WiFi Antenna Installation Issue

[Zydeco]

We need help with our WiFi Antenna. We are using a Bullet 2HP system with a POE injector. We mounted it at the masthead with about 90 feet of Cat 5 cable. (no, we do not have a 90 foot mast - the cable runs down the backstay, then to the middle of the cabin where the POE injector and router are mounted) It worked fantastic for awhile, then went haywire after about a week at anchor. It could still "see" numerous APs and connect, and according to the AirOS software it had a good connection (for example 80 dBm and good TX and RX rates (for example 36 Mbps) but when we opened a browser, it would not open any web pages. From time to time, it would connect very briefly, then stop. We tried different APs, different computers, varios settings on the Bullet, etc but no joy.

We spent last night in a marina and charged the battery bank. This morning the Bullet works perfectly. My guess is that the cable run is too long and the Bullet does not get quite enough power when the batteries are discharged. We have a 440 ah battery bank and it was down about 90 ah when we were having problems with the Bullet. Still reading a little over 12 volts.

We need to be able to use the WiFi antenna when we cruise the Bahamas this winter and we will be anchored for weeks at a time. The batteries will be down as much as 150 ah.

Any ideas?

[rwidman]

AH has nothing to do with it, it's the voltage. Your "guessing" may lead you down the wrong path. Measure the voltage that the unit sees when it is working correctly and again when it is not.

If it does turn out to be a low voltage issue, you might consider a seperate battery for the unit with solar panels to keep it charged.

[Zydeco]

Originally Posted by rwidman

AH has nothing to do with it, it's the voltage.

Yes, of course. I just gave the Ah for a rough estimate of the state of discharge. Another battery is an idea, but I'd rather avoid it if possible. Good suggestion though, assuming it's the voltage drop. I'm thinking more along the lines of a different installation with a shorter cable.

[colemj]

As long as you have >5V at the bullet, it should work fine. The current draw is ~100-200mA, so there shouldn't be a large enough voltage drop to cause the problem even at low battery voltage and 90' over thin ethernet cable.

How is your POE connected? Straight to the 12V house system? How is the router powered? When you are having these issues, have you tried plugging the ethernet cable straight into the computer?

If your router is connected directly to the 12V house system, I suspect the router is having trouble when the house voltage drops below 12V. You can buy very inexpensive DC/DC converters that will regulate the router supply at 12V. Google ASUS EeePc car adapters. I got one off eBay for $6.

Mark

[Zydeco]

Originally Posted by colemj

How is your POE connected? Straight to the 12V house system? How is the router powered? When you are having these issues, have you tried plugging the ethernet cable straight into the computer?

If your router is connected directly to the 12V house system, I suspect the router is having trouble when the house voltage drops below 12V. You can buy very inexpensive DC/DC converters that will regulate the router supply at 12V. Google ASUS EeePc car adapters. I got one off eBay for $6.

Mark

Thanks Mark. I tried plugging the ethernet cable straight into the computer thus eliminating the router but it didn't help. So it's not the router. Good suggestion, though.

Both the router and the POE injector are connected to the bus on the 12V house distribution panel so in that sense, yes, they are connected directly to the 12V house system.

[senormechanico]

I have the same system on my boat. I've experienced the same problem, but suspect either multipath interference or marginal reception. -80db is not all that great.
Also, just because you are receiving an unencrypted LAN doesn't mean it's actually connected to an internet source. Around the San Juans, some people take their DSL router home with them when they leave their vacation home leaving just an activated LAN.

[scotte]

Yep, it's not a problem with voltage or power - it's just a marginal wifi signal. SNR is a pretty broad estimate, and doesn't always mean you'll get a working signal. Also remember that just because you can hear the a/p it doesn't mean they can hear you.

What sort of an antenna do you have on your bullet?

[Zydeco]

The antenna is an 8dbi omni. I think we have more than just a bad or weak signal. We have been using this for awhile and connected to many APs. Then, suddenly, in three different anchorages it would not connect to any AP. There were at least half a dozen APs available in each anchorage and it would not connect to any of them. And by not connect I mean that the AirOS software indicated it was connected with good signal strength, but we could not get on the internet.

What senormechanico said is exactly what it was acting like - connected to an AP that was not connected to the internet. Which is possible, but seems unlikely that half a dozen APs in three different anchorages 20 miles apart would be disconnected from the internet.

I suppose its possible that there was some sort of internet outage in the area. Not long ago I spent the entire morning trying to fix my email account only to find out that it was the provider having issues.

[rwidman]

Originally Posted by Zydeco

The antenna is an 8dbi omni. I think we have more than just a bad or weak signal. We have been using this for awhile and connected to many APs. Then, suddenly, in three different anchorages it would not connect to any AP. There were at least half a dozen APs available in each anchorage and it would not connect to any of them. And by not connect I mean that the AirOS software indicated it was connected with good signal strength, but we could not get on the internet.

What senormechanico said is exactly what it was acting like - connected to an AP that was not connected to the internet. Which is possible, but seems unlikely that half a dozen APs in three different anchorages 20 miles apart would be disconnected from the internet.

I suppose its possible that there was some sort of internet outage in the area. Not long ago I spent the entire morning trying to fix my email account only to find out that it was the provider having issues.

You can't be completely sure it's a voltage issue unless you find a location where the system works and then lower the voltage, keeping track, until it no longer works.

There are several variables here. That's what I meant when I posted "Your "guessing" may lead you down the wrong path." in my previous post.

[senormechanico]

Ron,
My son (system administrator for his company) and I did just that before we installed mine. As I remember, the Ubiquiti Bullet is rated for 10.5 to 24 volts.
We hooked up a variable voltage power supply and found it would operate with good stability down to about 7.8volts (measured at the input pins of the bullet). The current draw increased as the voltage input decreased.
It became unstable at around 6.7 volts.

[rwidman]

Originally Posted by senormechanico

Ron,
My son (system administrator for his company) and I did just that before we installed mine. As I remember, the Ubiquiti Bullet is rated for 10.5 to 24 volts.
We hooked up a variable voltage power supply and found it would operate with good stability down to about 7.8volts (measured at the input pins of the bullet). The current draw increased as the voltage input decreased.
It became unstable at around 6.7 volts.

Most folks would not have gone to that much trouble.

That information should eliminate a voltage drop as the OP's problem. It also reinforces my statement that he might be heading down the wrong path with his "guess". When troubleshooting a problem, one must eliminate all but one variable if one hopes to be efficient.

[barnakiel]

We tried the following:
- antenna above, (powered) USB cable downwards,
- same, with network cable downward (all gear at PC),
- same, with coaxial downward and all gear at the PC.

The last solution gives us best signal.

b.

[RunningFish]

I'm looking for solutions to mount the bullet and would love to see what others have done.

[Zydeco]

Well, I've been down the wrong path before. Many times.

In any event, Ron-I'm doing just what you suggested. We are tied up in an area where it works great and connects to several APs. I'm letting the house batteries discharge and we'll see if the problem redevelops.

Maybe "guess" was a poor choice of words, but isn't troubleshooting little more than a series of guesses (sometimes educated, sometimes not) as to the source of a problem? Then you try to eliminate or test each possible source in some order based in part on how likely its the cause and how easy it is to eliminate or test for it. When I was in the Marine Corps working on airplanes, we called this "shotgunning". Just change parts until the problem goes away. I think the local Nissan dealership works the same way. At my expense, of course.

Anyway, the cable length wasn't just a WAG. The unit worked fine for months mounted lower with only 25' of cable. We just moved it to the masthead prior to this trip a couple weeks ago. Since that was the only thing that changed, it seems like a likely culprit. Maybe its just a red herring. Or, based on what I've read here, maybe the longer cable just makes the symptoms worse and the real cause is much more sinister. Like maybe I did a bad job of wiring it to begin with.

[senormechanico]

One more "data point" if you will.
My son and I actually did that voltage test to see how much ethernet cable would be acceptable for a bridge for a friend who lives in Eastern Washington. (Friend's DSL transmitting from a Bullet to another Bullet in the middle of a wheat field, there via ethernet cable to his house.) It handled 1000' of cable for the voltage supply just fine with average of a couple volts drop, but wouldn't pass the data at all at that length. 300' for data seemed about the limit for data transmission.