.comment-link {margin-left:.6em;}

Wireless Mesh

Sunday, April 30, 2006

ISO, Media Disk and Old PC - How far down the rabbit whole will you go?

You have to love this Open Source software with a passion or else it can drive you crazy, mad, or just make you feel plain stupid. Locustworld introduced me to the ISO CD burning. ISO is a standard in which data files can burn directly to a CD. The result is a bootable CD that in this case contains OS, application, and systems that will automatically configure a system for you. In Locustworld case it created wireless mesh. In the last two years I have been working with Asterisk PBX phone system. I was able to purchase as FXO FXS Dignam PCI card and start to configure these systems. Then Locustworld created a software module that attached Asterisk to Locustworld. I stop experimenting with Asterisk as a result. As people were moving to VoIP I believed that the Locustworld would have a mature product. Unfortunately it was used for experimentation and although it is functional I was looking for a more robust system and a system that I could upgrade. In two years the development of Asterisk has been outstanding. A product developed from called Asterisk@Home. This product is design similar to Locustworld in that you create a self contain ISO and simply take a used PC and automatically installs and format the PC. In which time the PC becomes a fully functioning PBX with VoIP system there is no wireless or meshing protocols. I was delighted because the original Asterisk was really difficult setup and configure.

So yesterday I figure I would reformat my existing Asterisk installation with the latest and greatest version. The nightmare started all over again. 99% of people out there who use this product will have absolute no difficulty installing the item except for me. Now I not complaining but I ran into some difficulties that I had to write about. Now you may ask yourself how this is related to Locustworld and wireless mesh. Well the problem that occurred was not unique to Asterisk@Home or Locustworld but general problems when using old equipment with new media technology and open source solutions. So if you are doing any open source project you may need to watch out for the following. If I can save you 16 hours of frustration maybe something good will come out of this.

The problem was that the media I used to burn the ISO image onto was 24X CD-R disk. When adding this type of disk to an old PC CD-ROM it tends not to work intermittently or not at all. To resolve the problem I found old 4X stock CD-R disk and reduce the speed of the CD-Writer to create the image. And after that it worked. Here are some of the symptoms I encounter. But before I go on I had exactly the right type make and model of PC that software was recommending.

After burning the ISO image onto the 24X CD I place it into the CD of the PC I that was being formatted for PBX system. It started to configure as expected. Almost near the end of the configuration it abruptly stopped and the software indicated it had a problem with writing some of the files. I rebooted the PC and tried to reinstall. The ISO disk was not being recognized by system. It indicated that there was an error on the hard drive.

I search Google for a solution. Eureka I found the solution. These older CD Readers sometime needs a command prior to launching the software. The command is

linux ide=nodma

Well I tried it and sure enough the boot loader once again worked and I was able to issue the command and yet again it started to work and I was very confident that this solved the problem. It crashed badly.

Then for next few hours the CD did not want to boot up the computer. I produced 4 more copies of the CD and nothing. So I quickly determine it must be the old PC. I remove the hard drive and found a new error message, Not enough memory. Sure enough I started to pull the RAM and replaced with other RAM.

I also added a few other CD ROM thinking it was the CD-ROM. These too were old and could not read the faster media. You know I was doomed.

Next I receive a problem with protected memory. Why on earth would the OS need protected memory? I rebooted constantly in various configuration. I even went to upgrade the Bios on the PC.

Sure nothing I did worked. I replace the PC and though maybe I could sell the PC. Someone could figure out this problem. So I went and took out my old Redhat 9.0 disks that were burnt on 4X CDR. It worked. Well I assumed that Redhat rules. I once again configure the PC and loaded asterisk once again. About 9 o’clock at night I started thinking? And I look at the media speed. I went and found an old unused media that for some reason I never need. I burnt the ISO image one more time, I believe this was copy five, and placed it into the old PC. It worked like a dream. Simple and effortless it configure itself and was up and running with in an hour. I even was able to configure it before I lost my Internet connection (that another story). Asterisk@Home has a good product. Remember when dealing with old equipment to ensure that use the appropriate media with that equipment.

Monday, April 24, 2006

Solar MeshAP 8 days of Operation

In the last week I have 8 days of operation. The temperature was in the high teen and low 20’s. I had sunshine for almost every day except for day 7 and 8. Temperature also went down to 5 degrees. On the 8 day the panel did not have enough efficiency to run the wireless mesh node and charge the battery. I can see that 7/24 solution did work however I can see that additional solar panel is required. Each panel gives me 30 W at 2 amps. Three were used in the experiment and I am adding a forth. That will give 120W at 8 amps. It should be noted that efficiency of the panels are still not optimized and that I not getting the full potential of the Solar array. I will need to angle the solar array to meet the sun’s angles.

From my earlier work last year I noticed that Toronto can get a whole week with out sunshine and that a few batteries will be required. I will also see if I can purchase some deep cycle batteries and controllers. Dual batteries are required to continue operations. I noticed that on sunny days that the batteries only take about 4 -6 hours to fully recharge. This is great news.

Last week I verified in WIANA that the operation was continuous. I noticed that during the middle of the day I had a crash of the system. I believe Locustworld’s Crash Guard system rebooted the Solar MeshAP.

Monday, April 17, 2006

Day 8 Solar Mesh or is it Day 2?

Wireless Mesh Solar Mesh continues to operate. Since the mesh unit died and resurrection on Saturday (you know it had to be Easter) the unit stop for approximately 1 hour. I let the unit continue operating and found that the it came back up and since it was a sunny day on Saturday and Sunday the unit continues to charge it self and run with our concern through the night. So really it day 2 than day 8. Lets call it Day 2.

Saturday, April 15, 2006

Pushing the envelope with Solar Mesh

Well, the solar Mesh went almost seven days with continuous power. Between 8-9 am today the unit went down. This was third day of cloudy or rainy period. I was amazed it lasted this long. By 9am the mesh came back up by itself because it a bright sunny day. I started to think continuous power put I knew that the battery can only hold out so long. It seems that I will need a second battery and possible another solar panel to keep the batteries charge while running the solar mesh.

Thursday, April 13, 2006

5 days and Wireless Solar Mesh keeps going and going

Day five of the Wireless Solar Mesh Experiment and it continues to run even though the weather has been predominately cloudy. The temperature is 20 C today and the sun finally came out this afternoon. My thoughts were yesterday that the solar panels were not able to fully charge the battery while running the wireless mesh. I currently believe that it should stop this evening.

Everything I have bench mark at the beginning leads me to believe that the battery was not fully charged and it need to run through a complete cycle. The cycle I'm talking about is the total discharge of the battery.

As soon as the Solar Mesh stops working then I will try to fully recharge the battery and rerun the bench marking times once again.

Wednesday, April 12, 2006

104 hours or Now 4 Days with continues Operation

It now 104 hours later or 4 days I have continues operation using Solar Wireless Mesh. Today it is about 16 degrees cloud with periods of rain. I didn't see sunshine today so suspect that I will run out of power shortly. Wireless Mesh is still running.

Tuesday, April 11, 2006

80 hours and Solar Mesh is still up

It is now the 80 hour mark and the Solar Wireless Mesh is still performing. The weather today was clear and bright. Temperatures were 15 degree Celsius. My expectation was to get 48 hours of non stop meshing and it has surpassed that. I continue the experiment until it fails.

Monday, April 10, 2006

56 Hours and Counting

The Solar MeshAP using Locustworld is now at the 58 hour mark. I achieve the goal of 48 hours. What is the difference between the second go around? First time I started in the evening about 6 PM this time around I start 10 am. Next I have had almost 3 days of solid sun shine with very little clouds. The Solar Panels are still laying on the ground and nothing has changed other than the weather and the start time.

I will keep this experiment going until the battery looses power. I will mount the solar panel to the pole shortly hopefully it should not stop the operation.

Sunday, April 09, 2006

Longest Link

I resent mention that I have put a link that is over 1 mile long and I had some initial problems with the link. It actually went further than a mile and is linking to most of mesh nodes. The Frenzel Zone and the elevation is high. What happens is that the signal has a clear line of site to most of my nodes; unfortunately some of the nodes are reflect off of large buildings. When I issue a "blocknode" command to the specific IP address I force this new node to route to other nodes rather than going straight to the uplink node. What this has done is maintain a steady amount of bandwidth. As the signal moves further away the amount of bandwidth decreases. As a result the connection is superb.

Clear line of site and no obstruction in the Frenzel Zone is the way to go with Wireless Mesh.

Saturday, April 08, 2006

New Node break record distance

As I was working on the Solar MeshAP I received a call to put up a new Node located about 1.3 Miles or about 1.8 KM away from the nearest MeshAP. The closest node was only 100 m away but due to obstruction of two large building I had to connect it the long way around. Installation went pretty smoothly and new Mesh node connected. Since it was located on the 19th floor and the nodes that connected to them are on hill and on the 7th floor it amazing that I received an 11mbps connection. The new node is saturated with noise I counted over 50 Wi-fi connections and also I noticed that between the new connection there is another 50 Wi-fi connection.

I amazed how well Locustworld OS software works with my equipment. I using a standard 8 dbi antenna with 200 mW radio card. Wow the performance is excellent.

You can see the new connection as it is most southern connection please see map. The new meshAP is located an apartment complex. If you switch to Satellite mode you can see the two building blocking the connection.

I was amazed when the frenzel zone is not obscured that connection can be that fantastic at that distance in a noisy environment. Way to go Locustworld!

Solar Back On

Well I've turn the Solar Power MeshAP and continue to test the Solar system. I have noticed that I getting pretty well good performance from the Wrap board and it is impressive. I just wished it was a little more powerful in MHz.

The solar collectors are still on the ground and I will need to build a frame to mounted them on my fancy pole. Currently the weather in Toronto has been cold and rainy and the performance of the solar panels is not very good. Each solar panel is 2 amps on a bright sunny day like today but on a cloudy day it is less than .5 amp. I currently have 3 and need possible to by a forth. The solar panel is trickling in with it charge and it is taking for ever to get the battery to 100%. This morning at 8 am EDST. I tried to turn the unit back on but alarm indicated that there was not enough power. I had to wait unit 10 AM and it currently working.

Two batteries and a large solar collector is probably required for continued operation. One of my contacts indicated to me that I should verify my work with others on google.

I believe that if I read and understood everything correctly that this unit has sufficient amps and watts to power and run the Solar MeshAP continuous; however, the weather, the location and the amount of sunlight I getting is skewing the problems.

I continue to experiment into the summer.

Tuesday, April 04, 2006

38 Hours and No Juice

Will it looks like the Wrap board, Powerbox (battery), an dsolar Panel went 38 hours before it lost power. The PC Engine Wrap board works with Locustworld build 25 dev 107 and the connection from the Powerbox to the MeshAP is POE. I noticed that it out performed the Via board by approximately 8 hours.

The WIANA reported that the connection was lost some time after 8 AM. I guess if I time it properly now I should start the experiment in the morning to take advantage of the Sun. Currently the temperature in Toronto is 2 degree C. This may have impact on the powerbox performance. I will try this experiment yet again. The power box will use the solar panel to generate as the the adapter that comes with it is very slow in recharging the unit.

I may be able to push 48 hours with this unit. Ideally I looking for perpetual generation.

Monday, April 03, 2006

Enclosure that I have used

Generally I don't like giving out where I bought equipment but here goes.

The metal was purchased at http://www.demarctech.com/products/reliawave-ode/reliawave-rw-ode-8585.html

The plastic enclosure was purchased from http://shop.defactowireless.com/s.nl/sc.2/category.233/.f


What is not shown is the enclosures I use often which is http://enclosures.fintix.com/vj.html This are a fibreglass or fibre composite and they are very strong.

Also I have used Scepter one from www.superpass.com but this does not come with a backing plate.

Check out www.packetgods.com they too have wrap enclosure outdoor and indoor.

Sunday, April 02, 2006

Get Connected

When I created the Video "Get Connected" it was to illustrate and help promote my wireless Mesh in my neighbourhood. Well, the video's call to action is hilarious and I made it to see if people would "Get Connected." Well they have started to use the wireless mesh and it seems that people contact me later to tell me that it a great advertisement.

I would like to start to use MeshAP as an instant out of the box video advertisement. I can see that this would be a fantastic idea to help sell anyone wireless mesh.

Oh here is the video once again

Moskaluk Website Update

In the last month I have updated my website to handle video as well as increase the capacity. Since I do not like to edit my pictures or video I have started to exceed bandwidth prevision of my Web host. As a result I have added more advertisements to my site to off load some of the costs.

Also if you notice that Wireless Mesh bloc I started a year ago January has had tremendous success. Wow Thank You.

If you are interested in any of the articles I have written please see research papers on at www.moskaluk.com/papers.htm

New Moskaluk Prototypes

As I continue with my Solar Experiment I noticed that I have to enclose the low power PC Engine wrap boards. I purchased two distinctly different enclosures one being plastic and the other being metal. These outdoor prototypes can be fasten to walls, pole, lightpoles, almost anything, as seen here.

I have also purchased a NTA's Thin Client and reprogram it with Locustworld's open source. This unit is extremely well built and contains, CF Flash card, PCI riser extension over the motherboard, two fans that can easily be disconnected, RAM, and are one of the best indoor enclosures I have ever seen. I added a senao radio card as well as pigtail and 8 dbi omni directional antenna. It works well. I noticed that the bios did not have the power setting to automatically start up in case of power failure. All in all NTA has very good thin client that can easily be modify to a MeshAP. The unit cost about $149 US. Add in the radio card, pigtail and antenna will drive the cost up another $100 US. So for under $250 dollars you have a very good indoor MeshAP.

I added a Morex case to one of the MeshAP I have previously mentioned as a size comparison.

Solar Experiment

Well I have continue with the Solar Wireless Mesh (SWM) experiment and I have come to the conclusion that I need two batteries and four solar panels. The wrap board has been truly a low power unit and has taken little energy. The simple fact is that 30 hours may not be enough to sustain continues operation in the northern or southern operations. The experiment also has to take in account the sun's angel. I currently have moved my solar panel from a 40 degree pitch to a flat scenario. This is to take in account the Sun's Winter angel, spring, summer and fall's variance in the angles. So far parallel to the earth seems to be working.

I have also taken a few photographs of the operations to give you an illustration of the setup.