Wireless Access

Hi!

 

I recently sold some IAPs to a client and I'm in the process of installing them. I'm not a Wireless guru as I usualy do cable networking and servers. So excuse me for what might seems to be stupid questions but I'm doing the best I can since I'm able to get those AP and sell them to him.

 

Here's some infos on the setup:

 

-IAPs = 1x IAP-104, 3x IAP-105, 3x IAP-134 (More to come once we figure out what are the best for us)

-I got the AP-ANT-1B for the 104 and 134 IAPs

-They are all setup on a Virtual Controller

-They all have Static IP

-I have different name for all of them for easy recognison

-They all seems to communicate correctly with each others, I can all see them on the VC

-They are all connected to a 24 port PoE Switch 802.3at/af compliant with new Cat6 cable runs

-They are all indoors

-The highest ceiling is ~20 feet, most of them are at ~12 feet high. And they are all mounted horizontally on the ceiling and in an open space.

-They operate at 2.4ghz and 5ghz and I let them choose their channel. My client would prefer the 5ghz band though...

-Their power is between 18 to 22 Db

-There will only be about 20 clients at peak time for inventory purpose of his warehouse.

 

So here's my couple questions for you guys :

 

1- How can you improve the range of IAPs? I mean, my client have some cheap 60$ TP-Link wireless N router that has WAY more range than the IAP-104, IAP-105 and IAP-134 I sold him.

 

I'm standing right under the IAP-134 and my Nexus 4 phone doesn't event get 100% signal. I'm pretty sure it's a configuration issue...

 

I understand that I might need more in the future to fully cover his space, but right now I'm more concerned about them outperforming the cheap routers he have so the price is justifyed.

 

**I'm not sure about the External Antenna gain setting. I put 4 for the 2.4ghz and 6 for the 5ghz, is this OK ?

 

2- What's the maximum lenght of a network cable to fully power the IAP?

 

3- The VC tells me all my IAP get a -90DBm noise signal. Is that good ?

 

4- Should I manually set the channel on them and should they all use the same channel?

 

Again, sorry for my dumb questions but Wireless isn't my thing but I DO want to learn more about this because I'll probably get more needs for solutions like that in the future. I really hope you can help me.

 

Thanks in advance!

1) You can change the minimum TX power but there might be a reason the power is not set a maximum. If the AP's are located close to each other they will set the power lower lowers, which will give you, better channel reuse. Just because an AP can blast a signal out further doesn’t mean the client can or should be connected at that distance from the AP. When a client is connected at the edge of wireless coverage (and you have the power turned up) that client will slow transmissions for all other clients attached to that AP. Also keeping the TX power high on the 2.4 could cause co-channel interference. 

 

2) 100M, which is the spec for POE

3) Are you asking Noise floor at -90Dbm? 

4) I would not recommend setting the channels manually. I would not use a single channel if you decide to manually set your channel settings. Why would you want set your channels manually? ARM (adaptive Radio Management) works pretty well out of the box if you are manually setting channel and power then something is wrong

 

Rarely use the external antenna versions of Aruba ap’s.

Here are the specs for your antennas I don’t see a gain in the datasheet. http://www.arubanetworks.com/pdf/products/ap-ant-1b_ss.pdf

I’m guessing here, as I don’t see the gain on the product datasheet, is your gain is set to high.

By the power level you reported, 18-22db, it’s a sparse deployment. Are the cheap ap’s 5Ghz? Are you comparing 2.4 to 5Ghz? 

Thanks alot for the fast answer!

 

1) I didn't change the setting of power, only the external antenna gain. In the PDF you found, it says Frequency/Gain like this :

 

2.4-2.5ghz / 3.8DBi

4.9-5.875ghz / 5.8dbi

 

I thought that was the externnal antenna gain setting I needed. That's why I put 4 and 6 because the interface doesnt accept a dot like 5.8

 

Just checked and all radio on all AP are set to use "ARM"

 

The Tp-Link router they have is a WR941ND with 3 antenna set on 802.11n with WPA, like you find in many household. I can see it pretty far, like 150 feet away from it, but ALL the Aruba AP lose signal at like 40 feet from them... I don't think for a 1k$ AP it's suposed to be this kind of coverage, right ? The TPLink are on 2.4ghz and I was comparing with 2.4Ghz on IAP.

 

2) Thank you

3) Yeah it's noise floor. Yesterday during work hours it was around -88 and -90. But when the shop closed it was around -92 and -96. So I guess everything is fine there.

4) I'll let them use ARM.

 

But what about the IAP-134 when I stand right under it I don't even get 100% signal?

 

I tried my best to put the AP as far as each other. But like I said, far=40feet in my case since I lose signal around that range. If I plug a standard home router instead of the IAP I get 2x the coverage. If only the IAP could get me a strong ~75feet that would be the best, but right now it isn't the case.

1)     It was late last night and I didn’t see the gains. You are correct the gains are correct per the datasheet.

What is the power level of the Aruba ap when you were testing?

 

Your question about coverage for a $1K ap. Your are comparing single radio home based ap’s to an enterprise grade duel radio ap for cost.  If you want to compare cost you need to look at a product that is closer in specs.. ap93 is a single radio ap but enterprise grade.

 

Are you strictly looking for coverage?

Or is performance and client density and feature sets a concern?  It sounds like client density isn’t an issue from your first post (20 client max).

 

How are you testing and showing your not getting 100% signal? What connection speeds are you seeing on that device? What connection speeds can that device support? What is the orientation of the antennas?

 

40 feet does sound low but Arm could be turning the power down. Try testing a single IAP turn off all other ap’s ARM will turn up the max power at that point. Another issue could be the ap’s on the ceiling are able to see each other (nothing in between the ap’s at the ceiling) so the power is turned way down. But when you get down on the floor and behind the racks/shelves (not sure what’s on the racks/shelves in the warehouse) but the low power could cause less penetration on the ground.  

 

Physically how far is one ap from the next ap?

First of all, thanks alot for helping and guiding me on this. Really appreciated!

 

I sold them those AP because I knew they were high quality product aimed at entreprise and they have the budget for such kind of devices.

 

I'm not really looking specificaly at coverage, it's just that I found it weird that the cheap router was being seen alot farter than the IAP. Density on the other hand can be a trouble because there is many customer coming in with Smart Phone and they have a wireless phone system for employee too that's work on 802.11g 2.4Ghz bandwith WEP.

 

We need something solid, fast and reliable. Coverage can be increased by adding more APs and that's ok. Right now we have 7 IAPs and we runned 13 cables accros the building.

 

For the signal strengh on my phone, Today,  I'm starting to test with the laptop that will be used with these AP. I'll let you know more about this.

 

I tried many orientation of the antennas, and finally let the antennas on 3 different orientation for maximum visibility. What would be the best ? All of them pointing down vertically?

 

The power is set by ARM, so I can't really tell what it is. Or tell me where to look.

 

Physically they are at ~40 to 60 feet from each other. Some are more in an open space without shelving (Customers area) and others are in dense stocking area (In the stocking warehouse). I should have let you know that this building is a flooring warehouse that stock lot's of hardwood flooring and ceramic, so not really wireless friendly, but it's not everywhere like that in the builing, only in the warehouse and everything is on the same floor at the same height.

 

Thanks alot for your time. I'm going back at testing signal.

There is a well used analogy to put wireless technologies into perspective. Imagine the radio comms to be sound waves and that the TP-Link can be heard over a large distance as it can shout very loud - the clients will hear it from far away, but what if the clients are only able to whisper in comparison? the TP-Link will hardly hear them - communications have to be matched evenly in both directions.

One access point means one single collision domain as they are all on the same channel and slower rates (they are slower because they can be sent and decoded over a longer distance). More time spent transmitting means more chance of another device transmitting and a collision occurring. Retransmissions are sent at a lower rate compounding the problem.

Imagine this as a picture - *( AP )* each asterisk is a client and the brackets are a circle showing the extent of the APs range when one client transmits the AP will hear it but not the client at the far end. If the second client begins transmitting there will be a collision at the AP. Do a google search for "near/far issue".

The IAPs in a cluster have their power lower to encourage only the clients closest to each of them to connect, and so that clients move onto other APs as they move away and closer to another one. High rates are achievable at closer distances so communications are shorter and more devices can co-exist in an area both in vicinity of one IAP which translates to more devices that can be supported across the whole cluster.

Smaller cells mean smaller collision domains so two groups at either end of say a warehouse can communicate with each other access points at the same time without interfering.

Single APs are great for homes but poor for companies trying to adequately cover large areas.

Thanks alot for the explanation. All this is starting to make sense.

 

So, wich AP would be the best for me? Should I stay with the IAP-104 or get those IAP-134 ?

 

I'll stop worrying about range, I'll just make sure we get coverage everywhere with as many AP as it need.

 

I did some testing with my phone and it seems like it has to Re-Auth everytime it switch AP. Any hint on having a better roaming? I think it's more client/device side than the AP fault, right?

You should not re-auth when roaming from ap to ap on the same virtual controller. The VC is the device that is doing the auth which is across all of the ap's (assuming we are only using one VC). How are you seeing re-auths on your device?Are you dropping traffic (setup a ping and then roam)?

Checking in the wireless setting of my phone it says "Auth" and showing a circle "working" when changing AP. But I'll get at you guys once I test with the laptops that will be used and a constant ping test. We also discused replacing the Wifi Card in the laptop if they wern't good.

 

Right now I'm busy on other things at the same office.

 

Thanks for your awesome fast replies!

Testing as proven to be succesful with a laptop!

 

I got rock solid wifi and speed.

 

Started 1000 pings and walked around the building for 20 minutes, until the ping finish and I only lost 3% (34 pings) during travel. But I didn't see a different while browsing the web using remote desktop on my terminal server. So for me everything seems perfect.

 

I disabled 5ghz and only use 2.4ghz. The laptop is getting way more signal than my phone. It "only" connects at 65mb/s but it seemed enough for our needs.

 

I only saw 2 place where signal dropped below 50%, so I'll order 2 more IAP.

 

I never lost connection to wifi, nor got affected by roaming while doing the test.

 

Hopefully everything will work as good as today when our inventory system go live.

 

Thanks alot for your time and helpful answers! It proven that my cellphone ISN'T a good tools for wifi testing...

 

You guys are awesome.

1. This is a bit of a false belief I suspect. The previous response from ddipert is bang on. Louder WiFi isn't always a good thing!

 

2. ddipert answered!

 

3. -90db is a good noise floor. Generally, I worry when it starts going over -80db in busy/challenging environments, and -75db in easy/light environments. The closer you get to 0, the worse it is. -100db is a theoretical. I've never seen better than -96db and didn't believe it!!!

 

4. No. Let ARM do it's thing, unless you can specifically identify a reason that means it's malfunctioning. Bad AP positioning is a normal reason this happens. Typically, it's where APs have direct line of sight to one another, but obstacles to all areas requiring coverage.

 

There's no dumb questions. Only dumb answers!