Dayton Hamvention

 

 

The Mesh Network -- We installed:

·        L-Com HG2409 2.4GHz Omni-Directional Horizontally Polarized wave-guide antennas with Ubiquiti Bullet M2HP radios located:

o   Outside:

§  On the 45 foot tower located on the north side of the East Hall

§  Outside, on the DARA Communications Van Mast

o   Inside the East Exhibit Hall

§  At the Mesh Networks Booth, located in the south-east corner of the hall, near the main East Hall Entrance

§  Hanging from the East Hall Ceiling, room center

·        Linksys 2.4 GHz Nodes (with antennas horizontally polarized, some with after-market 7dbi or 9dbi antennas), located:

o   In the Food Court area, on window sills common to the North Hall (three units)

o   On top of a wall ledge bordering the permanent food court sales area (east end of the food court)- one unit

o   Above the drop ceiling in the BallArena and Silver Arena (7 units)

o   At the Mesh Networks Booth, to support client laptop display computers (3 units)

·        Ubiquiti 5.8 GHz NanoStation nodes (60 degree beam width):

o   At the Mesh Networks Booth, pointed toward the food court

o   At the west end of the Food Court, pointed toward the mesh booth

·        A fixed-view TV Camera (1080p, 4 fps) on the East Hall tower, overlooking the Flea Market.  The camera was direct connected to a Linksys Node, at the Mesh Booth. 

·        Ubiquiti nodes on the tower and in the East Hall were direct-connected with an Ethernet switch located at the Mesh Networks Booth

 

We Learned: 

• Channel 1 (2.4 GHz) was never 100% saturated; I measured it as high as 85% during crowd peak periods.
• Signal to Noise Ratio measurements were made from various points in the East Hall, Food Court, North Hall, BallArena, and Silver Arena.  The Fluke AirCheck meter used enabled us to independently measure S/N ratios for each node.  We found:  
  • The Ubiquiti nodes with high-gain antennas (both 2.4 GHz and 5.8 GHz) had very high S/N ratios, even at crowd peak ..... 40-60 db. 
  • The S/N Ratio of the Linksys nodes did not fare as well, measuring 14 to 44db at crowd peak.
• The three East Hall Ubiquiti nodes (located at our booth, hanging from the East Hall ceiling, and on the outside tower on the north side of East Hall) were “hardwire direct connected”.   From any of them, we could see all 23 of our nodes on the mesh status screen at the start of each day.   All Linksys nodes were RF-connected only.
• By mid day, all the low power Linksys nodes in the BallArena and Silver Arena disappeared from the mesh status screen.  While the Fluke AirCheck showed they were still running ...... it appeared the link between the Linksys nodes and the Ubiquiti nodes in East hall was broken.  I believe links to the Linksys devices were overwhelmed by WiFi interference.  The following experiences support this belief.   
  • As an experiment, a visiting ham placed an additional Linksys node in the gap, between a BallArena Linksys Node and its closest Ubiquiti neighbor, causing most of the Linksys nodes to reappear on the central mesh status screen.   The ability of this link to bear traffic (bit rate) during peak crowd time was not measured.  
  • KD8RBP, Seth, attempted to run traffic over the mesh from the National Traffic System booth located in the ARRL Area of the BallArena, to the Preble County Communication Van.  Seth was using a 2.4 GHz NanoStation.  The path was from his node, to the Linksys nodes located above the ceiling in the BallArena, to the Ubiquiti nodes in East Hall, to the Ubiquiti tower node, to the DARA Van Node, to the Preble County Trailer.  The mesh connection failed, when the link between the BallArena Linksys nodes and the East Hall Ubiquiti nodes failed.  Seth subsequently attempted to pass traffic from the Mesh Networks Booth in East Hall; that worked, and he was able to successfully remote control an HF transceiver in the Preble County Van.  I believe Seth’s experience is consistent with our experience with the Linksys nodes.  
  • The Linksys nodes in the BallArena and Silver Arena were in close proximity to one another ...... and it appears they did function together as a mesh even at crowd peak, even though they became disconnected (isolated) from the East Hall Ubiquiti nodes.
• The 5.8 GHz mesh was never challenged .... no competing stations and no band loading.   The 5.8 GHz NanoStationlocated at the Mesh Networks Booth was “direct-connected” to the 2.4 Ubiquiti nodes.  It backed-up the 2.4 Ubiquiti nodes, and assured a workable, 5.8 GHz high speed communications channel throughout the East Hall and Food Court areas, regardless of what happened on 2.4 GHz.    Measured S/N ratios for the 5.8 Nodes were consistently high, above 40 db.  Unfortunately, we failed to field any clients that took advantage of the 5.8 GHz mesh (the MeshMobilewas intended to do this….).
• Outside, it appears our 2.4 Ubiquiti nodes (on the East Hall tower and on the DARA van mast) were never challenged; we measured high S/N ratios in the Flea Market, and in the vendor/staff parking areas, during crowd peak.
• In a sense, this experience supports the concept of deploying architecture with central Ubiquiti nodes supporting low power nodes around them.  I believe we would have a stronger mesh were we to put a Ubiquiti node in the center of each room.  Do away with using Linksys nodes placed as “Infrastructure”.... and instead use them as mesh gateways, on the floor, enabling people to connect to the mesh.  Maintain mesh integrity using the Ubiquiti node infrastructure across the complex.
• We planned to use proximity, power, and (horizontal) polarity to prevail over interfering WiFi signals.  I now believe “power” (on both 2.4 Hz and 5.8 GHz), the high gain antennas, and the higher receive sensitivity of the Ubiquiti nodes, ruled the day.  Proximity (and density) works, but only if it is high enough that you have no weak links.  The impact (or value) of horizontal polarization was not measured.
• Another lesson learned concerned the Cat5e Ethernet cable used to connect the node and camera on the East Hall tower to the Mesh Networks booth.  We initially installed low-cost copper-clad aluminum (CCA), 24 ga, shielded, twisted pair wire.  The cable tested OK using an Ethernet test device.  We used this wire successfully on other installations, at distances of up to 120 feet.  At Hamvention, we encountered problems with both Power over Ethernet, and with signal transmission, for the Ubiquiti node and the IP Camera located on the East Hall Tower.  Wire length was approximately 200 feet.  We tried moving the PoE units to the tower… and found we could power the devices in that way, but still could not copy digital data at the mesh booth.  Our problems were resolved only when we removed the CCA cable, and installed solid copper, 23 ga, shielded, twisted pair Cat6e cable.  
• Unfortunately, we failed to comprehensively demonstrate “Applications” of the mesh.  Bob Dixon’s demonstration of the integration of Mesh, WiFi, and UHF/VHF communications shined there, but otherwise, we had little to show that we were actually using the mesh to do something… to make something happen.  Only the tower camera video was available on the mesh; KD8RBP’s NTS link failed, we did not deploy the MeshMobile (roving scooter with 2.4 and 5.8 nodes active), nor did we take time to deploy telephones!  We had plans addressing these areas, but failed to follow through when we were consumed with booth traffic.  

The TAPR Forum:

Bob Dixon (W8ERD) and I presented Mesh at the TAPR Forum on Friday Morning; I believe the presentation was well received. I gave a brief introduction to mesh, after which Bob demonstrated (both at the Forum and at our booth) how UHF/VHF, Mesh, and WiFi technologies can be used together by Hams to deliver emergency communications significantly exceeding anything routinely done today.Perhaps more significantly, his work in no way diminished or dismissed the value of technology or capability Hams have deployed for years; rather, he showed how new technology can leverage existing skills and capabilities to greatly improve the service we offer.He caught the attention of many that took the time to look or inquire.Narrated presentation slides are attached to this email.

Surplus Wireless Gear Inc.:

Surplus Wireless Gear Inc., of Olyphant, PA (near Scranton), (www.swginc.net) accepted our invitation to exhibit in a booth adjacent to our Mesh Networks booth.It was their first time at Hamvention.They offered used Ubiquiti equipment for sale, as well as other gear.As far as I know, they were the only inside exhibitor offering broadband wireless equipment at Hamvention.They advised us that they sold out of the more popular Ubiquiti items the first day (Bullets, etc.), and were taking orders for fulfillment from their home base.As they were leaving, they mentioned to me that they definitely wanted to return next year, and would like to have more booth space.I believe this was a good experience for them, and also think their presence was good for mesh, good for Hamvention, and good for Ham Radio.

•  

We owe many thanks to all who made mesh happen at Hamvention; those who supported the planning, those who loaned equipment, those who helped install (and remove) our gear, and those who staffed the booth.We also thank the Hamvention organization, and Hara Arena, for their kind support.

A special vote of thanks is owed to Larry Baker, KB8EMD; he was at Hara Arena almost every day, all day, the week before Hamvention to support equipment install and troubleshooting.Many others gave of their time generously as well…. and that is very much appreciated!!

Conclusion:

All said, I think we learned much…… have much to think about….. and much more to learn.  

73

Bill

 

Bill Curtice - WA8APB

Home: 937-426-7082

Cell: 937-287-0871

Text:  937-234-7975

Skype: williamcurtice