omni's, don't amp them, they will cause you many sleepless nights.

OK, so where are omnis used then? If you have a 100foot tower location, then I can maybe see the need to use sector antennas (because an omni radiates a pattern around it, but not so much downward, correct?)

But if I want to go from the tower over to another location and broadcast the signal in a small area around that other location, then wouldnt an omni be appropriate there?

John

[dkii]

WISP- PM me your email address and I will forward it to you.

John-

The higher gain antenna you use, whether it's omni's or sectors, the narrower the beamwidth will be. A 8db sector may have a vertical beamwidth of 20 degrees, but a 16db sector may only be 7 degrees. Same with an omni, an 6db omni may have a vertical beamwidth of 60 degrees, giving decent coverage underneath it, but a 12db may only have a vertical beamwidth of 10 degrees. There are omnis with built in downtilt(done electrically) that help with this. Omni's are o.k. for small microcells in low-noise areas that you only plan to have a few customers, but every omni I've deployed with an amp has given me much trouble. I've also found much better range from a 12db sector than a 12db omni(I think most omni gain specs are bloated). Omni's also make it difficult to manage your channel usage, especially when amped. Put an amped omni at 100' and you can forget about re-using that channel for at least 15 miles. I currently have 3 omni's deployed, all in hotspot configurations. That is just about the only place I will use an omni. The main problem with omni's is they pick up noise from everywhere. They seem to find noise in places that you never knew existed.

There is another side to your first question. People look at a antenna with a 7 degree beamwidth and think that it won't reach the people directly under it. That's not necessarily true. Beamwidths are usually specified in a term called 3db beamwidth. That means outside that area, the gain will be reduced more than 3db, but not completely. Since you are so close, you generally don't need the extra gain anyways. I have some 13db sectors with a 12 degree vertical beamwidth, 480' in the air and I can get a -65 signal 100' from the base of the tower. I can also get a -65 signal 20 miles away from the tower. I have some 24db grids with a 7db vertical and horizontal beamwidth that I can aim the opposite direction, drive 10 miles away, and still pick up the signal from the backside of them.

[John Huszar]

dkii :

This is very interesting, thank you. This jives with what a local consultant was telling me the other day about the relationship between gain and beam width/angle. He was saying many times its better to go with a lesser gain omni (to distribute in a smaller area, like an apartment complex) to get better coverage.

Thanks!
John

[go.fast]

dkii :

This is very interesting, thank you. This jives with what a local consultant was telling me the other day about the relationship between gain and beam width/angle. He was saying many times its better to go with a lesser gain omni (to distribute in a smaller area, like an apartment complex) to get better coverage.

Thanks!
John

I agree with the theory that omni's shouldn't be used.
But,
They have there place in typical wisp operations.
If you were going to have a cell in an isolated area with not many subs then an omni is a good economical antenna to use. And if the omni is mounted low, like on the roof of a tall home in a subdivision, then a high gain omni is better for tree penetration.
And thenthe flat pancake pattern is not a problem if the pancake pattern covers what your target area is.
And it may be more desirable to have a flat pattern that hits the targets than a big cloud that goes everywhere .
Just my thoughts.
George

George