Author Topic: Wireless vs. CAT5 connect  (Read 1928 times)

0 Members and 1 Guest are viewing this topic.

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Wireless vs. CAT5 connect
« on: May 02, 2014, 08:35:26 PM »
I have discovered something that seems reproducible and therefore likely to be a real effect.

My RED is sitting in the dining room, with a 50' CAT5 UNshielded cable running down to my router.  About 5 feet away is a Netgear box that both extends my wireless, but also allows up to four hard-wired devices to plug in and have the NetGear convert them from a hard wire to being connected with the router via WiFi.  WN2000RPT Universal WiFi Range Extender and 4 Port WiFi Adapter.

This Netgear was purchased to allow me to set my controller in a little shed further from noise, and also allow the antennas to be moved even further away from my house.

The antennas are in a relatively waterproof box, which sits in a huge ziplock baggie that is really standing up to the rain.  This is connected to the controller via a SHIELDED cable.  The controller is grounded.

NOTHING CHANGES in this experiment but swapping the CAT5 from the router (and just laying the end down by the board) and picking up the UNshielded CAT5 short patch cord  that runs over to the NetGear. 

With CAT5 hard wired to the router, I have a relatively quiet appearing, but not perfect oscilloscope tracing.  When the huge storms moved through the midwest earlier, I was getting efficiencies upwards of 80%, usually never below 40% and sometimes was 3rd to 8th in rankings for the US.

I plugged the NetGear in, and rebooted the RED.  Now I'm running 4% or so efficiency and the baseline seems to be noiser, with similar autoset gains. 

I have swapped back and forth twice, and the huge falloff in efficiencies and the perceived noise in the baseline follow as I've described.

I am going to be ticked off if all the effort to move the antennas away from the house, and therefore dependent upon the wireless adapter NetGear to be in the picture, since there is no way I can run CAT5 out to that building.  Radio waves go there very nicely.

Anyone ever hear of this before or have a clue what I might try?  I'd think that using WiFi vs. the CAT5 would actually improve things, since there are no long UNshielded runs of CAT5 that might act like a noise antenna.

I have not touched the antennas, the preamp nor the run of shielded cable in, nor even the location of the board on my table, nor dressed or tweaked the wires coming and going to the board. 

Very strange, or so I think.
ECWx.info
&
ECWx.info/t/index.php

Offline jmcmurry

  • Jim McMurry
  • Forecaster
  • *****
  • Posts: 528
  • Davis Vantage Pro 2 Plus Wireless.
    • Juneau County Weather
Re: Wireless vs. CAT5 connect
« Reply #1 on: May 02, 2014, 09:33:57 PM »
Dale,

I have a similar setup and I found I had to run shielded to the NetGear.  Perhaps try that?

- Jim

Forum Search and Google Can be Your Best Friends

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Wireless vs. CAT5 connect
« Reply #2 on: May 02, 2014, 10:18:30 PM »
I've switched over and will see how things look in the morning. 

I see my B channel has a far lower gain setting right now.  Maybe it is all the nighttime vapors and swamp gas?  I'll see what numbers look like in the a.m.


Oh, has anyone else noticed that the Overview map, which usually showed the active stations as little green antennae, hasn't been showing up?  Weird, they were always there before.
ECWx.info
&
ECWx.info/t/index.php

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Wireless vs. CAT5 connect
« Reply #3 on: May 03, 2014, 10:13:48 AM »
Update:
Despite Jim's suggestion that a shielded cable may make a difference, I tried that and overnight my efficiency went to the single digits.  A few hours ago I plugged into the router directly and am  now up to 25%.

I will look for some type of wireless converter that may make a difference, and if I can find a snap on ferrite to put around the input I'll try that too.

Grrr.
ECWx.info
&
ECWx.info/t/index.php

Offline W3DRM

  • Forecaster
  • *****
  • Posts: 3360
    • Emmett Weather
Re: Wireless vs. CAT5 connect
« Reply #4 on: May 04, 2014, 04:58:10 PM »
I've switched over and will see how things look in the morning. 

I see my B channel has a far lower gain setting right now.  Maybe it is all the nighttime vapors and swamp gas?  I'll see what numbers look like in the a.m.


Oh, has anyone else noticed that the Overview map, which usually showed the active stations as little green antennae, hasn't been showing up?  Weird, they were always there before.

Dale,

It looks like the "Overview" map has been changed. The default listing on the left-side is now gone. You can turn on or off the station markers by clicking on the "Stations" icon at the top of the overview page. On the top there are also selections that let you turn on/off Lightning/Counter, Clouds and Weather. You can also switch from C to F temperature readings.

One other thing I've noticed lately about the "Overview" map is that I sometimes have difficulty getting it to display the map and overlays. Many times, I just get a blank page. It takes several reloads of the page to get the map to come up. I've seen this happen on Chrome, IE and FF and on three different computers. Not sure what would cause this to happen.
Don - W3DRM - Emmett, Idaho --- Blitzortung ID: 808 --- FlightRadar24 ID: F-KBOI7
Davis Wireless VP2, WD 10.37s150,
StartWatch, VirtualVP, VPLive, Win10 Pro
--- Logitech HD Pro C920 webcam (off-line)
--- RIPE Atlas Probe - 32849

Offline dfroula

  • Forecaster
  • *****
  • Posts: 551
Re: Wireless vs. CAT5 connect
« Reply #5 on: May 05, 2014, 09:05:10 PM »
The Ethernet controller chip on the Red controller board operates in half-duplex mode. This means that it shares one Ethernet pair for transmit and receive and is subject to collisions. Half-duplex was originally designed for "Thin net" coaxial cable Ethernet operation and is little-used today. There may be an issue with how your wireless bridge responds to the half-duplex traffic.

There is also a known hardware bug with the Ethernet chip that results in false collision detections in the chip caused by the link pulses. There is no real work-around. It results in lost UDP lightning detection traffic. The lost packet rate is very low, however. You can only see the false "late collisions" on a managed Ethernet switch, which I have. They will not show up on any of the Red web stats or anywhere else, except if the Red is plugged directly into a managed Ethernet switch, typically a Cisco. I worked extensively with Tobi a few months ago to characterize the issue. The solution would be to move to the built-in Ethernet controller on the Discovery board.

Wifi bridges also do not play very well unless all wifi devices on the lan are set to the same compatibility modes. I have an older Motorola Ethernet bridge that chokes significantly under high traffic on the same channel from other wifi networks in my neighborhood.

Regards,

Don

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Wireless vs. CAT5 connect
« Reply #6 on: May 05, 2014, 09:25:50 PM »
Don,
Thanks for this followup and to help explain one of the demons that some of us less-technically trained or equipped guys have.

I, much to my wife's concern, still have the CAT5 running from the controller down the stairs to the router.  I am 30 to 60 per cent efficient now.  I have but the one wireless converter, the NetGear.  I have a little TP link which to date I've not been able to coax into the proper mode, and will try this if I can play with it a bit more.  I assume all the innards are a little bit different, or use some custom chips and certainly custom firmware to run them, all of which may cause an interaction that is unpredicatable.
The sad part is for those of us with some room to spread out but don't have a lot of possible CAT5 runs would love to use wireless to jump from an out building to the router or at least an access point near the end of a run of CAT5 in the house.

Jim suggested a shielded CAT5 from the controller to the NetGear, but that didn't help, nor did another shielded different length cable.  Nor several different unshielded cables of different lengths and sources.

Frustrating, but still fun right now.

Thanks again.  Dale

ECWx.info
&
ECWx.info/t/index.php

Offline scarecrow93

  • Member
  • *
  • Posts: 26
Re: Wireless vs. CAT5 connect
« Reply #7 on: May 06, 2014, 01:23:22 PM »
Dale-

I'm using powerline networking to get the signal from the controller in a second floor spare bedroom to the router on the first floor.  The set up looks like this:

Controller --> 5 Port Switch --> Power Line Network Adapter (upstairs) --> Power Line Network Adapter (downstairs) --> 8 Port Switch --> Asus Wifi Router (one of the 4 ethernet ports on the back) --> Firewall --> Modem --> Internets

I've been using that setup for 3 months now without any problems.  You should be able to take the switches out and it should run just fine.  I have shielded CAT6 going from the controller to the 5 port switch and from the 5 port switch to the powerline network adapter (it was the only shilded stuff I had laying around).

This is the adapter I'm using: http://www.amazon.com/gp/product/B004DVEW8I/

Hope this helps!  I was thinking about the long CAT5 run from upstairs to downstairs too but didn't want to battle CINCHOUSE on the issue. :grin:  This was the best solution I could come up with.

Lance
#724 - Owings, MD
~Lance
#724 - Huntingtown, MD(Aug 2013 - Feb 2014)
Owings, MD(Feb 2014 - Feb 2015)
Annapolis, MD(Late Feb 2015)

Offline JonathanW

  • Engineer
  • Forecaster
  • *****
  • Posts: 323
Re: Wireless vs. CAT5 connect
« Reply #8 on: May 06, 2014, 02:41:57 PM »
The Ethernet controller chip on the Red controller board operates in half-duplex mode. This means that it shares one Ethernet pair for transmit and receive and is subject to collisions. Half-duplex was originally designed for "Thin net" coaxial cable Ethernet operation and is little-used today. There may be an issue with how your wireless bridge responds to the half-duplex traffic.

There is also a known hardware bug with the Ethernet chip that results in false collision detections in the chip caused by the link pulses. There is no real work-around. It results in lost UDP lightning detection traffic. The lost packet rate is very low, however. You can only see the false "late collisions" on a managed Ethernet switch, which I have. They will not show up on any of the Red web stats or anywhere else, except if the Red is plugged directly into a managed Ethernet switch, typically a Cisco. I worked extensively with Tobi a few months ago to characterize the issue. The solution would be to move to the built-in Ethernet controller on the Discovery board.

It would, indeed, be nice to go to a networking standard developed sometime in the last 20 years :)

From the latest system documentation, they discuss upcoming revisions to the controller board that will go to a lower-power, 100 mbps ethernet solution.  I suppose that means I'll be ordering another controller board in the near future (I haven't even received my first one yet!).
« Last Edit: May 06, 2014, 02:43:29 PM by n0ym »

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Wireless vs. CAT5 connect
« Reply #9 on: May 06, 2014, 06:05:53 PM »
Lance,
First, thank you for offering the experience you have gained with a similar situation. 

Second, another thanks for the link to the adapter you have used successfully, I sometimes struggle when someone doesn't give a brand name or model number and simply say, "Oh, there's lots out there, just go check ebay." 

A question for you.  I have a split entrance, as most people do, with 220 being split into two busses, and from old X10 controller experience have learned that if you aren't lucky enough to be on the same 'side' of the split, that connecting with the two units was tough, even with a little bridging capacitor that many recommended.

Were you able to find a socket on the correct side to communicate between the two alright?  Or was it trial and error?  Or does this thing have the problem of split entrances surmounted?

Thanks again.  My current solution isn't working and I see problems unless I can extend the LAN out to the shed with something like this.  Dale
ECWx.info
&
ECWx.info/t/index.php

Offline scarecrow93

  • Member
  • *
  • Posts: 26
Re: Wireless vs. CAT5 connect
« Reply #10 on: May 06, 2014, 08:39:35 PM »
A question for you.  I have a split entrance, as most people do, with 220 being split into two busses, and from old X10 controller experience have learned that if you aren't lucky enough to be on the same 'side' of the split, that connecting with the two units was tough, even with a little bridging capacitor that many recommended.

Were you able to find a socket on the correct side to communicate between the two alright?  Or was it trial and error?  Or does this thing have the problem of split entrances surmounted?

Hi Dale-

It looks like this house is split with the upper floor on one panel and downstairs being on another panel.  I haven't had problems with the power line adapter communicating between the two.  The only thing I have noticed is the speed between the adapters isn't maxed.  Doesn't matter because it's only the controller, printer and a lab computer on the other end anyways.  This is the second different house I've done this in and haven't had to troubleshoot anything.  Just plugged it in and it worked.

I bought those adapters a couple of years ago.  Netgear has newer models out there.

~Lance
~Lance
#724 - Huntingtown, MD(Aug 2013 - Feb 2014)
Owings, MD(Feb 2014 - Feb 2015)
Annapolis, MD(Late Feb 2015)

Offline dfroula

  • Forecaster
  • *****
  • Posts: 551
Re: Wireless vs. CAT5 connect
« Reply #11 on: May 06, 2014, 09:04:10 PM »
I think Jim McMurry in Mauston, WI is using a wireless Ethernet bridge successfully. You might PM him to see how it has been working.

He was using this one:

http://www.netgear.com/home/products/connected-entertainment/gaming-home-theater/WNCE2001.aspx

Regards,

Don

Offline DaleReid

  • Forecaster
  • *****
  • Posts: 2002
    • Weather at Eau Claire, WI
Re: Wireless vs. CAT5 connect
« Reply #12 on: May 06, 2014, 10:11:53 PM »
Don,
Yes he is, but despite his tutoring me, I've not found one of the tricks he used to help.

He suggested using a shielded CAT5 from the NetGear to the controller board, even though the NetGear does not have a grounded socket.  I've tried several different cables to make sure and no such luck as having success.

ECWx.info
&
ECWx.info/t/index.php

 

anything